May 24, 2005

A40-D1-M8-D-EB2 Document identification

Type of documents:

Glossary

Type of building:

1

Cx process 2

Type of commissioning: Initial-Cx Glossary check:

yes

MQC task: Other indications:

Program

Project

3 re-Cx

4 retro-Cx

Tool

Model 5

Continuous-Cx

no Design

Elaboration

Realisation

Operation

IEA Annex 40 "Commissioning of Building HVAC Systems for Improved Energy Performance"

Commissioning of Heat Pump Systems Oliver Baumann1, Haruyoshi Kibe2 1 Division for Building Design and Simulation, Ebert-Ingenieure München, Germany 2 Kajima Corporation, Tokyo, Japan; International Trainee at Ebert-Ingenieure München, Germany ABSTRACT / INTRODUCTION This paper shows a model-based commissioning process for heat pump systems by describing a series of steps like model configuration, calibration, application into the real building, verification of the model. The used model is based on MATLAB/Simulink, but can also be programmed in other languages. The performance of the Functional Performance Test is not included in this paper yet. GENERAL DESCRIPTION 1.1 Heat pump components A heat pump (HP) consists of different components like compressors, condensers, evaporators and refrigerant circuits with expansion valves. Low pressure liquid refrigerant enters the evaporator and is evaporated and superheated by the heat energy absorbed from chilled water (heat source) passing through the evaporator shell. Low pressure vapor enters the compressor where pressure and superheat are increased. Heat is given to the hot water (heat sink) by the water cooled condenser. Figure 1.1: Heat pump circuit. The fully condensed and sub-cooled liquid refrigerant then enters the expansion valve where pressure reduction and further cooling takes place before returning to the evaporator.

-1-

2. EXAMPLE GEBHARD-MUELLER-SCHOOL IN BIBERACH, GERMANY 2.1 Central heating system The model of the heat pump was applied for the commissioning of a school building in Biberach, Germany. The central heating system consists of two heat pumps (HPs) and a woodfired furnace. The HPs are ground water-cooled and supply hot water of 40°C (maximum) for heating. The temperatures shown in the scheme indicate design point temperatures. At these conditions, the COP can reach a maximum value at about 5 to 6 with the ground water temperature of around 10°C.

Figure 2.1: Heating system with heat pumps, boiler, pumps, and heat exchangers. Possible energy saving potential for the operation of the heat pump system, including the ground water pump and adjacent pumps and heat exchangers would be Demand control for the variable speed ground water pump to reduce the electrical power consumption. The plate-type heat exchanger is placed on the rooftop at 20m height from the level where the ground pump is placed. Maintain a maximum temperature difference between supply and return water in each loop to reduce energy consumption of the (variable speed) pumps. The maximum allowed temperature difference of 4 K between supply and return ground water must be observed to prevent ground water pollution. Optimal temperature settings for “total” energy saving of the central heating system. 2.2 Automation Control 2.2.1 Operation strategy in design-phase The HP’s are controlled by the heat demand signal from BEMS. The HP controller decides the appropriate stage (the number of compressors) and controls the speed of pumps, Pc1 and Pc2 in the cold loop, Ph1 and Ph2 in hot water loop, corresponding to the heat demand. The ground water pump PG1 is controlled by the supply water temperature Tcs, in regard of its setting point 8°C. 2.2.2 “3 °C” Problem An error in the control strategy was detected during the initial commissioning. The return water temperature Tcr in the cold loop decreased below 3°C, which is the minimum limit to prevent the HP’s from freezing, before the HP’s were forced to stop by the internal control. The reason was that the ground water flow rate was not able to follow the heat demand from the HP’s at the instant of the HP’s hopped up to the next stage because the pump speed was controlled by Tcs with its setting point at 8 °C. For example, in case that the HP’s stage -2-

hopped up from 25% to 50%, the HP’s needed twice heat from ground water, but the ground water pump kept on running at the 25% level because a signal from a sensor of Tcr still measured a return water temperature of 8 °C. This situation continued until Tcs decreased below 8 °C after the water on the point of Tcs sensor went around the loop and reached the point again. But actually Tcr was below 3 °C before the water went around the circuit and the HP’s were forced to stop before Tcs reached below 8 °C. This scenario was simulated with the model. The results are shown in following figures:

Figure 2.2: Ground water mass flow rate in kg/sec (Cyan: Ground water, Yellow: Chilled water)

