energies Article

Hybrid LSA-ANN Based Home Energy Management Scheduling Controller for Residential Demand Response Strategy Maytham S. Ahmed 1,2, *, Azah Mohamed 1 , Raad Z. Homod 3 and Hussain Shareef 4 1

2 3 4

*

Department of Electrical, Electronic and System Engineering, Faculty of Engineering and Built Environments, Universiti Kebangsaan Malaysia, Bangi, Selangor 43600, Malaysia; [email protected] General Directorate of Electrical Energy Production-Basrah, Ministry of Electricity, Basrah 61001, Iraq Department of Petroleum and Gas Engineering, Basrah University for Gas and Oil, Qarmat Ali Campus, Basrah 61004, Iraq; [email protected] Department of Electrical Engineering, United Arab Emirates University, Al-Ain 15551, UAE; [email protected] Correspondence: [email protected]; Tel.: +60-182-279-921

Academic Editor: Giovanni Pau Received: 15 June 2016; Accepted: 30 August 2016; Published: 6 September 2016

Abstract: Demand response (DR) program can shift peak time load to off-peak time, thereby reducing greenhouse gas emissions and allowing energy conservation. In this study, the home energy management scheduling controller of the residential DR strategy is proposed using the hybrid lightning search algorithm (LSA)-based artificial neural network (ANN) to predict the optimal ON/OFF status for home appliances. Consequently, the scheduled operation of several appliances is improved in terms of cost savings. In the proposed approach, a set of the most common residential appliances are modeled, and their activation is controlled by the hybrid LSA-ANN based home energy management scheduling controller. Four appliances, namely, air conditioner, water heater, refrigerator, and washing machine (WM), are developed by Matlab/Simulink according to customer preferences and priority of appliances. The ANN controller has to be tuned properly using suitable learning rate value and number of nodes in the hidden layers to schedule the appliances optimally. Given that finding proper ANN tuning parameters is difficult, the LSA optimization is hybridized with ANN to improve the ANN performances by selecting the optimum values of neurons in each hidden layer and learning rate. Therefore, the ON/OFF estimation accuracy by ANN can be improved. Results of the hybrid LSA-ANN are compared with those of hybrid particle swarm optimization (PSO) based ANN to validate the developed algorithm. Results show that the hybrid LSA-ANN outperforms the hybrid PSO based ANN. The proposed scheduling algorithm can significantly reduce the peak-hour energy consumption during the DR event by up to 9.7138% considering four appliances per 7-h period. Keywords: lightning search algorithm (LSA); home energy management system (HEMS); artificial neural network (ANN); load scheduling; residential demand response (DR)

1. Introduction In recent years, the peak demand has been increasing in the domestic sector and caused unwanted effects to the reliability and stability of power systems. The total energy demand is estimated to increase by 75% at the end of 2020 compared to 2000 [1]. Peak time loads occur in the grid when most end users are using electricity at the same time in a day [2]. In this case, power suppliers are forced to increase generation to meet the high demand, thereby increasing carbon dioxide emission [3], Energies 2016, 9, 716; doi:10.3390/en9090716

www.mdpi.com/journal/energies

Energies 2016, 9, 716

2 of 20

which can promote climate change [4]. Energy consumption in a residential building depends on many factors, such as the number of occupants living in the house, and usage pattern of household [5] as well as period of use, and appliance power [6]. The technology for home energy management system (HEMS) is efficient with data communication networks, which connect home appliances for remote management based on the internet and a combination of the home network to reduce the peak demand that leads to reduced risk of outages at the power distribution and transmission network [7]. A smart home enabled with residential demand response (DR) technologies features a function of HEMS that manages controllable appliances associated with smart socket and meters [8]. DR plays an important role in encouraging residential customers to participate in the distribution system. These programs are designed with an electricity tariff to persuade residential end users to voluntarily decrease their daily electrical consumption pattern or maximize their satisfaction by allocating available resources and effectively managing the electricity loads [9]. Participating customers in DR programs can save on electricity bills when they reduce their electricity usages during peak periods and shifting peak time load to off-peak time. HEMS can assist in the reduction of overall energy consumption through optimal residential load scheduling of appliances and to achieve various goals and functions in homes, such as automatic control, shift, or curtailment of the demand consumption [10]. Many optimization techniques can be used to solve many problems for different applications. Particle swarm optimization (PSO) has been used to minimize the annual total building energy consumption and to improve the building energy performance [11]. Similarly, fuzzy logic control was improved using the quantum lightning search algorithm and backtracking search algorithm to control an induction motor drive [12,13], and a quantum gravitational search optimization algorithm was used to solve the optimal power quality monitor placement problem in power systems [14]. A variety of methods and optimization techniques have been used recently to help end users create optimal appliance scheduling of energy usage based on different feed-in tariffs, pricing schemes, and comfort settings. Kang et al. [15] proposed long-term scheduling and real-time pricing to operate a framework of building an energy management system that included distributed energy storage systems and energy resources to achieve optimal decisions. Optimal energy consumption scheduling based on linear programming computations was applied to minimize the electricity bill and waiting time for each home appliance that operates with real-time pricing tariff [16]. While Haider et al. [17] presented dynamic residential load scheduling and used to achieve optimal scheduling of household appliances to allow end users to decrease energy bills and reduce the peak load. In some related works, neural networks have been used to save electric energy in residential lighting by implementing specific schedules [18]. Pedrasa et al. [19] used stochastic programming approach formulated for robust scheduling of four controllable residential distributed energy resources. The robust schedules were formulated using an improved version of PSO technique to maximize the net benefit of end users as the objective function to reduce electricity bill The method for scheduling home appliances was developed by using a mixed integer nonlinear optimization model built under time-of-use electricity tariff to minimize electricity costs, so that consumers were able to participate in a DR program by making a decision [20]. The PSO algorithm was applied to optimize desirable points during the appliance operation time [21]. Several studies combined the schedules of home appliances with renewable sources. Artificial neural network (ANN) with genetic algorithm has been applied for weekly appliance scheduling with optimized energy consumption in the residential sector to reduce energy demand during peak periods and to maximize the usage of renewable sources [22]. Gharghan et al. [23] hybridized the PSO with ANN to improve the ANN operation by selecting the optimum number of neurons in each hidden layer and learning rate, in which the selection of these parameters was formerly made using the trial and error approach and did not always provide optimum solutions. Neural network is an approach designed to handle any complex nonlinear functions with accuracy through training and learning system input and output. A major challenge faced in scheduling household appliances is the minimization of the energy consumption in a given period without affecting the comfort of customers. However, most previous researchers focused

Energies 2016, 9, 716

3 of 20

on alleviating electricity bills and saving energy without considering the convenience of residential end users. In this study, the lightning search algorithm (LSA) is used to improve the ANN for the home energy management scheduling controller (HEMSC) for the residential DR strategy by modeling the Energies 2016, 9, 716  3 of 19  household appliances. The main contribution of this study focuses on the modeling of household In this study, the lightning search algorithm (LSA) is used to improve the ANN for the home  appliances and developing HEMSC to achieve energy savings in the home on the basis of the energy management scheduling controller (HEMSC) for the residential DR strategy by modeling the  scheduled operation of several appliances according to a focuses  specific Four appliances, household  appliances.  The  main  contribution  of  this  study  on time. the  modeling  of  household  namely, air conditioner (AC), water heaterHEMSC  (WH), refrigerator (REF)savings  and washing machine (WM), arethe  developed appliances  and  developing  to  achieve  energy  in  the  home  on  the  basis  of  scheduled operation of several appliances according to a specific time. Four appliances, namely, air  in Matlab/Simulink according to customer preferences and priority of appliances. conditioner (AC), water heater (WH), refrigerator (REF) and washing machine (WM), are developed  in Matlab/Simulink according to customer preferences and priority of appliances.  2. Load Model of Home Appliances

Load2. Load Model of Home Appliances  modeling is necessary to evaluate residential DR at the distribution circuit and to study customer behavior. Thus, specific home appliance models that describe the dynamics of the process Load modeling is necessary to evaluate residential DR at the distribution circuit and to study  customer behavior. Thus, specific home appliance models that describe the dynamics of the process  to be controlled are important to design. Determining the operating conditions and characteristics to be controlled are important to design. Determining the operating conditions and characteristics of  of household appliances is necessary to develop a HEMSC with residential DR application. In this household  appliances  is  necessary  to  develop  a  HEMSC  with  residential  DR  application.  In  this  study, four selected electrical appliances, namely, AC, WH, WM and REF, are developed using study,  four  selected  electrical  appliances,  namely,  AC,  WH,  WM  and  REF,  are  developed  using  Matlab/Simulink, as shown in Figure 1. Matlab/Simulink, as shown in Figure 1. 

UTILITY AC HEMSC Smart meter

REF

WM

WH

Figure 1. Block diagram of the proposed home energy management scheduling controller (HEMSC) 

Figure 1. Block diagram of the proposed home energy management scheduling controller (HEMSC) system. AC: air conditioner; REF: refrigerator; WM: washing machine; WH: water heater.  system. AC: air conditioner; REF: refrigerator; WM: washing machine; WH: water heater. The signal of DR is assumed to come from the utility to the smart meter and then to the HEMSC  that  includes  and  of  load.  appliances  receive  a  DR  signal  from  the  HEMSC The signal of DRthe  is amount  assumed toduration  come from theAll  utility to thecan  smart meter and then to the HEMSC. The power consumptions of WH, AC, WM, and REF are 3, 2.3, 0.6, and 0.15 kW, respectively.  includes the amount and duration of load. All appliances can receive a DR signal from the The following subsections describe the model details of home appliances. 

that HEMSC. The power consumptions of WH, AC, WM, and REF are 3, 2.3, 0.6, and 0.15 kW, respectively. 2.1. Air Conditioner Modeling  The following subsections describe the model details of home appliances. This  section  presents  the  AC  load  model  development  to  produce  a  load  profile  at  the 

2.1. Air Conditioner distribution Modeling circuit  level.  To  calculate  the  parameters  that  can  be  used  with  a  physical‐based  AC 

model, the mathematical expressions should be derived to obtain an accurate AC load model. The 

This section presents the AC load model development to produce a load profile at the mathematical model is presented as a set of equations to obtain the relationship between the output  distribution circuit level. To calculate the parameters that can be used with a physical-based AC and input parameters, as shown in Figure 2.  model, the mathematical expressions should be derived to obtain an accurate AC load model. The AC unit parameters can be divided into three categories, namely, the characteristics of AC,  the set points of temperatures, and the building structures. The input parameters of the AC model  The mathematical model is presented as a set of equations to obtain the relationship between the are the occupant heat gain  H , room temperature at time t,  , , outside temperature , , set point  output and input parameters, as shown in Figure 2. temperature  , , and the signal of DR  , . The model outputs are the room temperature, which  The is  AC unit parameters can be divided into three categories, namely, the characteristics of used  as  an  input  to  the  model  at  the  subsequent  step  of  time  and  energy  consumption.  Other  AC, the set points of temperatures, and the building structures. The input parameters of the AC parameters, such as the number of people in the home, room size, solar radiation, season, number of  windows, house area, heat gain rate of the house, and cooling load capacity, should be considered    Tout,t , model are the occupant heat gain Hp , room temperature at time t, Tr,t , outside temperature in Simulink.  T , and the signal of DR Sn set point temperature . The model outputs are the room temperature, s,t ewh,t which is used as an input to the model at the subsequent step of time and energy consumption. Other parameters, such as the number of people in the home, room size, solar radiation, season, number of windows, house area, heat gain rate of the house, and cooling load capacity, should be considered in Simulink.

