ENERGY MANAGEMENT OF A CHILLED WATER PLANT USING THERMAL STORAGE

S. Bilodeau, PE, Ph.D. and J. Gagne, PE, M.Appl.Sc. Groupe Enerstat Inc 125, Turgeon, Bromptonville (Quebec) J0B 1H0 CANADA Tel: 1-819-846-1040 [email protected]

1.

INTRODUCTION

This paper illustrates a project implemented in a major IBM plant in Canada using a TES system integrated with other technologies to optimize the chilled water production. The implementation of the Novanergy system (thermal storage with phase change material) has given important peak load reductions as well as significant reduction in the energy consumption and in the environmental impact of the chilled water system operation. Natural Resource Canada and Hydro-Quebec have contributed in the project carried on by Groupe Enerstat Inc. A load-leveling approach has been exploited to minimize the required equipment and storage capacities (reducing the amount of refrigerant needs by about the equivalent of a 1000 tons unit) as well as GHG emissions for a given load because of energy efficiency improvement. The objective of the project was to demonstrate on-site the optimization of a major plant chilled water system using thermal storage. The IBM Bromont plant had been chosen as the implementation site because of the significant potential to reduce energy consumption and environmental impact of the operation with the technology. The Central Utility Plant at IBM Bromont has a capacity close to 10,000 tons to produce cooling water between 42 – 48 deg. F. In 2005, IBM Bromont needed to change its old Freon 12 chillers.

Figure 1: Thermal Storage Units The Novanergy thermal storage integrates a proprietary Phase Change Material (PCM) which the patent belongs to Groupe Enerstat Inc. The system includes a 1,400 tons VFD chiller to charge the thermal storage tanks, a 4,000

USGPM VFD pump, a 2,500 tons plate heat exchanger, a glycol loop and finally two 2,000 tons-h storage bank (for a total of 4,000 tons-h). The energy savings, heat flow into and out of the storage tanks, as well as the electric consumption are all being monitored. The work reported here is on the energy savings and subsequent financial benefits for the first year of operation while the monitoring is still undergoing for another 3 years.

Figure 2: Chiller Installation

2.

OVERVIEW OF THE IMPLEMENTATION

The Project at the IBM Bromont Central Utility Plant has shown that thermal storage may be used not only to reduce peak loads (kW), but also to reduce energy consumption (kWh). The coupling of a variable frequency drive compressor to the thermal storage tanks and a free cooling heat exchanger allows for an optimization of power and energy consumption. The thermal storage using PCM is used with an advanced control strategy to reduce the power consumption. The equipment typically runs at its full capacity for about 24 hours per day. When the load is less than the equipment output, the surplus energy is stored. When the load exceeds the equipment output, the additional requirement is discharged from storage. The size of the storage tanks (two 45 feet units) actually allows for a total storage capacity of more than 7,000 kWh (almost 50,000,000 BTU) per charge. A 4,000 USGPM glycol loop serves as the thermofluid which takes the “Efficient” kWh (BTU) from the storage tank to fulfill the load. The chiller is no longer following directly the load and operates in a stable pattern closer to its most efficient operating conditions; it is now charging the storage tank. The phase change process is sustainable and stable (no moving parts, no biological concerns, stable temperature, etc.) An innovative control strategy involving the real-time computation of the thermodynamic balance and a predictive model of the coming peak loads have been developed specifically for the project. The storage process is driven by different means: • Novanergy charging system NV1: High efficiency Compressor system The chiller (NV1) charge the low temperature storage tank (MCP2) and the moderate temperature tank (MCP1) at the same time using “unvalorized” energy rejection from the central energy plant • Free cooling: night free cooling is “moved” to follow day loads

NFC: by charging the system directly with a cooling tower at night – during mid-season – QFC: by precooling the NV1 evaporator to improve overall efficiency [Quasi-Free Cooling approach]

MCP2

MCP1 Heat Exchanger

Disharge

V1

Discharge

Charge &

V2

Cooling Tower or Heat Recovery

Figure 3: Operating Schematic The implementation is kept simple: only one thermofluid is required to charge and discharge the accumulator; the thermofluid takes the “Efficient BTU” from the tank, thanks to a circulating pump and modulating valve introduced in the main cold water loop of the energy plant. We have now a system with an instantaneous capacity of 2,500 tons that can modulate as low as 100 tons.

3. OVERVIEW OF MONITORING AND EXPERIMENTAL METHODOLOGY The project was to be carried out through 3 phases and 12 tasks realized from mid-2004 to Spring 2006 Phase I – Development and optimization of the implementation, build and install Task 1 : Final Design of the system. The scale-up of the unit was done according to the needs of the site. Task 2 : Manufacturing of the system. The unit was built according to the specifications. Task 3 : Installation. The unit was installed on site. Task 4 : Unit start-up and debugging. Phase II – Commissionning and Monitoring set-up Task 5 : Development of the test protocol. Task 6 : Experimental set-up (sensors and meters installed) in order to realize the test protocol. Task 7 : Implementation and test of instruments (Preliminary data acquisition). Task 8 : Commissionning (to optimize the integration in the chilled water plant) according to test protocol established in task 5. Task 9 : Preliminary analysis and report Phase III - Monitoring Task 10 : Long term testing was undertaken. At this stage, the system was working at the site under usual operating conditions and start-ups and shutdowns are made by the IBM site staff. The system transmits data periodically so that a complete working cycle can be analyzed. Task 11 : Data analysis in order to identify the advantages of the unit in terms of thermodynamic efficiency, environmental performances, economics, etc. Task 12 : Final report including all results, schematics, heat/mass balances and conclusions.