Figure 2.3: Ground water temperature in °C (Yellow: Ground water, Cyan: Outgoing Ground water, Magenta: Supply Chilled water, Red: Return Chilled water) To solve the problem, the control of the ground water pump was changed in a manner that the speed of the pump is first controlled by the signal that comes directly from the HP controller corresponding to the heat demand. After 6 minutes (time constant of secondary loop, determined by the simulation model) the control of the flow rate switches back to the temperature control (Tcr limited to a minimum of 3 °C). 3. MODELING 3.1 Configuration of HP component model A heat pump component model is already prepared as a block set in SIMULINK. The model can be defined and calibrated by setting several parameters. The parameters are calculated and set with manufacture’s performance data. 3.1.1 Schwamberger’s dynamic model Schwamberger shows an empirical dynamic model of a heat pump, which is based on the static performance diagrams regarding DIN 8900. The capacity of the heat pump is given as a function of the heat source temperature and the hot water temperature from the heat pump as follows.

-3-

Q& wp = K 1 ⋅ϑ A + K 2 ⋅ϑ V W p + K 3 Qwp Capacity of heat pump A Outdoor temperature VWp Outlet hot water temperature K1 Coefficient for primary inlet temperature K2 Coefficient secondary outlet temperature K3 Temperature independent heating power K1, K2 and K3 are given as the constant numbers when “e²” in the following equation is the minimum value in order to approximate the capacity. with

n

(

(i ) e 2 = ∑ Q& wp − K 1 ⋅ϑ A (i ) − K 2 ⋅ϑ V W p(i ) − K 3 i =1

)

2

The capacity is defined as the following equation with the inlet hot water temperature.

(

Q& wp = c F ⋅ m& F ϑV W p − ϑ R W p

)

cf Heat capacity mf Mass flow RWp Inlet hot water temperature An approximate value of the outlet hot water temperature at the operation point is drawn by transforming the equations.

with

∆ϑ

V Wp =

K1 c F ⋅ m& F ⋅ ∆ϑ R W p + ⋅ ∆ϑ A c F ⋅ m& F − K 2 c F ⋅ m& F − K 2

Delta (∆)means the change between certain operation point and approximate value. Figure 3.1 shows block diagram of heat pump model. A two minute time constant is recommended for the transport delay. ∆ϑ

A

K1 c F ⋅ m& F − K 2

1 1 + TA s

∆ϑ

V Wp

1 c F ⋅ m& F e −Tt s 1 + TV s c F ⋅ m& F − K 2 Figure 3.1: Block diagram of heat pump model ∆ϑ

RWp

3.1.2 Calibration of the model The next step is to calibrate the model with manufacture’s data. The performance of the heat pump model is characterized by K1, K2 and K3 of the Schwamberger model. These parameters are determined by using built-in functions of the CARNOT block set (see Figure 3.2).

-4-

Figure 3.2: Parameter setting mask. 3.2 Central heating system model, inclusive of HP’s, pumps and hydraulic networks 3.2.1 Central heating system model There is sometimes a lack of component models in SIMLINK when models are applied into the real building. In this project each of two HPs has two compressors but the component model in SIMLINK has only one stage. Therefore, we modelled ‘4-stage system’ with four single heat pump models as shown below. Storage tanks are omitted to simplify the model. Open loop

Closed loop Pb1 65°C Boiler 85°C Tcs=8°C

Tgs=10°C

GHX

HX Pc1

T

(67.5kW)

Thr=30°

30°C

HP1 Tgr=6°

T Tcr=4°C

Ths=40°C Pc2

Heating Load

Ph1

HP2

to Return well Pc3

Ph2 (67.5kW) HP3

PG1

Supply well

(67.5kW)

40°C

Pc4

(67.5kW)

Ph3

HP4 Ph4

Figure 3.3 Central heating system model with 4 single heat pumps. 3.2.2 Stage control strategy The following figures show the results of the 4-stage control of the heat pumps and pumps corresponding to the heat demand. The capacity of the heat pumps is designed to satisfy 80% of the peak load for each heating devices.

-5-

Figure 3.4: Heat pump capacity in kW (Yellow: heat pump capacity, Magenta: heat demand).

Figure 3.5: Water mass flow rate in kg/sec Ground water mass flow rate in kg/sec (Magenta: Hot water ,Yellow: Chilled water) (Cyan: Ground water). 3.2.3 Configuration of the 4 stage control model Figure 3.6 shows the overview of the control signal flow and water flow. The hot water loop is not closed in the model.

Figure 3.6: Signal flow and water flow in the heating system.