Energies 2016, 9, 716

4 of 20

Energies 2016, 9, 716 

4 of 19  DR signal (Sn,hvac)

Air condition load Model (HVAC)

Tout temp Tout, t (°C) Set point Temp (Ts) (°C)

Power consumption Phvac, t Room temp Tr,t+1

Occupant heat gain Hp Room temp Tr,t (°C)

  Figure 2. Flowchart of the AC load model. 

Figure 2. Flowchart of the AC load model.

In  the  initial  condition,  the  room  temperature  at  time  (t)  should  be  determined  based  on  the 

In the initial condition, the room temperature at time (t) should be determined based on the cooling load factor for glass/corrected cooling load temperature difference [24] as follows:  cooling load factor for glass/corrected cooling load temperature difference [24] as follows: ,

,

d



,

d

Chvac

d

(1) 

 

Qt



Tr,t+of  Tr,t + dt , m + the  house,    is  the  room  temperature  at  time  t ( ),  d   is  the  (1) where    is  the  heat  gain  rate  1 = hvac,t dc dc length of time slot,  d   is the energy that changes the air temperature in the Room 1 ),  ,   is  ◦   is the cooling load capacity (Btu/ ) [25].  the status of AC in the time slot, and  where Qt is the heat gain rate of the house, Tr,t is the room temperature at time t ( C), dt is the length The  output  room  temperature  is  used  as  an  input  room  temperature  to  the  model  at  the  of time slot, dc is the energy that changes the air temperature in the Room 1 (◦ C),AC  mhvac,t is the status following time step. From Equation (1), the heat gain rate of the house,  , is expressed as:  ◦ of AC in the time slot, and Chvac is the cooling load capacity (Btu/ C) [25]. room temperature is used as an input room temperature to the AC model at the The output following time step. From Equation as: (2)  (1), the heat gain rate of the house, Qt , is expressed

,

,




  = SHGC + Hp × Np    Afl Awal A Ace   is the occupant heat gain (btu/h);  win  is the number of people inside a room;    is the  (2) + + + K × S × V × T − T + + ( ) ( ) out,t r,t hos Rfl Rwal Rce Rwin

Qt

where  changes in room air in any time slot;  ,  ,    are the areas of floor, wall, ceiling, and  , + Awins × H solar 2), respectively;  window of the dwelling in (m is the solar heat gain coefficient of a window [26];  , ,  , are  the  average  thermal  resistance  of  the  floor,  wall,  ceiling,  and  window  in  where Hp is the occupant heat gain (btu/h); Np is the number of people inside a room; K is the changes ( m2 h/btu), respectively;  , is the outside temperature ( ) [27];  _   is the window area facing  in room air in2); any time slot; ( Afl , Awall , Ace Awin ) are the areas of floor, wall, ceiling, and window of 2   is the air heat factor (btu/ m, ); and  south (m   is the solar radiation heat power (W/m ).  2 ), respectively; S the dwelling in (m is the solar heat gain coefficient of a window [26]; ( R , R HGC fl wal , To change the room temperature by 1 °C to btu/ , the specific heat of air needs to be specified.  ◦ 2 Rce , A of the floor, wall, ceiling, and window ( C·m ·h/btu), The specific heat capacity of air, C p, is 0.2099/m , and the house volume,  win ) are the average thermal resistance , in  min, is included  in Equation (3):  respectively; Tout,t is the outside temperature (◦ C) [27]; Awin_s is the window area facing south (m2 ); 2 S is the air heat factor (btu/◦ C·m3 ); andbtu Hsolar is the btu solar radiation heat power (W/m ). (3)  d m   ◦ C, the specific To change the room temperature by 1 ◦ C to btu/ heat of air needs to be specified. m The specific heat capacity of air, Cp , is 0.2099/m2 ·◦ C, and the house volume, Vhos , in m3 , is included in The amount of AC power consumed in kW with a thermostat operating in OFF or ON mode and  Equation (3): running at its rated power when switched on at a given interval,  , can be expressed as:       , btu btu ◦ 3 dc ◦ = C,p C × V hos m (4)  (3) C m3

where 

  is the status of the device; 

  = 1 means that the device is turned on, and 

  = 

The amount of AC power consumed in kW with a thermostat operating in OFF or ON mode and 0 means that the device is turned off.    is the AC rated power in kW.  running atThe differences between the set point of AC and the lower or upper limit of the temperatures  its rated power when switched on at a given interval, Phavc,t , can be expressed as: are  called  dead  band.  If  the  room  temperature  decreases  below  a  set  point  minus  the  dead  band  Phvac,t = mhvac × Phvac (4) temperature, then the AC unit is switched OFF. If the room temperature reaches its maximum set  point  plus  the  dead  band  temperature,  then  the  AC  unit  is  switched  ON.  However,  if  the  room  where mhvac is is  thewithin  statusits oftolerable  the device; mhvacthe = AC  1 means thatsame  the status  deviceas  isdescribed  turned on, temperature  band,  then  keeps  the  and mmathematically in Equation (5):  = 0 means that the device is turned off. P is the AC rated power in kW. hvac hvac

The differences between the set point of AC and the lower or upper limit of the temperatures are called dead band. If the room temperature decreases below a set point minus the dead band temperature, then the AC unit is switched OFF. If the room temperature reaches its maximum set point plus the dead band temperature, then the AC unit is switched ON. However, if the room temperature

Energies 2016, 9, 716

5 of 20

is within its tolerable band, then the AC keeps the same status as described mathematically in Equation (5):  Energies 2016, 9, 716 

mhvac Energies 2016, 9, 716 

 =

 Thvac,t < ( Ts,t − ∆T )  Thvac,t > ( Ts,t + ∆T )  × Snhvac,t ∆ , , Ts,t − ∆T ≤ Thvac,t ≤ Ts,t + ∆T

0, 1,

0, 0, 1, 1, ,

mhvac,t−1 ,

5 of 19 

5 of 19 

(5)

(5)  ∆∆ ,   ∆ ∆ ,   (5)  , The electric power demand of the AC load model depends on the DR signal, Snhvac,t . During a DR , ∆ ∆ , , , , . During a  The electric power demand of the AC load model depends on the DR signal,  ,

,

,

∆,

,

, , , ,

, event, this signal, which originates from the revised thermostat set point, can be changed by end users. DR event, this signal, which originates from the revised thermostat set point, can be changed by end  The electric power demand of the AC load model depends on the DR signal,  , . During a  Ts,t is the set point temperature, and ∆T is the dead band temperature (±2 ◦ C). users.  ,   is the set point temperature, and  ∆   is the dead band temperature (±2 °C).  DR event, this signal, which originates from the revised thermostat set point, can be changed by end  By using Equations (1), (4) and (5), we simulate the component models of AC and the mathematical using  Equations  (1),  (4)  and  (5),  we  simulate  the  component  models  of  AC  and  the  users.  By  ,   is the set point temperature, and  ∆   is the dead band temperature (±2 °C).  model inmathematical model in Matlab, as shown in Figure 3.  Matlab, as shown in Figure 3.

By  using  Equations  (1),  (4)  and  (5),  we  simulate  the  component  models  of  AC  and  the  mathematical model in Matlab, as shown in Figure 3. 

  Figure 3. Matlab block for the simulation model of the AC load model. 

Figure 3. Matlab block for the simulation model of the AC load model.   Figure 3. Matlab block for the simulation model of the AC load model.  2.2. Electric Water Heater Modeling

2.2. Electric Water Heater Modeling

The electric water heaters (EWHs) are among the major appliances that consume high energy in  2.2. Electric Water Heater Modeling

Theresidential areas. Energy consumption depends on the amount of hot water that people use at home.  electric water heaters (EWHs) are among the major appliances that consume high energy in The electric water heaters (EWHs) are among the major appliances that consume high energy in  EWH  should  To  obtain  accurate  model  of  EWH  and  to on reflect  goal  of  residential areas. an  Energy consumption depends the any  amount of DR  hot strategies,  water that people usebe  at home. residential areas. Energy consumption depends on the amount of hot water that people use at home.  modeled for use in HEMSC by calculating the input and output parameters, as shown in Figure 4.  To obtain an accurate model of EWH and to reflect any goal of DR strategies, EWH should be modeled To  obtain  an  accurate  model  of  EWH  and  to  reflect  any  goal  of  DR  strategies,  EWH  should  be  for use in HEMSC by calculating the input and output parameters, as shown in Figure 4. modeled for use in HEMSC by calculating the input and output parameters, as shown in Figure 4.  DR signal Sn,ewh

Ambient DR signal temp Tam Sn,ewh (°C)

Ambient temp Flow Tam rate (°C) Flr,t Set point Flow rate temp Tse Flr,t (°C)

Set point Inlet temp temp Tse Tinl (°C) (°C)

Electric water heater Model Electric (EWH) water heater Model (EWH)

Power (Pewh)

Power (Pewh)

(Tout,t+1)

(Tout,t+1)

Tank temp InletTout,t temp (°C) Tinl (°C)

Tank temp Tout,t (°C)

Figure 4. Flowchart of the electric water heater (EWH) load model. DR: demand response. 