The experimental set-up was established to allow for redundancy and accuracy. Continuous measurement (sampling rate is 10 minutes) of the energy balance is done through more than 30 different variables ranging from temperature in 14 locations to the electrical consumption of the chiller and pumps, as well as the flowrates in the different loops and the position of the 3-way valves (V1 and V2). The data acquisition was including the outdoor conditions (dry-bulb temperature, wet bulb temperature) as well as some production data (plant electrical and chilled water consumptions) to allow for good comparison of the results. To evaluate the electrical savings, a model estimate of the original system use based on the analysis the 2003-2004 years is compared with the actual use of the TES system on a monthly basis. The reduced data set was assumed to be representative of the entire month, so the daily average was used to estimate a month’s savings. To properly credit the use of the chilled water production energy, all of the existing chillers were monitored before and after the implementation. Before starting the continuous monitoring, extensive experimental validations have been performed. The operating parameters have been measured in the first few weeks of operation. The results are shown in the following table. Table 1: Operating parameters

Parameters

Before (2004)

After implementation (2005)

Nominal Storage Capacities

(no storage)

MCP1 : 28,179 MJ (2,226 tons-h) MCP2 : 25,265 MJ (1,995 tons-h)

Nominal Chiller Capacities

2 x 1000 tons

4,384 kW (1,247 tons)

Reduction in energy consumption ---

5.3 kWh/day/tons (45.9%)

The typical daily evolution of the chilled water production including the thermal storage contribution shows that during peak loads, the complete system was operating at full delivery loads – around 2,550 tons - while the electric consumption was kept as low 510 kW. Those kinds of numbers correspond to an instantaneous coefficient of performance of 0.23 kW/ton during the peak shaving process. The results have shown two general operating patterns. First, the peak shaving contribution of the thermal storage tanks, around 1,200 tons. Second, the instantaneous dampening effect of the MCP1 tank. This last pattern stabilizes the energy supplied by the chiller to follow a fluctuating load and helps keep its efficiency at an optimum level.

800000

2400

700000

2100

600000

1800

500000

1500

400000

1200

300000

900

200000

600

100000

300

0

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Reduction in consumption

Free Cooling Hrs

Figure 2: Monthly Energy Savings and Free Cooling Hours

Free Cooling (number of hrs)

Consumption (kWh)

Reduction in energy consumption

The on-site results have shown that the energy efficiency improvement is significant for the production of chilled water. Globally, the energy savings are fluctuating month by month from 300,000 kWh per month to 600,000 kWh. The best results are obtained in mid-season (fall and spring), because of the conjugated contribution of the Free cooling and higher Peak loads (resulting in higher Peak Shaving opportunities). The following chart illustrates the yearly evolution of the reduction in energy consumption. The following table presents a synthesis of the general results of the monitoring. Table 2: Monitoring results

General results

Energy production and storage for the chilled water loop Before (2004)

After (2005-2006)

Chiller water Production

18,728 tons-h/jr

37,537 tons-h/jr

Daily Consumption

16,706 kWh/jr

15,572 kWh/jr

Average Instantaneous Consumption Average Production (MCP + VFD Chiller) Average Production (before)

696 kW

648.8 kW

--

1564.1 tons

780.3 tons

--

Free Cooling

945 tons (average)

Total Chilled Water Production

750 tons (for 90% of time) 1455 tons

Efficiency

0.892 kW/tons

0.415 kW/tons

Efficiency (including Free Cooling)

0.478 kW/tons

0.259 kW/tons

2509 tons

Different quick observations have been made from the results. First, the reduction in peak load is more than 1.0 MW representing the elimination of the equivalent of a 1000 tons chiller. Second, the general performance before the implementation was 0.89 kW/ton. Presently, the efficiency of the chilled water production is actually less than 0.42 kW/ton, which represents a reduction of more than 0.5 kW/ton. The annual energy savings are evaluated to 5,311,950 kWh per year. 4.

DISCUSSION OF RESULTS

A detailed analysis was performed on the system as operated with TES compared with the previous twelve month period. It was assumed that the first period represented a control similar to the original system, while during the second period, the thermal management of the chilled water production was carried on using the advance controls system and the TES. The improvement in the efficiency is shown in the next Figure. It was found that the best savings are produced during the mid-season, where the storage allows for using the night free cooling all day long. Interesting results, but lower saving, were achieved in the summer and in the winter compared with the spring and fall periods. The remaining savings are those due to the operational efficiency of the integrated system.

