Optimization of chiller sequencing control for district cooling system at the Hong Kong-Zhuhai-Macao Bridge Hong Kong Port

Abstract

To mitigate the climate change issue, the Government of the Hong Kong Special Administrative Region (HKSAR) proposed the “Hong Kong Climate Change Plan 2030+” with an ambitious goal of reducing the carbon intensity to 65%–70% below 2005 levels by 2030. To achieve this target, the HKSAR Government has developed two District Cooling Systems (DCS) to promote the use of district energy system, enhancing the overall energy efficiency for cooling purpose. One of the government-owned DCSs is located at the Hong Kong-Zhuhai-Macao Bridge Hong Kong Port (HZMB-HKP), producing a total design cooling capacity of 24.16 MW to passenger clearance building and ancillary buildings there. This paper demonstrates how the DCS at HZMB-HKP could enhance overall energy efficiency for cooling purpose, through proper chiller sequencing control, and reduce CO₂ emission significantly. An optimal seasonal chiller operation mode is also suggested to be adopted at the DCS at HZMB-HKP from cost-effectiveness perspective.

Practical application: District Cooling Systems are considered as a promising solution to alleviate the climate change issue. An optimized chiller sequencing control strategy of DCS can effectively reduce overall cooling energy consumption, so as to achieve carbon neutrality. This paper outlines a case study on a cost-effective operational control strategy of the DCS at HZMB-HKP. The analysis of both the theoretical and practical sequencing control strategy can be taken as a reference for DCS development in the future.

Keywords

district cooling system chiller sequencing energy efficiency retro-commissioning

Introduction

District cooling systems (DCS) present one of the most promising solutions with regard to the climate change mitigation and the development of a carbon-neutralized community. It can achieve a considerable amount of energy saving of up to 35% and 20% when comparing with the traditional Air-cooled Air-Conditioning System (AACS) and Water-cooled Air-Conditioning System (WACS), respectively.¹ The major difference between a typical centralized cooling system and DCS is the concentration effect. Concentration effect means that buildings with similar cooling load profile will be grouped and connected to the same chiller plants. This effect can result in a larger chiller with higher COP performance. Normally, chiller with larger capacity will have a higher COP and when the chillers operate at a part-load ratio with its highest COP, overall energy efficiency will then be improved. Indeed, many regions have already adopted the DCS to maximize the overall energy efficiency. Countries with prominent DCS cooling capacity are indicated in Table 1.^2,3

Table 1.

Various DCSs and their installed cooling capacity.

Region	Installed DCS cooling capacity (MW)
Canada	281.0
France	395.0
Hong Kong, China	308.2
Japan	1032.6
Mainland China	4200.0
United Kingdom (UK)	68.0
United States of America (USA)	1131.0

Considering the rapid growth of energy consumption as well as the drastic increase of the carbon dioxide (CO₂) emission in the territory, the Government of the Hong Kong Special Administrative Region (HKSAR) has proposed the “Hong Kong Climate Change Plan 2030+” to collaborate with other countries globally, in which, DCS development in Hong Kong is also one of the major measures for the reduction of carbon intensity to 65%–70% below 2005 levels by 2030.⁴ The current installed cooling capacity of the DCSs, owned by the HKSAR Government, is about 308.16 MW including 284 MW for Kai Tak Development and 24.16 MW for Hong Kong-Zhuhai-Macao Bridge Hong Kong Port (HZMB-HKP).⁵ To further reduce the CO₂ emission, a comprehensive analysis of energy performance of the DCS at the HZMB-HKP is conducted based on real operational data and seasonal optimized chiller sequencing control pattern is also recommended for further enhancement of the DCS’ overall energy efficiency.

State-of-the-art DCS development in the HKSAR

There are currently two government-owned DCSs implemented in the HKSAR, namely the DCS at Kai Tak Development Area and the DCS at the HZMB-HKP.