A Figure 4. Flowchart of the electric water heater (EWH) load model. DR: demand response.  domestic  EWH  consists  of  a  thermostat  to  sense  temperature  and  OFF/ON  switch  to  heat  Figure 4. Flowchart of the electric water heater (EWH) load model. DR: demand response. water. The WH parameters can be divided into three categories, namely, the set point temperature,  the  use  of  hot  water,  and  the  characteristics.  The  input  and  parameters  the  ambient  A  domestic  EWH  consists  of  a device  thermostat  to  sense  temperature  OFF/ON are  switch  to  heat  temperature,  , water flow rate,  , temperature of inlet water, Tinl,  the set point temperature,  A domestic EWH consists of a thermostat to sense temperature and OFF/ON switch to heat water. , water. The WH parameters can be divided into three categories, namely, the set point temperature,  , temperature of water tank, , and signal of residential DR,  . The model output is the  , , , The WH parameters can be divided into three categories, namely, the set point temperature, the use of the  use  of  hot  water,  and  the  device  characteristics.  The  input  parameters  are  the  ambient 

temperature,  device , water flow rate,  Tinl, the the set point temperature,  hot water, and the characteristics. The input parameters are ambient temperature, Tamp , , , temperature of inlet water, ,

, temperature of water tank,

,

, and signal of residential DR, 

, . The model output is the 

Energies 2016, 9, 716

6 of 20

water flow rate, Flr,t , temperature of inlet water, Tinl, the set point temperature, Tse,t , temperature of water tank, Tout,t , and signal of residential DR, Snewh,t . The model output is the temperature of the EWH tank that is used as an input to the model at the subsequent step of time and power consumption. The data for the EWH model depends on additional parameters, such as the surface area of storage Energies 2016, 9, 716  tank WH, cross-sectional area, and volume that contribute to the design of accurate models. 6 of 19  In the initial condition, the water temperature at time (t) of the EWH should be calculated based temperature of the EWH tank that is used as an input to the model at the subsequent step of time and  on the usage pattern [25]. First, the outlet water temperature of the tank is considered, which is power consumption.  expressed as   The  data  for  the  EWH  model  on  additional  such  as  the  surface  area  of  Tout,tdepends  ×(Voltank − V× dt)+ Tinl × Fparameters,  lr,t ×dt dt Tout,t+1 = + 60 storage tank WH, cross‐sectional area, and volume that contribute to the design of accurate models.  Voltank    In the initial condition, the water temperature at time (t) of the EWH should be calculated based  Atank ×( Tout,t − Tamp ) Pewh,t 3412 BTU − × on  the  usage  pattern  [25].  First,  the  water  kwhtemperature  of Rthe  Voloutlet  tank tank  is  considered,  which  is  tank expressed as 

(6)

3 where Tinl is the inlet water temperature (◦ C), Flr,t isdthe hot water flow rate d at a given interval (m /s), d , , 3 Voltank is the volume of ,the tank (m ), Atank is the surface area of the60 tank, Tamp is the ambient (6)  temperature, Rtank is the heat resistance of ,the 3412 tank BTU (◦ C·m3 ·h/btu), and dt is the duration of the time ,   kwh slot t. The differences between the temperature  set point lower upper limits offlow  the rate  tankat temperature are called   is  the  inlet  water  ( ),  and the  hot  water  a  given  interval  where  ,   is  3/s),  3),  (m   is  the  volume  of  the  tank  (m   is  the  surface  area  of  the  tank,    is  the  dead band. If the water tank temperature drops below the set point lower limit minus the dead band 3 h/btu), and  dt  is the duration of    is the heat resistance of the tank ( ambient temperature,  temperature range, then the EWH coils are switched ON. If the m water tank temperature is raised to its the time slot t.  set point upper limit plus the dead band temperature, then the heating coils of EWH are switched OFF. The differences between the set point lower and upper limits of the tank temperature are called  The EWH operation depends on the device status, mewh , which is mathematically expressed as: dead band. If the water tank temperature drops below the set point lower limit minus the dead band 

  temperature range, then the EWH coils are switched ON. If the water tank temperature is raised to  1, Tewh,t < Tse,t − ∆T its set point upper limit plus the dead band temperature, then the heating coils of EWH are switched    mewh =  0, Tewh,t > Tse,t + ∆T OFF. The EWH operation depends on the device status,  , which is mathematically expressed as:   × Snewh,t mewh,t1,−1 ,



(7)

Tse,t − ∆T, ≤ Tewh,t ∆≤ Tse,t + ∆T ,

0,

,

,

∆

,  

(7) 

where Snewh,t is the DR signal, Tse,t, is ,the set point temperature, ∆ ∆and ∆T is the dead band temperature , , , (±2 ◦ C).where  The electric power demand of the EWH load model depends on the DR signal Snewh,t . , is  the  DR  signal,  ,   is  the  set  point  temperature,  and  ∆   is  the  dead  band  The amount of EWH power consumed in kW depends on the thermostat that operates in the temperature (±2  ). The electric power demand of the EWH load model depends on the DR signal  OFF/ON and runs at its rated power [28]. The power of EWH at a given time is calculated by: states .  , The amount of EWH power consumed in kW depends on the thermostat that operates in the  OFF/ON states and runs at its rated power [28]. The power of EWH at a given time is calculated by:  Pewh,t = mewh × Pewh ,

 

(8) 

(8)

where mewh is the status of the device; and mewh = 1 means that the device is switched on, where    is the status of the device; and    = 1 means that the device is switched on, and    and mewh = 0 means that the device is switched off. Pewh is the EWH rated power in kW. = 0 means that the device is switched off.    is the EWH rated power in kW.  By using Equations (6)–(8), we simulate the component models of EWH and the mathematical By using Equations (6)–(8), we simulate the component models of EWH and the mathematical  model in Matlab, as shown in Figure 5. model in Matlab, as shown in Figure 5. 

  Figure 5. Matlab block for the simulation model of the EWH load model. 

Figure 5. Matlab block for the simulation model of the EWH load model.

Energies 2016, 9, 716 Energies 2016, 9, 716  Energies 2016, 9, 716 

7 of 20

7 of 19  7 of 19 

During a DR event, the signal, which originates from the revised thermostat set point, can be  During a DR event, the signal, which originates from the revised thermostat set point, can be  During a DR event, the signal, which originates from the revised thermostat set point, can be changed by the homeowner. The data for the EWH model depend on the storage tank WH and other  changed by the homeowner. The data for the EWH model depend on the storage tank WH and other  changed by the homeowner. The data for the EWH model depend on the storage tank WH and other parameters that contribute to the design of the EWH physical model.  parameters that contribute to the design of the EWH physical model.  parameters that contribute to the design of the EWH physical model. 2.3. Water Heater and Refrigerator Modeling  2.3.2.3. Water Heater and Refrigerator Modeling  Water Heater and Refrigerator Modeling WM is a home appliance operated by a motor connected to the agitator through a unit called  WM is a home appliance operated by a motor connected to the agitator through a unit called  WM is a home appliance operated by a motor connected to the agitator through a unit called transmission.  WMs  are  divided  into  two  categories  depending  on  the  positioning  of  their  axis,  transmission.  WMs  divided  categories  depending  on positioning the  positioning  of  axis, their namely, axis,  transmission. WMs areare  divided intointo  twotwo  categories depending on the of their namely, vertical axis WM and horizontal axis WM. The power consumption of WM at the residential  namely, vertical axis WM and horizontal axis WM. The power consumption of WM at the residential  vertical axis WM and horizontal axis WM. The power consumption of WM at the residential sector sector accounts for approximately 7.2% of the total electricity consumption [29] and usually consists  sector accounts for approximately 7.2% of the total electricity consumption [29] and usually consists  accounts for approximately 7.2% of the total electricity consumption [29] and usually consists of of an induction motor. In addition, the main electrical component of REF is the compressor, which is  of an induction motor. In addition, the main electrical component of REF is the compressor, which is  anagain an induction motor. Many different approaches are used to model the WM and REF. Real data  induction motor. In addition, the main electrical component of REF is the compressor, which is again an induction motor. Many different approaches are used to model the WM and REF. Real data  again inductionby  motor. Many different approaches areto  used to model the WM REF. Real data are are an measured  using  a  power  quality  analyzer  obtain  accurate  WM and and  REF  models.  are  measured  by  using  a  power  quality  analyzer  to  obtain  accurate  WM  and  REF  models.  measured by using a power quality analyzer to obtain accurate WM and REF models. Matlab/Simulink Matlab/Simulink is developed using resistors and reactance, as shown in Figures 6 and 7, for WM  Matlab/Simulink is developed using resistors and reactance, as shown in Figures 6 and 7, for WM  is developed using resistors and reactance, as shown in Figures 6 and 7, for WM and REF, respectively. and REF, respectively.  and REF, respectively. 

   Figure 6. Simulation model of WM.  Figure 6. Simulation model of WM. Figure 6. Simulation model of WM. 

    Figure 7. Simulation model of REF.  Figure 7. Simulation model of REF.  Figure 7. Simulation model of REF.

The rated power of WM depends on the stage of washing cycles, including washing, rinsing,  The rated power of WM depends on the stage of washing cycles, including washing, rinsing,  The rated power of WM depends on the stage of washing cycles, including washing, rinsing, and spinning, which last typically 55 min to finish the WM job at full load with 53 L of water.  and spinning, which last typically 55 min to finish the WM job at full load with 53 L of water.  and spinning, which last typically 55 min to finish the WM job at full load with 53 L of water.

Energies 2016, 9, 716

8 of 20

Energies 2016, 9, 716 

8 of 19 

Energies 2016, 9, 716 

8 of 19 

3. Gathering Data for Household Appliance Models 3. Gathering Data for Household Appliance Models 

3. Gathering Data for Household Appliance Models  Data were obtained in December 2015 by using a power quality analyzer to measure the power Data were obtained in December 2015 by using a power quality analyzer to measure the power  consumption of AC, WH, WM, and REF as shown in Figure 8 by considering a sample apartment Data were obtained in December 2015 by using a power quality analyzer to measure the power  consumption of AC, WH, WM, and REF as shown in Figure 8 by considering a sample apartment  Energies 2016, 9, 716  8 of 19  houseconsumption of AC, WH, WM, and REF as shown in Figure 8 by considering a sample apartment  inhouse in the town of Kajang, Malaysia as a case study.  the town of Kajang, Malaysia as a case study. house in the town of Kajang, Malaysia as a case study.  3. Gathering Data for Household Appliance Models 

Data were obtained in December 2015 by using a power quality analyzer to measure the power  consumption of AC, WH, WM, and REF as shown in Figure 8 by considering a sample apartment  house in the town of Kajang, Malaysia as a case study. 

  Figure 8. Power quality analyzer to measure the power consumption. 