Energy consumption for Chilled Water 1500000

kWh

1000000

500000

Mois

Ju ly Au gu Se st pt em be r O ct o No ber ve m De be r ce m be r

ay Ju ne

M

Ap ril

Ja nu a Fe ry br ua ry M ar ch

0

Actual Chiller Consump. kWh

Global Plant Consump. kWh

Figure 4: Monthly Energy Consumption Before and After Implementation Those results outlined that to reduce energy consumption and to recharge the “thermal battery”, different items have been instrumental: High efficiency compressor system (VFD chiller) • Reduction of part loads (minimizing part loads to maintain performance between 0.4 to 0.6 kW/ton) • Reduction in condensing temperature (the operation of the compressor is done favorably at night, at lower outdoor temperatures) with average daily temperature fluctuation of about 20 °F, it represents an improvement of up to 15 % in kW/ton; Free cooling heat exchanger. • An existing free cooling heat exchanger (FCHE) was already in use and had a capacity of about 800 tons when outdoor conditions were favorable. Now the project permits the operation of this FCHE even when outdoor temperature normally commands stopping the operation. The outlet of the FCHE is now directed to the thermal storage system. In fact, this added more than 3,000 hours per year (from September to May) with a “Quasi Free Cooling” approach. Even if the outdoor temperatures were above average during the months of January and February this winter, the “Quasi free cooling” was used, allowing IBM to recuperate more than 950 tons of free cooling even during warmer mid-season days. The total electric savings were 5,312 MWh. The dollar value of the savings at a marginal cost of $0.0253 was $134,400. Additional savings from peak load shaving (kW) was $162,400 for a grand total of $296,800. Assuming an annual dollar savings of $300,000 (the electricity marginal cost has increased by almost 10% since 2004), the present value of the cost avoidance with a twenty year probable usefulness life and a real rate of interest of four percent is equal to $3.9 million dollars. This analysis does not include an estimate of expected reduced cost in maintenance and equipment replacement which was not measured but could well increase savings substantially. Essentially, on the operator side, the system has proven to be simpler to operate and maintain compared to the standard chiller approach. The storage material is confined in a sealed enclosure; the TES tank requires no maintenance and operates on a stable basis. Also, the operator does not need to juggle anymore with the decision to start a 1000 ton chiller when demand is only a few hundred tons or when the pressure is too low. Low demand is met with the thermal storage and low pressure with a new variable frequency drive pump. The reliability of the operation has also been demonstrated when, at different times, fluctuating electric supply had caused chiller shutdown. Every time, the storage unit has taken the load in a few minutes and maintained the chilled water delivery to the plant. The reduction in production losses due to eliminated fluctuations (in supply temperatures) is not considered here. But this safety feature is an important benefit of the approach.

5.

CONCLUSION

This load-leveling approach minimizes the required equipment and storage capacities (reducing the amount of refrigerant needs by about the equivalent of a 1000 tons unit) as well as GHG emissions for a given load because of energy efficiency improvement. A reduction of about 35% in GHG emission for the production of chilled water has been observed at the IBM plant in Bromont. The outcome of the project was also to demonstrate the suitability of such a system in the industrial production of chilled water. A description of the benefits that have resulted from the project includes energy efficiency, environmental impact, costs and paybacks. The results confirm that the developed Novanergy can realize adequately the thermal management of efficient chilled water production. The performance analysis and on-site testing have highlighted many benefits of using the storage for thermal management. Benefits of the system include: • Dramatically reduced energy consumption (about 5,300 MWh per year) • Reduced maintenance and improved reliability • Increased cooling output in severe environments (fluctuating demands) • The system’s performance is virtually independent of weather conditions. The coupling of the regenerator (chiller) and thermal storage stabilizes the operating temperatures. • The Novanergy system is flexible enough to be readily adaptable to the needs of different categories of similar processes.

ACKNOWLEDGMENTS We wish to express our deep appreciation to IBM Bromont for the continuous involvement of its staff and to Natural Resources Canada (NRCan) for funding this project. Our special thanks go to MM. P. Bisset and Mr. D. Pare from IBM, as well as Mr. J. Guerette and his staff from NRCan, for all their contribution.

REFERENCES Bilodeau, S., “An Innovative Approach to Thermal Energy Storage”, Moving forward on Climate Change, United Nations Climate Change Conference, Montreal (Quebec), December 2005. Hewitt, G.F., G.L., Shires and T.R. Bott, Process Heat Transfer, New-York : CRC Press. Chapter 2, 1994. Bilodeau, S., “Thermal Storage to Optimize Passenger Shuttle Cooling systems Operating in Severe Conditions”, SAE International, SAE Commercial Vehicule Engineering Congress, Chicago (Illinois), November 2005. Bilodeau, S., Mercadier Y. and Brousseau P., “Numerical and experimental investigation on frosting of energyrecovery ventilator”, 6th International Symposium on Thermal Engineering and Sciences for Cold Regions, ISTESCR, Darmstadt (Germany), August 1999. NASA Goddard Space Flight Center, International Solar Terrestrial Physics (ISTP), NASA GSFC Technical Report Server, 1997. Bilodeau, S., “Integrated Mechanical System Approach to optimize a hybrid HVAC system”, Hybrid Ventilation 2002, The International Energy Agency, IEA Annex 35, Montreal (Quebec), May 2002