District cooling system at Kai Tak development area

The one at Kai Tak Development Area is the first-ever DCS project in the HKSAR, which has been commissioned since January 2013. There are total of three plant houses including North Plant, South Plant, and seawater pump house. The current installed cooling capacity for this system is around 284 MW. Annual estimated electrical energy consumption of air-conditioning at non-domestic floor area in Kai Tak Development Area was around 243 million kWh with adoption of conventional air-cooled air-conditioning system (AACS) and 197 million kWh with adoption of water-cooled air-conditioning system (WACS),¹ while the energy consumption for DCS was only around 158 million kWh. Instead of adopting traditional AACS and WACS, the annual estimated amount of energy reduction for utilizing DCS was about 85 million kWh (35% reduction) and 39 million kWh (20% reduction), respectively, contributing to 59,500 tons and 27,300 tons of CO₂ emission, respectively, using the carbon dioxide emission factor of 0.57kgCO₂/kWh.⁶ It is worth mentioning that an additional DCS is under construction phase, tailor-made to serve the increasing cooling demand of the extension of the Kai Tak Sport Park and the increase in the scale of the new Acute hospital.⁷ Upon full utilization of the additional DCS, it is estimated that an annual amount of 37,000 tons of CO₂ can be further reduced.⁷

District cooling system at Hong Kong-Zhuhai-Macao Bridge Hong Kong Port

The DCS at the HZMB-HKP is the second DCS in the Territory, commissioned in 2018, designed to serve all facilities in the 130-hectare reclaimed island including Passenger Clearance Building (PCB) and ancillary buildings for various authorities. The simplified layout of the DCS at HZMB-HKP is shown in Figure 1. The design cooling capacity of this DCS is around 24.16 MW. The central chiller plant is located at the basement of the PCB, in which the chilled water is distributed through the underground piping network to the heat exchanger at discrete user buildings scattering at the island. In the meantime, seawater is utilized as a cooling medium for cooling the refrigerant, R134a, in the condensers of the chiller. The seawater is being pumped from the West side of the island to the condensers for refrigerant cooling and then being discharged at the North side of the island as indicated in Figure 1. The detailed system configuration, and specifications for the DCS at HZMB-HKP is highlighted in Configuration and specifications of DCS at HZMB-HKP of this paper.

Figure 1.

Simplified layout of district cooling system at HZMB-HKP.

Methodology

All data used for analysis and optimization of the chiller sequencing are the real operational data obtained from the Supervisory Control and Data Acquisition (SCADA) system of the HZMB-HKP for this study. The data are logged at 15-min intervals and 1 year of operational data in the period of 2020–2021 was collected for the analysis. Since the SCADA system can only record yearly operational data, including seawater temperature, only the data starting from February 2020 to February 2021 can be obtained for this study. In view of a lockdown due to outbreak of COVID-19 and only government authorities were operated at the HZMB-HKP. Hence, the cooling load calculated by using the operational data can be considered as a base load, without taken occupant load into account. According to the Hong Kong Observatory (HKO), Spring, Summer, Autumn, and Winter can be divided as the period of March–May, June–August, September–November, and December–February of the next year, respectively.⁸ For the DCS energy analysis, it is generally accepted to consider the centralized water-side equipment including the condensing water pumps, chillers, and chilled water pumps.^{9, 10} The equipment considered for the chiller sequencing analysis in this paper includes the seawater pumps, seawater-cooled chillers, primary chilled water pumps, and the main secondary chilled water pumps.

Based on the analysis of the real operational data obtained from the SCADA system at the HZMB-HKP and seasonal weather in Hong Kong, optimal chiller sequencing pattern is then recommended.

Configuration and specifications of DCS at HZMB-HKP

Configuration of the DCS at HZMB-HKP

Configuration of the DCS at HZMB-HKP can mainly be divided into three loops, namely condensing water loop, primary chilled water loop, and secondary chilled water loop. For the condensing water loop, once-through seawater cooling is adopted, in which, the seawater is being pumped from the West side, entering the condensers, and being discharged at the North side of the island as indicated in Figure 1.

There is no heat exchanger provided in between the condensers and seawater inlets and certain filtration and screening processes are required before entering the condenser. Those processes include arresting small debris contained in seawater through the traveling band screens and filtration through automatic backwash strainers. After minimizing the debris contained in the seawater, the seawater enters the condensers for refrigerant cooling purpose.