Figure 8. Power quality analyzer to measure the power consumption.  Figure 8. Power quality analyzer to measure the power consumption.  The input physical model of AC considered three important factors, namely, AC characteristics,  temperature,  and  building  characteristics.  The  temperature  consisted  of  outdoor  and  indoor  set  The input physical model of AC considered three important factors, namely, AC characteristics, The input physical model of AC considered three important factors, namely, AC characteristics,    points. The outside temperature was measured by a temperature and humidity sensor wireless data  temperature, and building characteristics. The temperature consisted of outdoor and indoor set points. temperature,  and  building  characteristics.  The  temperature  consisted  of  outdoor  and  indoor  set  logger connected inside and outside residences, as shown in Figure 9.  Figure 8. Power quality analyzer to measure the power consumption.  The outside temperature was measured by a temperature and humidity sensor wireless data logger points. The outside temperature was measured by a temperature and humidity sensor wireless data  connected inside and outside residences, as shown in Figure 9. logger connected inside and outside residences, as shown in Figure 9.  The input physical model of AC considered three important factors, namely, AC characteristics,  temperature,  and  building  characteristics.  The  temperature  consisted  of  outdoor  and  indoor  set  points. The outside temperature was measured by a temperature and humidity sensor wireless data  logger connected inside and outside residences, as shown in Figure 9. 

  Figure 9. Temperature and humidity wireless data logger. 

 

Solar irradiation (H solar) was measured by using the Apogee Instruments pyranometer. The  Figure 9. Temperature and humidity wireless data logger.  pyranometer is a silicon cell that is sensitive to a portion of the solar spectrum that estimates the total  Figure 9. Temperature and humidity wireless data logger.   radiation across the entire solar spectrum, as shown in Figure 10. The REF data were obtained with a  Solar irradiation (H solar) was measured by using the Apogee Instruments pyranometer. The  freezer temperature of −18    and refrigerator temperature of 3  , as shown in Figure 11. Thermal  Figure 9. Temperature and humidity wireless data logger .  Solar irradiation (H solar) was measured by using the Apogee Instruments pyranometer. pyranometer is a silicon cell that is sensitive to a portion of the solar spectrum that estimates the total  mass in the freezer and REF was combined and includes five bottles of water, two chickens, and 6 kg  radiation across the entire solar spectrum, as shown in Figure 10. The REF data were obtained with a  The pyranometer is a silicon cell that is sensitive to a portion of the solar spectrum that estimates the of fruits and vegetables in the REF at a working time of 24 h. The simulation model output result of  Solar irradiation (H solar) was measured by using the Apogee Instruments pyranometer. The  and refrigerator temperature of 3  , as shown in Figure 11. Thermal  total freezer temperature of −18  radiation across the entire  solar spectrum, as shown in Figure 10. The REF data were obtained REF after calculating the equivalent circuit of the motor depended on the measured data.  pyranometer is a silicon cell that is sensitive to a portion of the solar spectrum that estimates the total  radiation across the entire solar spectrum, as shown in Figure 10. The REF data were obtained with a  with mass in the freezer and REF was combined and includes five bottles of water, two chickens, and 6 kg  a freezer temperature of −18 ◦ C and refrigerator temperature of 3 ◦ C, as shown in Figure 11. of fruits and vegetables in the REF at a working time of 24 h. The simulation model output result of  freezer temperature of −18    and refrigerator temperature of 3  Thermal mass in the freezer and REF was combined and includes, as shown in Figure 11. Thermal  five bottles of water, two chickens, mass in the freezer and REF was combined and includes five bottles of water, two chickens, and 6 kg  REF after calculating the equivalent circuit of the motor depended on the measured data.  and 6 kg of fruits and vegetables in the REF at a working time of 24 h. The simulation model output of fruits and vegetables in the REF at a working time of 24 h. The simulation model output result of  result ofREF after calculating the equivalent circuit of the motor depended on the measured data.  REF after calculating the equivalent circuit of the motor depended on the measured data.

  Figure 10. Apogee Instruments pyranometer solar radiation sensors. 

  Figure 10. Apogee Instruments pyranometer solar radiation sensors.    Figure 10. Apogee Instruments pyranometer solar radiation sensors. 

Figure 10. Apogee Instruments pyranometer solar radiation sensors.

Energies 2016, 9, 716

9 of 20

Energies 2016, 9, 716 

9 of 19 

Energies 2016, 9, 716 

9 of 19 

Figure 11. REF power consumption measurement. 

Figure 11. REF power consumption measurement. Each case of washing, rinsing and spinning has different power consumptions. By contrast, the  Figure 11. REF power consumption measurement.  REF power consumption depends on two cases: the first case has doors closed, and the second case  Each case of washing, rinsing and spinning has different power consumptions. By contrast, denotes frequent opening of the doors. The temperature settings on the REF and freezer are 3 °C and  Each case of washing, rinsing and spinning has different power consumptions. By contrast, the  the REF power consumption depends on two cases: the first case has doors closed, and the second −18 °C, respectively.  REF power consumption depends on two cases: the first case has doors closed, and the second case  case denotes frequent opening of the doors. The temperature settings on the REF and freezer are 3 ◦ C To simulate the model in Matlab, the equivalent circuit of MW and REF should be calculated to  denotes frequent opening of the doors. The temperature settings on the REF and freezer are 3 °C and  ◦ and −18reflect the behavior of the real operation and to obtain an accurate result for both appliances. The  C, respectively. −18 °C, respectively.  To equivalent circuit of the motor can be represented as impedances, as shown in Figure 12.  simulate the model in Matlab, the equivalent circuit of MW and REF should be calculated To simulate the model in Matlab, the equivalent circuit of MW and REF should be calculated to  to reflect the behavior of the real operation and to obtain an accurate result for both appliances. reflect the behavior of the real operation and to obtain an accurate result for both appliances. The  The equivalent circuit of the motor can be represented as impedances, as shown in Figure 12. equivalent circuit of the motor can be represented as impedances, as shown in Figure 12. 

  Figure 12. Equivalent circuit of WM and REF. 

  4. Artificial Intelligent Techniques Used for Home Energy Management Scheduling Controller  Figure 12. Equivalent circuit of WM and REF.  In this section, ANN, LSA, and the proposed hybrid LSA‐ANN are discussed.  Figure 12. Equivalent circuit of WM and REF.

4. Artificial Intelligent Techniques Used for Home Energy Management Scheduling Controller  4.1. Artificial Neural Network Technique 

4. Artificial Intelligent Techniques Used for Home Energy Management Scheduling Controller In this section, ANN, LSA, and the proposed hybrid LSA‐ANN are discussed.  An ANN is an information processing paradigm that models nonlinear systems and attempts to 

In this section, ANN, LSA, of  and the proposed are discussed. simulate  the  functionality  the  human  brain. hybrid Neural LSA-ANN networks  have  many  unique  benefits,  4.1. Artificial Neural Network Technique  especially with the complex nonlinear relationships between system input and output, which handle 

4.1. Artificial Neural Network Technique any complex nonlinear functions through training and learning system input and output. 

An ANN is an information processing paradigm that models nonlinear systems and attempts to  In  this  study,  a  feed‐forward  neural  network  type  and  the  Levenberg–Marquardt  training  simulate  functionality  of processing the  human paradigm brain.  Neural  have  many  unique and benefits,  Analgorithm are selected for training the ANN in the Matlab toolbox. The ANN structure consists of  ANNthe  is an information thatnetworks  models nonlinear systems attempts especially with the complex nonlinear relationships between system input and output, which handle  to simulate the functionality of the human brain. Neural networks have many unique benefits, five  inputs  ( , , , , DR),  two  hidden  layers  with  the  activation  function  as  sigmoid  any complex nonlinear functions through training and learning system input and output.  function,  and complex four  outputs  (AC,  WH,  WM,  and  between REF).  The system actual  data  collected  the  handle especially with the nonlinear relationships inputare  and output,from  which In  this  study,  a  feed‐forward  neural  network  type  and  the  Levenberg–Marquardt  training  simulation system, which represents the training data of the ANN, as shown in Figure 13.  any complex nonlinear functions through training and learning system input and output. algorithm are selected for training the ANN in the Matlab toolbox. The ANN structure consists of  In this study, neural type with  and the  theactivation  Levenberg–Marquardt training five  inputs  ( ,a feed-forward , , , DR),  two network hidden  layers  function  as  sigmoid  algorithm are and  selected for training ANN theREF).  Matlab Theare  ANN structure function,  four  outputs  (AC, the WH,  WM, inand  The toolbox. actual  data  collected  from  consists the  of five inputs (T , T , T , T , DR), two hidden layers with the activation function as sigmoid ac ot simulation system, which represents the training data of the ANN, as shown in Figure 13.  im wh

function, and four outputs (AC, WH, WM, and REF). The actual data are collected from the simulation system, which represents the training data of the ANN, as shown in Figure 13.

Energies 2016, 9, 716

10 of 20

Energies 2016, 9, 716 

10 of 19  Input layer

Hidden layer1

Tac Twh

Tot

Tim

Hidden layer 2

1

1

2

2

3

3

4

4

5

5

6

6

N

M

Output layer WH AC WM

REF

DR

 

Figure 13. Architecture of the artificial neural network (ANN) algorithm. 

Figure 13. Architecture of the artificial neural network (ANN) algorithm.

All inputs and outputs of ANN‐based HEMSC can be expressed by Equations (9) and (10): 

All inputs and outputs of ANN-based HEMSC can be expressed by Equations (9) and (10): DL DL : Tim1 DL1   Tac1 Twh1 Tot1  T T :  ac2 wh2 Tot2 Tim2 DL2 DL 





Input =   

     

Input

:

AC WH REF WM : AC WH REF WM T T T whn otn : Timn DL  n Output acn :   WH 1 REF AC1 WH AC REF1WM WM1

(9) 

(9) (10) 

); 2 WH   is the WH temperature;    is the total power  where    denotes the room temperature ( 2 REF2 WM2   AC   Output =  (10) : DR   is the signal of a DR event.  consumption;    denotes the time of the system; and     In this system, the DR starts from between 16 h and 23 h, and the value of demand limit (DL) is    : assumed to be 3 kW. If    is higher than DL, then the controller will switch OFF the lower priority  ACn WHn REFn WMn appliance;  otherwise,  the  system  works  normally.  The    and    of  the  system  are  used  to  whereevaluate the comfort level of the end users for AC and WH, respectively. The outputs of the ANN are  Tac denotes the room temperature (◦ C); Twh is the WH temperature; Tot is the total power consumption; Tim denotes the time of the system; and DR is the signal of a DR event. the signals to turn the four home appliances ON or OFF according to customer preferences, comfort  level, and priority of appliances. Sudden changes in the home appliances can be predicted by using ANN.  In this system, the DR starts from between 16 h and 23 h, and the value of demand limit (DL) is





assumed to be 3 kW. If Tot is higher than DL, then the controller will switch OFF the lower priority 4.2. Overview of Lightning Search Algorithm  appliance; otherwise, the system works normally. The Tac and Twh of the system are used to evaluate Optimization is a process to find the best solution to problems depending on the input variables  the comfort level of the end users for AC and WH, respectively. The outputs of the ANN are the after  determining  objective  function  to  constraints.  The  objective  function  is  often level, signals to turn the fourthe  home appliances ONsubjected  or OFF according to customer preferences, comfort formulated  based  on  a  certain  application  and  can  take  the  form  of  minimal  error,  minimal  cost,  and priority of appliances. Sudden changes in the home appliances can be predicted by using ANN.