-6-

Figure 3.7: SIMULINK model of heating system with 4 single heat pumps. 4. OPTIMIZATION STUDY The following shows the optimisation of the heat pumps’ dynamic performance using the SIMULINK model. Verification of the model and re-calibration with measured data are not described explicitly in this study. We studied two methods to enhance the COP of the heat pumps as shown in figure 4.1 (bold paths). Both of them cause a trade-off relationship between higher COP of the heat pumps and additional power consumption for the pumps (primarily the ground water pump). The other three methods are not practical because it’s neither possible to raise the ground water temperature nor the intention to change the heat exchanger’s capacity or coils. The increased water flow rate can be realized without new components, just by adjusting the existing balancing valves. Further optimisation methods including the exact schedules for the heat demand from AHU’s, heating systems, etc. are not considered in this study, since the used model does not include these systems.

Figure 4.1: Optimization strategies to enhance the COP of heat pump systems. The following chapters show some particular results of the studies. -7-

4.1 Relation between COP of HPs and heat source water flow rate 4.1.1 Conditions Heat load profile 0 to 300 kW (see Figure 2.4) Simulation time 10 hours HP capacity 270 kW with 4 stages Heat exchanger capacity 270 kW Ground water temperature 10 °C Hot water flow rate max. 20.4 kg/sec Supply hot water temperature 40 °C (set point) 4.1.2 Parameters and Results

Case 1 Case 2 Case 3

chilled water flow rate

ground water flow rate

100% = 16.0 kg/sec 110% = 17.6 kg/sec 120% = 19.3 kg/sec

100% = 8.34°C 16.1 kg/sec 110% = 8.48°C 17.7 kg/sec 120% = 8.60°C 19.3 kg/sec

mean supply chilled water temperature

mean COP of HP’s

HPs power consumption

4.16

445.8 kWh (100%) 446.7 kWh (100.2%) 447.6 kWh (100.4%)

4.17 4.18

heat generation

real HPs power consumption for 100% heat generation 1,861 kWh 445.8 kWh (100%) (100%) 1,860 kWh 444.7 kWh (100.4%) (99.8%) 1,877 kWh 443.9 kWh (100.8%) (99.6%)

4.1.3 Conclusion This study shows, that the mass flow rate of the chilled water in the secondary loop and the ground water has no distinctive influence on the dynamic operation and thus the COP of the heat pump system (less than 1 %). Since the energy consumption of the ground water and circulating pumps would increase with higher flow rates, the COP of the entire system would decrease dramatically. The optimal control strategy therefore is to reduce the ground water and chilled water flow rate to a minimum that prevents under-cooling of the secondary loop temperature and maintains a maximum temperature difference for the supply and return ground water of 4 K. The COP of the heat pump system of 4.2 shows an efficient operation of the system at the given conditions. Anyway, the targeted COP of 5 or even 6 can not be reached with a hot water temperature of 40 °C, but with a hot water temperature of 28 °C what is the maximum supply temperature for the low temperature heating system. 4.2 Relation between COP of HPs and hot water flow rate 4.2.1 Conditions Heat load profile 0 to 300 kW (see Figure 2.4) Simulation time 10 hours HP capacity 270 kW with 4 stages Heat exchanger capacity 270 kW Ground water temperature 10 °C Chilled water flow rate max 16.04 kg/sec Ground water flow rate max 16.05 kg/sec Supply hot water temperature 40 °C (set point)

-8-

4.2.2 Parameters and Results Hot water flow rate

Mean Supply Mean COP HPs power hot water of HPs consumption temperature 40.93°C 4.16 445.77 kWh (100%)

Case 1

100% = 20.40 kg/sec

Case 2

110% = 22.44 kg/sec

40.84°C

4.17

445.47 kWh (99.93%)

Case 3

120% = 24.48 kg/sec

40.78°C

4.17

445.22 kWh (99.87%)

Heat generation 1,861.28 kWh (100%) 1,862.42 kWh (100.06%) 1,863.39 kWh (100.11%)

Real HPs power consumption for 100% heat generation 445.77 kWh (100%) 445.20 kWh (99.87%) 444.70 kWh (99.76%)

4.2.3 Summary The mass flow rate in the hot water loop has no distinctive influence on the COP of the heat pump system, regarding this study. A higher flow rate would improve the heat transport within the load utilities (heating coils, low temperature heating system), but also cause higher return temperatures and higher energy consumption for the circulation pumps. Therefore, it is not recommended to increase the mass flow rate in the hot water loop of the heating system. 5. CONCLUSION Dynamic models can be applied profitable within the initial commissioning and/or optimization phase of the described heating system. Most of the shown measures are evidently in their effects; at least when seen separately. But the effective value of measures with opposite impacts on the energy efficiency need to be evaluated exactly with numeric methods. A further advantage is that the model can be applied already within the design phase of systems to improve components and an appropriate control strategy in a combined procedure.

-9-