For the chilled water loops, a constant-primary variable-secondary pumping system is adopted for chilled water distribution in which, the primary chilled pumps run at constant speed to maintain stable flow through the chillers and the secondary chilled water pumps, delivering the chilled water to the heat exchanger of PCB and ancillary buildings, run at various speed based on cooling load demand.

The schematic diagram of the condensing water loop, primary chilled water loop, and the main secondary chilled water loop is indicated in Figure 2.

Figure 2.

Schematic for condensing water loop, primary chilled water loop and the main secondary chilled water loop for DCS at HZMB-HKP.

Specifications of the equipment of DCS at HZMB-HKP

There are several major items of the DCS at HZMB-HKP, namely traveling screen bands, seawater pumps, automatic backwash strainers, seawater-cooled chillers, primary chilled water pumps, secondary chilled water pumps, and plate heat exchangers for PCB and ancillary buildings. The specifications of the major equipment for the DCS at HZMB-HKP are summarized in Table 2

Table 2.

DCS configuration at HZMB-HKP.

DCS equipment	Capacity and quantity
Traveling band screen	1500 L/s × 2
Seawater pump	476 L/s × 4
Automatic backwash strainer	1500 L/s × 2
Seawater-cooled chiller	1000 RT × 7
Seawater-cooled chiller	435 RT × 3
Primary chilled water pump	210 L/s × 4
Primary chilled water pump	100 L/s × 2
Secondary chilled water pump	266 L/s × 3
Secondary chilled water pump	191 L/s × 2
Plate type heat exchanger	400 kW–2500 kW × 12

Centrifugal pumps for DCS

N = \frac{H*Q*ρ*g}{(1000 *η)}

(1)

There are three types of pumps adopted in the energy performance analysis and system coefficient of performance (SCOP) calculation, namely the seawater pumps (SWP), primary chilled water pumps (PCHWP), and secondary chilled water pumps (SCHWP). Since the SCADA system only takes log of supply water pressure, return water pressure and water flow rate, power consumption of the pumps¹⁰ is evaluated by adopting equation (1), in which N is the electric power consumption in kW, H is the difference between the supply water and return water pressure in m, Q is the water flow rate in m³/s, ρ is the water density in kg/m³, g is the gravitational acceleration in m/s², and η is the efficiency of the pump. The full-load operating efficiency of the seawater pumps, primary chilled water pumps, and secondary chilled water pumps ranges from 80.0% to 85.1%, which were evaluated during testing and commissioning stage of HZMB-HKP.

Seawater-cooled chillers

There are two types of seawater-cooled chillers (SWCCH) adopted for the DCS, that is, 435RT and another with 1000RT cooling capacity. The Coefficient of Performance (COP) characteristic curve for 435RT and 1,000RT SWCCH are shown in Figure 3. All the COP data are evaluated during testing and commissioning stage of HZMB-HKP.

(a) Variation of COP with cooling load of 435RT SWCCH

(b) Variation of COP with cooling load of 1,000RT SWCCH

Figure 3.

Variation of COP with cooling load at different seawater temperature for 435RT SWCCH and (a) Variation of COP with cooling load of 435RT SWCCH (b)Variation of COP with cooling load of 1,000RT SWCCH 1,000RT SWCCH.

Anti-corrosion measures taken for DCS at HZMB-HKP

The DCS at HZMB-HKP utilizes seawater as a cooling medium for refrigerant, in which corrosion is a critical problem that may depreciate the chiller operating performance. Several anti-corrosion measures are implemented on seawater contact parts of the DCS in this case study which are stated as follows:

The condenser tubing of all chillers at HZMB-HKP is made of titanium which has excellent corrosion-resistant property

Zinc metal is installed inside the surface of the water box of condensers of all chillers at HZMB-HKP to further enhance the anti-corrosion performance

Fouling is a common problem in seawater condensing loop. Salt deposition and other impurities might easily build up inside the condenser tubes. From time to time, accumulating scaling and fouling will highly affect the performance of the condensers. An automatic tube cleaning system is integrated to the condensers of all chillers at HZMB-HKP to cater the problem of fouling. The basic principle is to circulate certain amount of soft sponge balls in the condensing water loop through the condenser tubes to minimize the formation of fouling and scaling, slowing down the depreciation of seawater condensers.