optimal design, and optimal management. LSA is a new optimization algorithm based on the natural  phenomenon  of  lightning  and  it  is  inspired  by  the  probabilistic  nature  and  sinuous  4.2. Overview of Lightning Search[29],  Algorithm characteristics of lightning discharges during a thunderstorm. LSA is organized from the mechanism  Optimization is a process to find the best solution to problems depending on the input variables of step leader propagation. This algorithm considers the participation of fast particles (projectiles) in  after the figuration of the binary tree structure of a step leader. Similar to other metaheuristic algorithms,  determining the objective function subjected to constraints. The objective function is often formulated based ona apopulation  certain application and can take the formsuggests  of minimal error, minimal LSA  also  needs  to  begin  the  search.  The  projectile  random  solutions  for cost, corresponding problems to be solved by LSA. More details about LSA and basic ideas can be found  optimal design, and optimal management. LSA is a new optimization algorithm based on the natural in [30,31]. The step leaders are formed in the first phase because transition projectiles are ejected from  phenomenon of lightning [29], and it is inspired by the probabilistic nature and sinuous characteristics the thunder cell in a random direction. Thus, the formula of uniform probability distribution is used  of lightning discharges during a thunderstorm. LSA is organized from the mechanism of step leader for a random number of step leaders. The standard uniform distribution can be formulated as follows: 

propagation. This algorithm considers the participation of fast particles (projectiles) in the figuration of the binary tree structure of a step leader. 1Similar to aother metaheuristic algorithms, LSA also , for b   (11)  b a needs a population to begin the search. The projectile suggests random solutions for corresponding 0, elsewhere problems to be solved by LSA. More details about LSA and basic ideas can be found in [30,31]. The step leaders are formed in the first phase because transition projectiles are ejected from the thunder cell in a random direction. Thus, the formula of uniform probability distribution is used for a random number of step leaders. The standard uniform distribution can be formulated as follows:

Energies 2016, 9, 716

11 of 20



(



f xT =

1 , (b−a)

0,

for a < x T ≤ b

)

elsewhere

(11)

where x T is a random number that may provide a solution; and a and b are the lower and upper bounds of the solution space, respectively. The probability density function f ( x s ) of an exponential distribution [31] is shown by Equation (12): ( s

f (x ) =

xs

1 ( (µ) ) µe

0,

, for x s ≥ b for x s ≤ b

) (12)

After the initial is evaluated, the position and direction are updated with Equation (13): pis−NEW = pis ± exprand (µi )

(13)

where pis−NEW is the new projectile, and pis is the old projectile. The projectiles and the step leaders that have traveled close to the ground do not have adequate potential to ionize large sections in front of the leading edge. In this way, the lead projectile can be formulated as a random number taken from the standard normal distribution. The normal probability density function f ( x L ) is expressed as: L

f (x ) =



2 L 2 1 √ e−( x −µ) /2σ σ 2π

 (14)

where f ( x L ) is the normal probability density function, σ is the scale parameter, and µ is the shape parameter. From Equation (14), the randomly generated lead projectile can search in all directions from the current position defined by the shape parameter. The scale parameter σ decreases exponentially to find the best solution. Therefore, the position of p L in step + 1 can be shown in Equation (15) [31]: L pNEW = p L + normrand (µ L , σ L )

(15)

L where pNEW is the new lead projectile.

4.3. Proposed Hybrid Lightning Search Algorithm-Based Artificial Neural Network The ANN algorithm can be used to control the appliances in HEMSC. The comfort level of end users can be utilized as the inputs of the ANN to determine and improve the suitable ON/OFF status of appliances and schedule another time without affecting the convenience of end users in the devices. The learning rate and the neurons in each hidden layer in the ANN architecture are the significant parameters. However, the selection of the learning rate and the neurons are subject to trial-and-error processes, which do not give the optimal solution. LSA addresses such a problem to enhance the ANN performance by finding the optimum learning rate and the best value of neurons in each hidden layer of the neural network that can be used in home energy scheduler controller. The implementation starts by resetting the LSA parameters, namely, number of iterations (T), population size (N), problem dimension (D), and channel time. Each step leader in this algorithm contains three components, namely, learning rate (LR), number of neurons in the first hidden layer (N1), and number of neurons in the second hidden layer (N2). The obtained values of LR, N1 and N2 are used in the ANN training to minimize the error of ON/OFF devices status in HEMSC. The flow chart of the proposed hybrid LSA-ANN

Energies 2016, 9, 716

12 of 20

is shown in Figure 14. The mean absolute error (MAE) is used as objective function to improve the performance of the ANN by minimizing the error function, as shown in the following equation: n

MAE =

1 (error )2 n i=1

∑

(16)

where error = sa − se , sa is the actual status, se is the estimated status of HEMSC, and m is the number of samples. Energies 2016, 9, 716  12 of 19  Start Reset LSA parameters: number of iteration (T), size of the population (N), number of problem dimension (D), channel time Generate initial population of step leaders Yes

i>N Reach Maximum population? No i=i+1

Choice N1, N2, LR and Run ANN Calculate objective function (MAE) for each step leader (Pij) using Eq. (16) Yes

t>T Reach maximum iteration? No

t=t+1

Reach maximum channel time? No Update best and worst step leaders

Eliminate bad channel (Move step leader from worst to best)

Yes

Reset channel time

Update kinetic energy, Ep and Update direction Yes

i>N Reach maximum step leaders?

i=i+1

No Calculate distance (dist) between populations

Calculate objective function (MAE) for each step leader (Pij) using Eq. (16) Choice N1, N2, LR and Run ANN

j>D Reach maximum problem dimension? No (dist) = 0

Yes

Update position using Eq.(15) j=j+1

Yes No

Update position using Eq.(13)

Check the boundary control mechanism for new position Extend channel (Ep>Esl)? No Remain position

Yes

Forking occur? No New position

Yes

Create two symmetrical channels at fork point Eliminate channel which has lower energy

Output the optimal N1, N2 and LR

  Figure 14. Flowchart of the proposed hybrid LSA-ANN. Figure 14. Flowchart of the proposed hybrid LSA‐ANN.  5. Overall Proposed Home Energy Management Scheduling Controller System  5. Overall Proposed Home Energy Management Scheduling Controller System The proposed HEMSC algorithm is developed such that it can control and schedule the WH, 

The proposed HEMSC algorithm is developed such that it can control and schedule the WH, AC, WM and REF, and switch customer load to decrease the costs of electrical power consumption  AC, WM and REF, and switch customer load to decrease the costs of electrical power consumption during DR event. The HEMSC algorithm starts by reading the data and information of all the above‐ during mentioned  DR event.appliances.  The HEMSC algorithm starts byon  reading thepoints,  data including  and information of all the Each  appliance  is  compared  several  set  load  priority,  power consumption, and customer preference, by settings on the room temperature of AC and the  water temperature of WH. The entire system is implemented by Matlab/Simulink, as shown in Figure 15. 

Energies 2016, 9, 716

13 of 20

above-mentioned appliances. Each appliance is compared on several set points, including load priority, power consumption, and customer preference, by settings on the room temperature of AC and the water temperature of WH. The entire system is implemented by Matlab/Simulink, as shown in Figure 15. Energies 2016, 9, 716  13 of 19  Energies 2016, 9, 716 

13 of 19 

  Figure 15. Matlab block implementation for the simulation model of the overall system.    Figure 15. Matlab block implementation for the simulation model of the overall system. Figure 15. Matlab block implementation for the simulation model of the overall system. 

6. Results and Discussion  6. Results and Discussion 6. Results and Discussion  This section describes the results of the simulation models of the home appliances with experimental  This section describes the results of the simulation models of the home appliances with results and the hybrid LSA‐ANN results for the home energy management scheduling controller.  This section describes the results of the simulation models of the home appliances with experimental  experimental results and the hybrid LSA-ANN results for the home energy management results and the hybrid LSA‐ANN results for the home energy management scheduling controller.  scheduling controller. 6.1. Home Appliance Simulation Result  6.1. Home Appliance Simulation Result  6.1. Home Appliance Simulation Result The following subsections describe the simulation model results of the modeled home appliances.  The following subsections describe the simulation model results of the modeled home appliances.  The following subsections describe the simulation model results of the modeled home appliances. 6.1.1. Water Heater Simulation Result  6.1.1.A case study is conducted to illustrate the performance model of the WH. This case study shows  Water Heater Simulation Result 6.1.1. Water Heater Simulation Result  the  hot  water  usage  of  WH  at  different the times,  as  in  Figure  16a.  In  WH. Figure  16b,  the  maximum  A case study is conducted to illustrate performance model of the This case study shows A case study is conducted to illustrate the performance model of the WH. This case study shows  temperature is assumed to be 48 °C, and the minimum temperature of the WH setting is assumed to  the hot usage of WH at different times,times,  as in Figure In Figure 16b,Figure  the maximum the  hot water water  usage  of  WH  at  different  as  in 16a. Figure  16a.  In  16b,  the temperature maximum  be 42 °C. These values can be altered in the physical model according to the preference of customers.  is assumed to be 48 ◦ C, and the minimum temperature of the WH setting is assumed to be 42 ◦ C. temperature is assumed to be 48 °C, and the minimum temperature of the WH setting is assumed to  These values can be altered in the physical model according to the preference of customers. be 42 °C. These values can be altered in the physical model according to the preference of customers. 

(a) 

(b)

   

(a)  (b) Figure 16. Simulation model of EWH load: (a) flow rate of the hot water in gpm; and (b) hot water  temperature within 42–48 °C with the power consumption pattern.  Figure 16. Simulation model of EWH load: (a) flow rate of the hot water in gpm; and (b) hot water  Figure 16. Simulation model of EWH load: (a) flow rate of the hot water in gpm; and (b) hot water temperature within 42–48 °C with the power consumption pattern. temperature within 42–48 ◦ C with the power consumption pattern. 

When the hot water is used at 7:00 a.m. and the temperature reaches its minimum allowable set  point of 42  , the WH will switch ON to keep the water temperature at its comfortable range. When  When the hot water is used at 7:00 a.m. and the temperature reaches its minimum allowable set  the hot water is used between 4:00 p.m. and 6:00 p.m., the WH will switch ON again to maintain the  point of 42  , the WH will switch ON to keep the water temperature at its comfortable range. When  water temperature in the tank until the temperature reaches its maximum allowable set point of 48  the hot water is used between 4:00 p.m. and 6:00 p.m., the WH will switch ON again to maintain the  ,  to switch  OFF  the  WH. When  the  temperature  of  the  water  in  the  tank  is within 42–48  ,  the  water temperature in the tank until the temperature reaches its maximum allowable set point of 48  heater switch status will maintain the previous device state.  ,  to switch  OFF  the  WH. When  the  temperature  of  the  water  in  the  tank  is within 42–48  ,  the 

Energies 2016, 9, 716

14 of 20

When the hot water is used at 7:00 a.m. and the temperature reaches its minimum allowable set point of 42 ◦ C, the WH will switch ON to keep the water temperature at its comfortable range. When the hot water is used between 4:00 p.m. and 6:00 p.m., the WH will switch ON again to maintain Energies 2016, 9, 716  14 of 19  the water temperature in the tank until the temperature reaches its maximum allowable set point of 48 ◦ C, to switch OFF the WH. When the temperature of the water in the tank is within 42–48 ◦ 14 of 19  C, Energies 2016, 9, 716  6.1.2. Air Conditioner Simulation Results  the heater switch status will maintain the previous device state.

5 6 4 5 3 4 2 3 1 2 0 10 12 14 16 18 20 22 241

25 30 20 25 15 20 10 15 5 100

2

4

6

8

Time ( hour)

Power consumption Power consumption (kW) (kW)

C C Temperature, Temperature,

In the simulation, the maximum and minimum temperatures of AC are set at 28 °C and 18  ,  6.1.2. Air Conditioner Simulation Results  6.1.2. Air Conditioner Simulation Results respectively. These values can be changed in the physical model according to the preference of the  In the simulation, the maximum and minimum temperatures of AC are set at 28 °C and 18  ,  In the simulation, the maximum and minimum temperatures of AC are set at 28 ◦ C and 18 ◦ C, customer. Real data are measured to be used as input for the AC load model that includes the H solar  respectively. These values can be changed in the physical model according to the preference of the  These values can be in the physical model according to the preference of the and  respectively. outside  temperature,  both  of changed which  are  measured  in  Kajang,  Malaysia.  According  to  the  customer. Real data are measured to be used as input for the AC load model that includes the H solar  customer. Real data are measured to be used as input for the AC load model that includes H homeowner comfort setting of the room temperature, which is set between 28 °C and 18  the , the AC  and  solar outside  temperature,  both  of  which  are  measured  in  Kajang,  Malaysia.  According  to  and outside temperature, both of which are measured in Kajang, Malaysia. According to the the  Simulink is shown in Figure 17.  homeowner comfort setting of the room temperature, which is set between 28 ◦ C and 18 ◦ C, the, the AC  AC homeowner comfort setting of the room temperature, which is set between 28 °C and 18  30 Simulink is shown in Figure 17. 6 Simulink is shown in Figure 17. 

 

5 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Figure 17. Simulation model of AC power consumption pattern with room temperature.

 

Time ( hour)

 

Figure 17. Simulation model of AC power consumption pattern with room temperature.   Figure 17. Simulation model of AC power consumption pattern with room temperature. Figure 17 shows that if the room temperature reaches its minimum set point temperature of 18  ,  then  the  AC  is  switched  OFF.  When  the  room  temperature  reaches  its  maximum  set  point  Figure 17 shows that if the room temperature reaches its minimum set point temperature of 18  Figure 17 shows that if the room temperature reaches its minimum set point temperature of 18 ◦ C, temperature of 28  , then the AC is switched ON to keep the room temperature in its comfortable  ,  then  the AC AC  switched  the  room  temperature  reaches  its  maximum  set  point  then the is is  switched OFF.OFF.  When When  the room temperature reaches its maximum set point temperature range. When the room temperature is within 18–28  , the switch status will maintain the previous  ◦ temperature of 28  , then the AC is switched ON to keep the room temperature in its comfortable  of 28 C, then the AC is switched ON to keep the room temperature in its comfortable range. When the device state.  range. When the room temperature is within 18–28  room temperature is within 18–28 ◦ C, the switch status , the switch status will maintain the previous  will maintain the previous device state. device state.  6.2. Experimental Measurement Data 

6.2. Experimental Measurement Data

C C Temperature, Temperature,

The temperatures outside and inside the building were measured by using a wireless data logger  The temperatures outside and inside the building were measured by using a wireless data logger 6.2. Experimental Measurement Data  sensor,  and and the the measured  outdoor  is shown shown inin Figure Figure  The  solar  irradiation  sensor, measured outdoortemperature  temperature is 18.18.  The solar irradiation was was  The temperatures outside and inside the building were measured by using a wireless data logger  measured using the Apogee instrument pyranometer sensors and the measured data are shown in measured using the Apogee instrument pyranometer sensors and the measured data are shown in  sensor,  and  measured  outdoor  temperature  is  shown  in  Figure  solar  irradiation  was  Figure 19.the  Both the temperature and solar irradiation were used as inputs18.  to The  the AC model. Figure 19. Both the temperature and solar irradiation were used as inputs to the AC model.  measured using the Apogee instrument pyranometer sensors and the measured data are shown in  36 Figure 19. Both the temperature and solar irradiation were used as inputs to the AC model.  Temp 34 36 32 34 30 32 28 30 26 28 24 260

Temp

2

4

6

8

10

12

14

Time (hour)

16

18

20

22

24

 

24 0 2 4 6 8 10 12 14 16 18 20 22 24 Figure 18. Measured outdoor temperature. 

Figure 18. Measured outdoor Time (hour) temperature.  

Solar irradiation radiation (w/m2) (w/m2)

1000 Figure 18. Measured outdoor temperature.  H solar 800 1000 600 800 400 600 200 400

H solar

Temper

30 28 26 24 0

2

4

6

8

Energies 2016, 9, 716

10

12

14

Time (hour)

16

18

20

22

24

 

15 of 20

Figure 18. Measured outdoor temperature. 

Solar irradiation (w/m2)

1000

H solar

800 600 400 200 0 0

2

4

6

8

10

12

14

Time (hour)

16

18

20

22

24

 

Figure 19. Measured solar irradiation data.  Figure 19. Measured solar irradiation data.

Other data from WM and REF power consumption were also considered. According to the real  Other data from WM and REF power consumption were also considered. According to the real

measurement  of  WM,  55.5  min  was  needed  to  complete  the  job  with  three  different  energy  Energies 2016, 9, 716  15 of 19  Energies 2016, 9, 716  15 of 19  measurement of WM, 55.5 min was needed to complete the job with three different energy consumption

Power consumption in kW Power consumption in kW

levels during washing, rinsing, and spinning. In the washing cycle, three intervals at 11.5, 2.25 and consumption levels during washing, rinsing, and spinning. In the washing cycle, three intervals at  consumption levels during washing, rinsing, and spinning. In the washing cycle, three intervals at  2.5 min were needed to finish the task. In the rinsing cycle, three duration times were needed to 11.5, 2.25 and 2.5 min were needed to finish the task. In the rinsing cycle, three duration times were  11.5, 2.25 and 2.5 min were needed to finish the task. In the rinsing cycle, three duration times were  finish the with a job,  1 min interval. Finally, the last cyclethe  needed 2.5, 2.5 and 6min to complete needed  to job, finish  the  with  a  1  min  interval.  Finally,  last  cycle  needed  2.5,  2.5  and  6min the to  needed  to  finish  the  job,  with  a  1  min  interval.  Finally,  the  last  cycle  needed  2.5,  2.5  and  6min 20. to  job. The experimental measurement data of the WM power consumption are shown in Figure complete the job. The experimental measurement data of the WM power consumption are shown in  complete the job. The experimental measurement data of the WM power consumption are shown in  The power consumption curve of the REF was measured every minute, as shown in Figure 21. Figure 20. The power consumption curve of the REF was measured every minute, as shown in Figure 21.  Figure 20. The power consumption curve of the REF was measured every minute, as shown in Figure 21.  0.7 0.7 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 00 0

5 10 15 20 25 30 35 40 45 50 55 60 5 10 15 20Time 25 (minute) 30 35 40 45 50 55 60 Time (minute)

   

Figure 20. Actual power consumption curve of WM load.  Figure 20. Actual power consumption curve of WM load. Figure 20. Actual power consumption curve of WM load.  Power consumption (kW) Power consumption (kW)

0.17 0.17 0.15 0.15 0.1 0.1

0.05 0.05 0 00 0

5 10 15 20 25 30 35 40 45 50 55 60 5 10 15 20Time 25 (minute) 30 35 40 45 50 55 60 Time (minute)

Figure 21. Actual power consumption curve of REF load.  Figure 21. Actual power consumption curve of REF load.  Figure 21. Actual power consumption curve of REF load.

6.3. Results of the Hybrid Lightning Search Algorithm‐Based Artificial Neural Network  6.3. Results of the Hybrid Lightning Search Algorithm‐Based Artificial Neural Network  6.3. Results of the Hybrid Lightning Search Algorithm-Based Artificial Neural Network The inputs of ANN include the room temperature  , WH temperature  , time of the system  The inputs of ANN include the room temperature  , WH temperature  , time of the system    and total power consumption of the system . The signals are output to ON/OFF the WH, AC,  The inputs of ANN include the room temperature Tac , WH temperature Twh , time of the system   and total power consumption of the system . The signals are output to ON/OFF the WH, AC,  WM, and REF. By using the ANN training and testing data, we can find all other parameters, such as  Tim and total power consumption of the system Tot . The signals are output to ON/OFF the WH, AC, WM, and REF. By using the ANN training and testing data, we can find all other parameters, such as  inputs, number of neurons in each hidden layer, number of hidden layers, weights, learning rate, and  WM, and REF. By using the ANN training and testing data, we can find all other parameters, such as inputs, number of neurons in each hidden layer, number of hidden layers, weights, learning rate, and  output. LSA searches the best values for learning rate and the number of neurons in each hidden  inputs, number of neurons in each hidden layer, number of hidden layers, weights, learning rate, and output. LSA searches the best values for learning rate and the number of neurons in each hidden  layer to enhance the ANN performance. The objective function for 10, 20, 30, 40 and 50 population  output. LSA searches the best values for learning rate and the number of neurons in each hidden layer layer to enhance the ANN performance. The objective function for 10, 20, 30, 40 and 50 population  sizes can be obtained, as shown in Figure 22. Several populations are executed to permit the LSA to  to enhance the ANN performance. The objective function for 10, 20, 30, 40 and 50 population sizes can sizes can be obtained, as shown in Figure 22. Several populations are executed to permit the LSA to  select the population size to achieve the minimum error and the consumption time.  be obtained, as shown in Figure 22. Several populations are executed to permit the LSA to select the select the population size to achieve the minimum error and the consumption time.  population size to achieve the minimum error and the consumption time. -7

Objective function Objective function

x 10 6 x 10-7 6 5 5 4 4 3 3 2 2 1

LSA10 LSA10 LSA20 LSA20 LSA30 LSA30 LSA40 LSA40 LSA50 LSA50

WM, and REF. By using the ANN training and testing data, we can find all other parameters, such as  inputs, number of neurons in each hidden layer, number of hidden layers, weights, learning rate, and  output. LSA searches the best values for learning rate and the number of neurons in each hidden  layer to enhance the ANN performance. The objective function for 10, 20, 30, 40 and 50 population  sizes can be obtained, as shown in Figure 22. Several populations are executed to permit the LSA to  Energies 2016, 9, 716 16 of 20 select the population size to achieve the minimum error and the consumption time.  -7

Objective function

6

x 10

LSA10 LSA20 LSA30 LSA40 LSA50

5 4 3 2 1 0 0

20

40 60 Iterations

80

100

 

Figure 22. Objective function with iteration of the hybrid LSA‐ANN for different population sizes.  Figure 22. Objective function with iteration of the hybrid LSA-ANN for different population sizes.