Preventive maintenance is also carried out monthly, quarterly and yearly for the condensers and relevant parts to ensure a proper operation of condensing water system. The works include condenser tube and water box cleaning, coating repainting, replacing zinc anodes, and making good joint of condenser covers.

The seawater contact parts of the seawater pumps are made of various types of stainless steel. The volute casing, impeller, shaft is made of duplex steel, and the shaft protecting sleeves are made of cold rolled (CR) steel. Both the duplex steel and CR steel are renowned for corrosion-resistant in the industry.

There is anti-corrosion protection applied to the pipelines and fitting in the seawater piping network at HZMB-HKP. Cement mortar lining is also applied to all the straight pipes, with all the fittings incorporated with fusion epoxy coating, to minimum the possibility of corrosion.

Energy performance indicators for DCS

The following three indicators are generally accepted for assessing general energy performance of the DCS.

Coefficient of Performance of chiller plants

{COP}_{Plant} = \frac{CC}{W_{Plant}}

(2)

W_{Plant} = W_{SWCCH} + W_{PCHWP} + W_{SWP}

(3)

The Coefficient of Performance of chiller plants (COP_Plant) has been widely used for assessing energy performance of the chiller systems and pumps, that is, SWCCH, PCHWP, and SWP. The governing equation is stated in equations (2) and (3), in which, CC is the instantaneous cooling demand in kW, W_Plant, W_SWCCH, W_PCHWP, and W_SWP are the power consumption of DCS plant, seawater-cooled chillers, primary chilled water pumps, and seawater pumps in kW, respectively.⁹ The COP curve of different seawater temperature for 1,000RT SWCCH and 435RT SWCCH are indicated in Figure 3.

Water transfer factor of network distribution

{WTF}_{Distribution} = \frac{CC}{W_{Distibution}}

(4)

The Water Transfer Factor of the cooling network distribution (WTF_Distribution) is adopted for evaluating the energy performance of the cooling distribution network in the chilled water system, in which, CC and W_Distribution represent instantaneous cooling demand in kW and power consumption used for distributing the chilled water (SCHWP) in kW, respectively.⁹ The governing equation is stated in equation (4).

System coefficient of performance

The System Coefficient of Performance (SCOP) is the overall energy performance indicator for DCS, in which the CC_Effective is the instantaneous cooling capacity accounting for the cooling loss of pipework, with governing equations stated in equation (5). The energy consumption of chilled water pumps will eventually be converted into heat gains in the water flow during distribution process and 80% of the pump energy is usually assumed.⁹ By combining the equations (2) to (6), the SCOP for assessing the energy performance of DCS is indicated at equation (7)

SCOP = \frac{{CC}_{Effective}}{W_{Plant} + W_{Distribution}}

(5)

SCOP = \frac{CC - 0.8 W_{Distribution}}{W_{Plant} + W_{Distribution}}

(6)

SCOP = \frac{({WTF}_{Distribution} - 0.8) ({COP}_{Plant})}{{WTF}_{Distribution} + {COP}_{Plant}}

(7)

Results and recommendations

Annual cooling load and distribution of power consumption

CC = m \times c_{p} \times (T_{CHWR} - T_{CHWS})

(8)

The peak cooling load of each month is indicated in Figure 4, calculated by using equation (8), in which, m is the mass flow rate of the chilled water through the SWCCH, c_p is the specific heat capacity of water (4.198 kJ/kg°C), T_CHWR and T_CHWS are chilled water return temperature (°C) and chilled water supply temperature (°C), respectively. Since a control of chiller sequencing is only based on the peak cooling load of each season, only seasonal peak cooling load is indicated in Figure 4. The peak cooling load for Spring (March–May 2020), Summer (June 2020–August 2020), Autumn (September 2020–November 2020), and Winter (December 2020–February 2021) are 7,596 kW, 13,027 kW, 9,488 kW, and 3,056 kW, respectively. The distribution of the power consumption for the annual maximum cooling load extracted from the SCADA system is shown in Figure 5, in which the SWCCH, SWP, PCHWP, and main SCHWP accounts for 62% (2,378 kW), 9% (331 kW), 14% (559 kW), and 15% (594 kW) of total power consumption for the DCS at HZMB-HKP, respectively.