Figure 22 illustrates that the number of populations (N) as 40 can give the best result for the LSA  Figure 22 illustrates that the number of populations (N) as 40 can give the best result for the LSA than the population size numbers of 10, 20 and 30 because a high error and the objective function are  than the population size numbers of 10, 20 and 30 because a high error and the objective function are −9 after 40 iterations. The population size of 50 needs more working time achieved in error 9.128 × 10 achieved in error 9.128 × −9 10 after 40 iterations. The population size of 50 needs more working time  Energies 2016, 9, 716  16 of 19  than the population size of 40. The ANN parameters based on the results of the hybrid LSA-ANN are than the population size of 40. The ANN parameters based on the results of the hybrid LSA‐ANN are  shown in Table 1. The PSO algorithm is also implemented to obtain the same objective for 10, 20, 30, shown in Table 1. The PSO algorithm is also implemented to obtain the same objective for 10, 20, 30,  40 and 50 population sizes for comparison with the results from the hybrid LSA‐ANN, as shown in  40 and 50 population sizes for comparison with the results from the hybrid LSA-ANN, as shown in Figure 23.

Figure 23. 

Table 1. ANN‐designed parameters.  Table 1. ANN-designed parameters. Parameter Value Parameter Value Number of inputs  5  Number of inputs 5 4  Number of outputs  Number of outputs 4 Number of hidden layers  2  Number of hidden layers 2 Number of neurons in hidden layer N1  6  Number of neurons in hidden layer N1 6 Number of neurons in hidden layer N2  Number of neurons in hidden layer N2 4 4  Number of iterations 1000 Number of iterations  1000  Learning rate 0.6175 Learning rate  0.6175 

Type Type ANN inputs  ANN inputs ANN outputs  ANN outputs ANN hidden layer  ANN hidden layer Obtained from LSA  Obtained from LSA Obtained from LSA  Obtained from LSA ANN iterations ANN iterations  Obtained from LSA Obtained from LSA 

-7

Objective function

6

x 10

PSO10 PSO20 PSO30 PSO40 PSO50

5 4 3 2 1 0 0

20

40 60 Iterations

80

100

 

Figure 23. Objective function with iteration of the hybrid particle swarm optimization (PSO)‐ANN  Figure 23. Objective function with iteration of the hybrid particle swarm optimization (PSO)-ANN for for different population sizes.  different population sizes.

The obtained results from the hybrid LSA‐ANN are compared with those from the hybrid PSO‐ The obtained results from the hybrid LSA-ANN are compared with those from the hybrid ANN to validate the developed algorithm. The result of the hybrid LSA‐ANN achieves a MAE error  PSO-ANN to validate the developed algorithm. The result of the hybrid LSA-ANN achieves a MAE −9 after 40 iterations at a population size of 40. The hybrid PSO‐ANN obtains a MAE error  of 9.128 × 10 error of 9.128 × 10−9 after 40 iterations at a population size of 40. The hybrid PSO-ANN obtains a MAE −8 after 81 iterations at a population size of 40, as illustrated in Figure 24.  error of 1.195 × 10−8 after 81 iterations at a population size of 40, as illustrated in Figure 24. of 1.195 × 10 -8

Objective function

5

x 10

PSO-ANN LSA-ANN

4 3 2 1 0 0

20

40 60 Iterations

80

100

 

for different population sizes.  for different population sizes. 

The obtained results from the hybrid LSA‐ANN are compared with those from the hybrid PSO‐ The obtained results from the hybrid LSA‐ANN are compared with those from the hybrid PSO‐ ANN to validate the developed algorithm. The result of the hybrid LSA‐ANN achieves a MAE error  ANN to validate the developed algorithm. The result of the hybrid LSA‐ANN achieves a MAE error  −9 of 9.128 × 10 −9 after 40 iterations at a population size of 40. The hybrid PSO‐ANN obtains a MAE error  of 9.128 × 10  after 40 iterations at a population size of 40. The hybrid PSO‐ANN obtains a MAE error  Energies 2016, 9, 716 17 of 20 −8 of 1.195 × 10 of 1.195 × 10−8 after 81 iterations at a population size of 40, as illustrated in Figure 24.   after 81 iterations at a population size of 40, as illustrated in Figure 24.  -8

Objectivefunction function Objective

x 10-8 55 x 10

PSO-ANN PSO-ANN LSA-ANN LSA-ANN

44 33 22 11 00 00

20 20

40 40Iterations60 60 Iterations

80 80

100 100

  

Figure 24. Performance comparison of LSA and PSO.  Figure 24. Performance comparison of LSA and PSO. Figure 24. Performance comparison of LSA and PSO. 

The regression coefficient (R) close to unity and the value of R for training is 1. The performance  The regression coefficient (R) close to unity and the value of R for training is 1. The performance  The regression coefficient (R) close to unity and the value of R for training is 1. The performance of the hybrid LSA‐ANN is shown in Figure 25.  of the hybrid LSA-ANN is shown in Figure 25. of the hybrid LSA‐ANN is shown in Figure 25. 

Output~= ~=1*Target 1*Target++8.5e-08 8.5e-08 Output

11 0.8 0.8

Training: Training: R=1 R=1 Y Y= =T T Fit Fit Data Data

0.6 0.6 0.4 0.4 0.2 0.2 00 00

0.2 0.2

0.4 0.6 0.4 0.6 Target Target

0.8 0.8

11

  

Figure 25. Performance of the hybrid LSA‐ANN.   Figure 25. Performance of the hybrid LSA‐ANN. Figure 25. Performance of the hybrid LSA-ANN.  

6.4. Results of the Proposed Hybrid LSA-ANN Based Home Energy Management Scheduling Controller A DR event is usually imposed by the power utility to reduce the total power consumption at the peak period time with DL. The DR event is assumed to start from between 4:00 P.M. and 11:00 P.M., and DL is assumed to be 3 kW. If the total electrical power consumption is greater than the DL, then the HEMSC will turn OFF the appliance according to priority, starting with REF, and force the loads to shift and schedule their operating time after the DR event to keep the total power consumption below its DL. The HEMS issues a control signal to turn ON the appliance when the total household load is below its DL level. In this way, the HEMSC will optimize the scheduling of the appliances while maintaining the total power consumption below its DL. Two case studies are considered to describe the implementation of the HEMSC algorithm. The first case does not apply the DR signal and the second case applies the DR signal by using the hybrid LSA-ANN, as shown in Figure 26. The second case using the hybrid PSO-ANN is shown in Figure 27 to clarify the performances of WH, AC, WM and REF, and to calculate the power saving. Figures 26 and 27 show that AC, WM, and REF have to be switched OFF, and one appliance, which is the WH, can be operated and draws 3 kW according to priority. The AC and REF require their schedules to be shifted to another period. The results explain the performance of the proposed HEMSC with the reduction of the total power consumption of the four home appliances at a specific time below the DL value. The algorithm prevents the total power consumption from exceeding the selected DL value. The energy saving for the total power consumption is 9.7138% per 7 h without any effect on the comfort level of the end users, whereas the energy saving for the total power by using the hybrid PSO-ANN is 2.3817% per 7 h. The power saving performance of the proposed hybrid LSA-ANN based HEMSC is better than that of the hybrid PSO-ANN.

Power consumption (kW) Power consumption (kW)

load is below its DL level. In this way, the HEMSC will optimize the scheduling of the appliances  load is below its DL level. In this way, the HEMSC will optimize the scheduling of the appliances  while maintaining the total power consumption below its DL.  while maintaining the total power consumption below its DL.  Two case studies are considered to describe the implementation of the HEMSC algorithm. The  Two case studies are considered to describe the implementation of the HEMSC algorithm. The  first case does not apply the DR signal and the second case applies the DR signal by using the hybrid  first case does not apply the DR signal and the second case applies the DR signal by using the hybrid  LSA‐ANN, as shown in Figure 26. The second case using the hybrid PSO‐ANN is shown in Figure 27  Energies 2016, 9, 716 18 of 20 LSA‐ANN, as shown in Figure 26. The second case using the hybrid PSO‐ANN is shown in Figure 27  to clarify the performances of WH, AC, WM and REF, and to calculate the power saving.  to clarify the performances of WH, AC, WM and REF, and to calculate the power saving.  7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 00 0

Before DR signal Before DR signal

Demand limit Demand limit

Air conditioner Water Air conditioner heater Water heater Refrigerator

After DR signal After DR signal

Washing machine Washing machine

Refrigerator

1 1

2 2

3 3

4 4

5 5

6 6

7 7

8 8

9 9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time12 (hour) 10 11 13 14 15 16 17 18 19 20 21 22 23 24 Time (hour)

Power consumption (kW) Power consumption (kW)

Figure 26. Total power consumption before and after DR signal with the hybrid LSA‐ANN.  Figure 26. Total power consumption before and after DR signal with the hybrid LSA-ANN. Figure 26. Total power consumption before and after DR signal with the hybrid LSA‐ANN.  7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 0 0 0

Before DR signal Before DR signal

After DR signal After DR signal

Demand limit Demand limit Water heater Water heater Refrigerator Refrigerator

1 1

2 2

3 3

4 4

5 5

Washing machine Washing machine

Air conditioner Air conditioner

6 6

7 7

8 8

9 9

10 10

11 12 13 Time12 (hour) 11 13 Time (hour)

14 14

15 15

16 16

17 17

18 18

19 19

20 20

21 21

22 22

23 23

24 24

Figure 27. Total power consumption before and after DR signal with the hybrid PSO‐ANN.  Figure 27. Total power consumption before and after DR signal with the hybrid PSO‐ANN.  Figure 27. Total power consumption before and after DR signal with the hybrid PSO-ANN.