Figure 4.

Annual peak cooling load profile of the DCS at HZMB-HKP.

Figure 5.

Distribution of power consumption at annual peak cooling load of HZMB-HKP.

Theoretical optimal chiller sequencing control at HZMB-HKP

Since the seawater-cooled chillers account for around 62% of the total power consumption of DCS, an optimal chiller staging strategy can effectively enhance the overall energy efficiency for the HZMB-HKP. Chiller sequencing control is then proposed.

Taking into account the average seawater temperature for Spring, Summer, Autumn, and Winter of around 24°C, 29°C, 27°C, and 19°C, respectively, the first option is sole operation of the chiller with 1000RT capacity (hereafter called “baseline”). The second option is to deal with the cooling demand with both 1,000RT and 435RT chillers (hereafter called “S+B”). If the cooling capacity of the 435RT chillers is insufficient, one more 1,000RT chiller will be operated and so on. The third option is almost the same as the second pattern, in which only difference is maximum cooling capacity of each chiller (hereafter called “S+B (MAX COP)”). The maximum cooling capacity of both 435RT and 1,000RT chillers which are set according to the maximum COP, indicated in the COP curve in Figure 3, (i.e., when the maximum cooling capacity with the highest COP of one chiller is reached, one more chiller will then be operated to cater the cooling demand). The 435RT chiller will also be put into operation before the start of the 1,000RT chillers. The minimum output (Turn-down ratio) of all chillers of DCS at HZMB-HKP is 25% of full cooling capacity as provided by the manufacturer. This study of “Baseline,” “S+B,” and “S+B (MAX COP)” chiller sequencing control have taken the turn-down ratio into account. Figure 6 and Figure 7 show the operating pattern for 435RT chiller and 1,000RT chiller, respectively. The starting point of each line represent the minimum cooling output of the operating chiller and the region between dark-blue dotted lines indicate the number of chillers that are operating and share equal amount of cooling load. All the above-mentioned three configurations in this study have taken the minimum turn-down ratio into account, in which, each chiller is operating above the 25% of the full cooling capacity.

Figure 6.

Minimum turn-down ratio and operation pattern of 435RT Chiller for different cooling load distribution of DCS at HZMB-HKP.

Figure 7.

Minimum turn-down ratio and operation pattern of 1000RT Chiller for different cooling load distribution of DCS at HZMB-HKP.

The SCOP of each chiller system combination is evaluated with polynomial regression model, which can be used to indicate the approximate SCOP for each cooling load. The chiller sequencing control for DCS is formulated based on the actual power consumption of each combination of chiller systems for all four seasons.

Chiller sequencing control for spring

The peak cooling demand in Spring was about 7,596 kW. The power consumption and the related SCOP of the chiller operation configurations for Spring are indicated in Figure 8 and Figure 9, respectively. In Figure 8, it shows that the configuration “S+B (Max COP)” performs the best theoretically as it had the least power consumption in Spring, and the highest SCOP of this configuration can be up to 4.2. Hence, the combination of operating both 435RT and 1,000RT chillers is recommended for Spring.

Figure 8.

Power consumption of different chiller operation pattern in Spring

Figure 9.

SCOP of different chiller operation patterns in Spring

Adopting this theoretical chiller sequencing control, the total energy consumption was about 1,505MWh, contributing to a reduction of 8.5% energy consumption when comparing with the “Baseline” configuration for Spring. The related CO₂ reduction was around 80 tons, with the territory-based CO₂ emission factor of 0.57 kgCO₂/kWh.¹¹

Chiller sequencing control for Summer

The peak cooling demand in Summer was about 13,027 kW. The power consumption and the related SCOP of the chiller operation configurations are indicated in Figure 10 and Figure 11, respectively. In Figure 10, for the cooling load ranging from 0 to 5,000 kW, configuration “S+B (Max COP)” and “S+B” combination have a similar power consumption, which was lower than that of the “Baseline” operation mode. However, when the cooling load increased to 5,000 kW or above, the power consumption for the “S+B (MAX COP)” chiller system configuration kept the lowest and the related highest SCOP along that region was about 3.6. It is recommended to adopt the “S+B (MAX COP)” configuration in Summer theoretically. The total energy consumption of adopting this energy-saving sequencing control was about 7,380MWh, accounting for an energy reduction of 7% comparing with the “Baseline” operating mode.

Figure 10.

Power consumption of different chiller operation pattern in Summer

Figure 11.

SCOP of different chiller operation patterns in Summer

Chiller sequencing control for Autumn

The peak cooling load in Autumn was 9,488 kW. The power consumption and the related SCOP curve of the chiller operation configurations are indicated in Figure 12 and Figure 13, respectively. From Figure 12, it can be observed that the “Baseline” chiller configuration consumed the least power for most of the cooling load range in Autumn.

Figure 12.

Power consumption of different chiller operation pattern in Autumn

Figure 13.

SCOP of different chiller operation patterns in Autumn

Although the “S+B” sometimes overlap with the “Baseline” chiller operation mode, Figure 13 shows that the Baseline mode has a better performance regarding the overall system energy performance including the relevant SWP, PCHWP, and SCHWP, having largest SCOP of 4.0. Therefore, it is recommended to adopt the “Baseline” operation mode in Autumn. The total energy consumption for utilizing this strategy was around 4663 MWh with CO₂ emission of about 3264 tons. Comparing with the “S+B” configuration, 7.7% (205 tons) of CO2 emission could be reduced.

Chiller sequencing control for Winter

The peak cooling load for Winter was 3,056 kW. The power consumption and the related SCOP of the chiller operation configuration are indicated in Figure 14 and Figure 15, respectively. In Figure 14, the “S+B (MAX COP)” performed the best, having the least power consumption in almost all cooling load range in Winter. In Figure 15, the highest SCOP of this control strategy was around 3. Hence, the “S+B (MAX COP)” is recommended to be adopted for Winter theoretically. The related energy consumption is about 650 MWh, contributing to a reduction of 49 tons of CO₂ emission. It should be noted that the SCOP of the chiller operation configurations in Winter was the least among Spring, Summer, and Autumn in the year, which may be highly related to low part-load ratio of the chiller operation.

Figure 14.

Power consumption of different chiller operation pattern in Winter

Figure 15.

SCOP of different chiller operation patterns in Winter

Recommendations for an optimal chiller sequencing control based on different design criteria

In Chiller sequencing control for spring to Chiller sequencing control for winter, a seasonal analysis on various chiller sequencing control is conducted. In this section, an annual analysis is conducted on overall energy efficiency and cost-effectiveness, tabulated in the configurations in Table 3. The first case is to adopt “S+B” configuration for every season in a year, while the second case is to apply “S+B (MAX COP)” control for every single season. For the third case and forth case, “Baseline” chiller control is adopted for Autumn as the analysis in Chiller sequencing control for autumn showed that both “S+B” and “S+B (MAX COP)” consume more energy than that of “Baseline.”

Table 3.

Cases considered for energy efficiency and cost-effectiveness analysis.

	Spring	Summer	Autumn	Winter
1	S+B	S+B	S+B	S+B
2	S+B (MAX COP)	S+B (MAX COP)	S+B (MAX COP)	S+B (MAX COP)
3	S+B	S+B	Baseline	S+B
4	S+B (MAX COP)	S+B (MAX COP)	Baseline	S+B (MAX COP)

Optimal chiller sequencing control base on energy efficiency

Figure 16 showed the total energy consumption of the studied chiller sequencing control for each season. As the most energy-efficient operational mode, it is proposed to adopt configuration number 4, that is, “S+B (MAX COP)” in Spring, Summer, and Winter and “Baseline” in Autumn, leading to an annual potential energy saving 785,316kWh.

Figure 16.

Seasonal energy consumption for each chiller sequencing control.