Figures 26 and 27 show that AC, WM, and REF have to be switched OFF, and one appliance,  Figures 26 and 27 show that AC, WM, and REF have to be switched OFF, and one appliance,  7. Conclusions which is the WH, can be operated and draws 3 kW according to priority. The AC and REF require  which is the WH, can be operated and draws 3 kW according to priority. The AC and REF require  their schedules to be shifted to another period. The results explain the performance of the proposed  This paper presents the application of LSA to solve the problems of ANN by finding the optimum their schedules to be shifted to another period. The results explain the performance of the proposed  HEMSC with the reduction of the total power consumption of the four home appliances at a specific  learning rate and the best value of neurons in each hidden layer of the neural network that can be HEMSC with the reduction of the total power consumption of the four home appliances at a specific  time below the DL value. The algorithm prevents the total power consumption from exceeding the  used in HEMSC. A comparison of results shows that the hybrid LSA-ANN used in HEMSC is better time below the DL value. The algorithm prevents the total power consumption from exceeding the  selected DL value. The energy saving for the total power consumption is 9.7138% per 7 h without any  than the hybrid PSO-ANN in terms of scheduling household appliances and reducing the peak load selected DL value. The energy saving for the total power consumption is 9.7138% per 7 h without any  effect on the comfort level of the end users, whereas the energy saving for the total power by using  while guaranteeing end user comfort associated with the operation of loads. The ANN ON/OFF effect on the comfort level of the end users, whereas the energy saving for the total power by using  the hybrid PSO‐ANN is 2.3817% per 7 h. The power saving performance of the proposed hybrid LSA‐ estimation status is enhanced by minimizing the MAE. The hybrid LSA-ANN achieves a MAE error the hybrid PSO‐ANN is 2.3817% per 7 h. The power saving performance of the proposed hybrid LSA‐ ANN based HEMSC is better than that of the hybrid PSO‐ANN.  of 9.128 × 10−9 , whereas the hybrid PSO-ANN achieves a MAE error of 1.195 × 10−8 . The proposed ANN based HEMSC is better than that of the hybrid PSO‐ANN.  algorithm shows a better response in switching the status in HEMSC. Therefore, the energy saving 7. Conclusions  for the total power by using the hybrid LSA-ANN is 9.7138% per 7 h, whereas that by using the 7. Conclusions  This  paper  presents  the  per application  LSA  to  solve the the capability problems of of the ANN  by  finding  the  hybrid PSO-ANN is 2.3817% 7 h. Theof  results explain proposed HEMSC This  paper  presents  the  application  of  LSA  to  solve  the  problems  of  ANN  by  finding  the  optimum learning rate and the best value of neurons in each hidden layer of the neural network that  algorithm to maintain the total electrical energy consumption below the DL value during a DR event. optimum learning rate and the best value of neurons in each hidden layer of the neural network that  Moreover, the algorithm easily deals with the DR signals and is more effective in energy saving. Acknowledgments: The authors gratefully acknowledge the University Kebangsaan Malaysia for the financial support for the project under research Grant No. DIP-2014-028. Author Contributions: Maytham S. Ahmed is a Ph.D. student implementing the project and he is the corresponding author of the manuscript. Azah Mohamed is the main supervisor of the student who leads the project and edits the manuscript. Hussain Shareef is the co-supervisor of the student who has edited the manuscript and given valuable suggestions to improve the manuscript. Raad Z. Homod is the co-supervisor of the student who has edited the manuscript. Conflicts of Interest: The authors declare no conflict of interest.

Energies 2016, 9, 716

19 of 20

References 1. 2. 3.

4. 5. 6. 7. 8.

9. 10.

11.

12.

13. 14.

15. 16. 17. 18.

19.

20. 21.

Deploying a Smarter Grid through Cable Solutions and Services, 2010. Available online: http://www. nexans.com/Corporate/2010/WHITE_PAPER_SMART_GRIDS_2010.pdf (accessed on 9 August 2016). Hammond, G.P.; Pearson, P.J. Challenges of the transition to a low carbon, more electric future: From here to 2050. Energy Policy 2013, 52, 1–9. [CrossRef] Barton, J.; Huang, S.; Infield, D.; Leach, M.; Ogunkunle, D.; Torriti, J.; Thomson, M. The evolution of electricity demand and the role for demand side participation, in buildings and transport. Energy Policy 2013, 52, 85–102. [CrossRef] Huaman, R.N.E.; Tian, X.J. Energy related CO2 emissions and the progress on ccs projects: A review. Renew. Sustain. Energy Rev. 2014, 31, 368–385. [CrossRef] Yun, G.Y.; Kim, H.; Kim, J.T. Effects of occupancy and lighting use patterns on lighting energy consumption. Energy Build. 2012, 46, 152–158. [CrossRef] Richardson, I.; Thomson, M.; Infield, D.; Clifford, C. Domestic electricity use: A high-resolution energy demand model. Energy Build. 2010, 42, 1878–1887. [CrossRef] Horst, G.R.; Zhang, J.; Syvokozov, A.D. Total home Energy Management System. U.S. Patent 7561977 B2, 14 July 2009. Arif, M.T.; Oo, A.M.T.; Stojcevski, A. An investigation for improved home energy management. In Proceedings of the 2014 Australasian Universities Power Engineering Conference (AUPEC), Perth, Australia, 28 September–1 October 2014. Vardakas, J.S.; Zorba, N.; Verikoukis, C.V. A survey on demand response programs in smart grids: Pricing methods and optimization algorithms. IEEE Commun. Surv. Tutor. 2015, 17, 152–178. [CrossRef] Patel, K.; Khosla, A. Home energy management systems in future smart grid networks: A systematic review. In Proceedings of the 2015 1st International Conference on Next Generation Computing Technologies (NGCT), Dehradun, India, 4–5 September 2015; pp. 479–483. Farzamkia, S.; Ranjbar, H.; Hatami, A.; Iman-Eini, H. A novel PSO (particle swarm optimization)-based approach for optimal schedule of refrigerators using experimental models. Energy 2016, 107, 707–715. [CrossRef] Abd Ali, J.; Hannan, M.A.; Mohamed, A. A novel quantum-behaved lightning search algorithm approach to improve the fuzzy logic speed controller for an induction motor drive. Energies 2015, 8, 13112–13136. [CrossRef] Ali, J.A.; Hannan, M.; Mohamed, A.; Abdolrasol, M.G. Fuzzy logic speed controller optimization approach for induction motor drive using backtracking search algorithm. Measurement 2016, 78, 49–62. [CrossRef] Ibrahim, A.A.; Mohamed, A.; Shareef, H. Optimal power quality monitor placement in power systems using an adaptive quantum-inspired binary gravitational search algorithm. Int. J. Electr. Power Energy Syst. 2014, 57, 404–413. [CrossRef] Kang, S.J.; Park, J.; Oh, K.-Y.; Noh, J.G.; Park, H. Scheduling-based real time energy flow control strategy for building energy management system. Energy Build. 2014, 75, 239–248. [CrossRef] Mohsenian-Rad, A.H.; Leon-Garcia, A. Optimal residential load control with price prediction in real-time electricity pricing environments. IEEE Trans. Smart Grid 2010, 1, 120–133. [CrossRef] Haider, H.T.; See, O.H.; Elmenreich, W. Dynamic residential load scheduling based on adaptive consumption level pricing scheme. Electr. Power Syst. Res. 2016, 133, 27–35. [CrossRef] Hernandez, C.A.; Romero, R.; Giral, D. Optimization of the use of residential lighting with neural network. In Proceedings of the 2010 International Conference on Computational Intelligence and Software Engineering (CiSE), Wuhan, China, 10–12 December 2010. Pedrasa, M.A.; Spooner, E.D.; MacGill, I.F. Robust scheduling of residential distributed energy resources using a novel energy service decision-support tool. In Proceedings of the 2011 IEEE PES Innovative Smart Grid Technologies (ISGT), Anaheim, CA, USA, 17–19 January 2011. Setlhaolo, D.; Xia, X.; Zhang, J. Optimal scheduling of household appliances for demand response. Electr. Power Syst. Res. 2014, 116, 24–28. [CrossRef] Wang, Z.; Yang, R.; Wang, L. Multi-agent control system with intelligent optimization for smart and energy-efficient buildings. In Proceedings of the IECON 2010—36th Annual Conference on IEEE Industrial Electronics Society, Glendale, AZ, USA, 7–10 November 2010; pp. 1144–1149.

Energies 2016, 9, 716

22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

20 of 20

Yuce, B.; Rezgui, Y.; Mourshed, M. ANN–GA smart appliance scheduling for optimised energy management in the domestic sector. Energy Build. 2016, 111, 311–325. [CrossRef] Gharghan, S.K.; Nordin, R.; Ismail, M.; Ali, J.A. Accurate wireless sensor localization technique based on hybrid PSO–ANN algorithm for indoor and outdoor track cycling. IEEE Sens. J. 2016, 16, 529–541. [CrossRef] Homod, R.Z.; Sahari, K.S.M.; Almurib, H.A.; Nagi, F.H. RLF and TS fuzzy model identification of indoor thermal comfort based on PMV/PPD. Build. Environ. 2012, 49, 141–153. [CrossRef] Shao, S.; Pipattanasomporn, M.; Rahman, S. Development of physical-based demand response-enabled residential load models. IEEE Trans. Power Syst. 2013, 28, 607–614. [CrossRef] U.S. Department of Energy. Building Energy Codes Progeam. Available online: http://www.Energycodes. Gov/support/shgc_faq.Stm (accessed on 15 May 2016). Kandar, M.; Nikpour, M.; Ghasemi, M.; Fallah, H. Study of the effectiveness of solar heat gain and day light factors on minimizing electricity use in high rise buildings. World Acad. Sci. Eng. Technol. 2011, 73, 73–77. Ahmed, M.S.; Shareef, H.; Mohamed, A.; Ali, J.A.; Mutlag, A.H. Rule base home energy management system considering residential demand response application. Appl. Mech. Mater. 2015, 785, 526–531. [CrossRef] Bertoldi, P.; Hirl, B.; Labanca, N. Energy Efficiency Status Report 2012; Joint Research Centre, European Commission: Ispra, Italy, 2012. Mutlag, A.H.; Mohamed, A.; Shareef, H. A nature-inspired optimization-based optimum fuzzy logic photovoltaic inverter controller utilizing an eZdsp F28335 board. Energies 2016, 9, 120. [CrossRef] Shareef, H.; Ibrahim, A.A.; Mutlag, A.H. Lightning search algorithm. Appl. Soft Comput. 2015, 36, 315–333. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).