Optimal chiller sequencing control base on cost-effectiveness

In this section, a simple payback calculation for different chiller combinations is performed to figure out the most cost-effective chiller sequencing control. The installation cost of a SCADA system for the DCS at HZMB-HKP is around HKD15 million, while the potential cost for reprogramming and interfacing SCADA system to chillers would be around HKD3 million on top of the original installation cost of SCADA system. The average electricity tariff rate in the studied period was around HKD1.22/kWh.

Table 4 indicates the net annual energy reduction when comparing with the “Baseline” control for the four studied cases in different seasons. The positive sign indicates the energy consumption of that particular configuration is larger than that of “Baseline.” In contrast, the negative sign indicates that particular configuration consumes less energy than that of “Baseline.”

Table 4.

Net energy saving of various configurations comparing with “Baseline” control of chiller in kWh

	Spring	Summer	Autumn	Winter	Annual net energy saving
1	−93,162	−505,824	+358,830	−79,826	−319,982
2	−140,167	−558,740	+225,530	−86,409	−559,786
3	−93,162	−505,824	0	−79,826	−678,812
4	−140,167	−558,740	0	−86,409	−785,316

Positive (+): The configuration consumed less energy than that of “Baseline” Negative (−): The configuration consumed more energy than that of “Baseline”

Table 5 shows the payback period of the four studied cases that considered the installation cost and annual energy cost saving. The analysis showed that configuration 4 may not be the most cost-effective control, even though it is the most energy efficient mode among all the studied chiller sequencing control models, as the payback period is around 21 years. Considering configuration 3, utilizing “S+B” for Spring, Summer, and Winter and applying “Baseline” for Autumn, the payback period is around 18 years, which is less than the other three control strategies. Based on the estimation of payback period, it is proposed to adopt configuration three if the system designers put emphasis on cost-effectiveness. The payback calculation was performed based on data collected during the COVID-19 lockdowns and it is expected that the maximum cooling demand is likely to increase mainly due to a significant increase in the occupant cooling load in the post-COVID-19 period. The energy saving is expected to be increased as all chillers in the HZMB-HKP can be operated at part-load ratio of at least 70%, contributing to a high chiller COP. With capital cost remaining unchanged, an increase in the annual energy cost saving would contribute to a shorter payback period. For future analysis on the SCADA logged data post-COVID-19, it is suggested to the studied cases should be determined seasonally on top of the “Baseline,” “S+B,” and “S+B MAX COP.”

Table 5.

Summary of payback period of four studied configurations

	Capital cost	Annual energy cost saving (HKD)	Payback period (Year)
1	15,000,000	390,378	38
2	20,000,000	682,939	29
3	15,000,000	828,151	18
4	20,000,000	958,086	21

Summary of recommended chiller sequencing configurations

The energy reduction level and recommended chiller sequencing control strategy for each season considering energy efficiency aspect and cost-effectiveness is indicated in Table 6. The energy-efficient and cost-effective approaches for the seasonal chiller operating pattern are formulated in Table 7 and Table 8, respectively.

Table 6.

Percentage energy reduction and recommendations of chiller sequencing control for each season regarding the energy efficiency and cost-effectiveness

	Spring	Summer	Autumn	Winter
S+B	−5.7%	−6.4%	7.7%	−10.8%
S+B (Max COP)	−8.5%	−7.0%	4.8%	−11.7%
Energy-efficient approach	S+B (Max COP)	S+B (Max COP)	Baseline	S+B (Max COP)
Cost-effective approach	S+B	S+B	Baseline	S+B

Remark: Negative (−) = consume less energy comparing with baseline mode Positive (+) = consume more energy comparing with baseline mode

Table 7.

Summary of seasonal optimal energy efficient chiller sequencing control for DCS at HZMB-HKP

	Spring		Summer		Autumn		Winter
% of cooling load	Cooling load (kW)	Chiller sequencing	Cooling load (kW)	Chiller sequencing	Cooling load (kW)	Chiller sequencing	Cooling load (kW)	Chiller sequencing
0.1	760	S	1303	S (0.75)S (0.75)	949	B	306	S
0.2	1519	S (0.6)S (0.6)	2605	S (0.75)S (0.75)B	1898	B	611	S
0.3	2279	S (0.6)S (0.6)B	3908	S (0.75)S (0.75)B	2846	B	917	S
0.4	3038	S (0.6)S (0.6)B	5211	S (0.75)S (0.75)B (0.75)B	3795	BB	1222	S (0.6)S
0.5	3798	S (0.6)S (0.6)B	6514	S (0.75)S (0.75)B (0.75)B	4744	BB	1528	S (0.6)S
0.6	4558	S (0.6)S (0.6)B (0.75)B	7816	S (0.75)S (0.75)B (0.75)B (0.75)B	5693	BB	1834	S (0.6)S
0.7	5317	S (0.6)S (0.6)B (0.75)B	9119	S (0.75)S (0.75)B (0.75)B (0.75)B	6642	BB	2139	S (0.6)S
0.8	6077	S (0.6)S (0.6)B (0.75)B	10,422	S (0.75)S (0.75)B (0.75)B (0.75)B (0.75)	7590	BBB	2445	S (0.6)S (0.6)B
0.9	6836	S (0.6)S (0.6)B (0.75)B	11,724	S (0.75)S (0.75)B (0.75)B (0.75)B (0.75)	8539	BBB	2750	S (0.6)S (0.6)B
1	7596	S (0.6)S (0.6)B (0.75)B (0.75)B	13,027	S (0.75)S (0.75)B (0.75)B (0.75)B (0.75)B	9488	BBB	3056	S (0.6)S (0.6)B

Remark: S = 435RT chiller, B = 1,000RT chiller

Table 8.

Summary of seasonal optimal cost-effective chiller sequencing control for DCS at HZMB-HKP.

	Spring		Summer		Autumn		Winter
% of cooling load	Cooling load (kW)	Chiller sequencing	Cooling load (kW)	Chiller sequencing	Cooling load (kW)	Chiller sequencing	Cooling load (kW)	Chiller sequencing
0.1	760	S	1303	S	949	B	306	S
0.2	1519	S	2605	SS	1898	B	611	S
0.3	2279	SS	3908	SSB	2846	B	917	S
0.4	3038	SSB	5211	SSB	3795	BB	1222	S
0.5	3798	SSB	6514	SSB	4744	BB	1528	S
0.6	4558	SSB	7816	SSBB	5693	BB	1834	SS
0.7	5317	SSB	9119	SSBB	6642	BB	2139	SS
0.8	6077	SSB	10,422	SSBBB	7590	BBB	2445	SS
0.9	6836	SSBB	11,724	SSBBB	8539	BBB	2750	SS
1	7596	SSBB	13,027	SSBBB	9488	BBB	3056	SSB

Remark: S = 435RT chiller, B = 1000RT chiller

In fact, there are certain constraints for this study. The regular scheduled preventive maintenance for the SCADA system may lead to data loss for some period of time. Also, as the SCADA system does not log the instantaneous power consumption data directly, corresponding power consumption for all HVAC pumps and chillers, it can only be calculated based on theoretical equations and assumptions from academic thesis. Hence, it would be better to conduct retro-commissioning individually for each premise to tailor-made the most technically feasible and cost-effective energy saving solution.

Proposed configurations in different dimensions

Optimal chiller sequencing control is one of the most essential components in both design and operation and maintenance stage to minimize the CO₂ emission from reducing energy consumption. It is recommended to apply chiller configuration 4, adopting “S+B (MAX COP)” for Spring, Summer, and Winter and “Baseline” for Autumn to the DCS at HZMB-HKP if the aim of the design is to achieve higher system energy efficiency. With consideration shortest payback period, it is suggested to adopt configuration 3 with payback period of 18 years, utilizing “S+B (MAX COP)” for Spring, Summer, and Winter and “Baseline” for Autumn for the DCS at HZMB-HKP. The payback calculation was performed based on data collected during the COVID-19 lockdown and it is anticipated that the annual energy cost saving will increase in the post-COVID-19 period as the higher PLR of the chillers contribute to a higher COP. The whole study demonstrates a simple method that assist the system designers with the decision making on various chiller sequencing controls with different design criteria.

Practical application on energy assessment of various chiller sequencing design for new DCS development

District energy systems are a popular means of enhancing regional energy efficiency. This study demonstrates a simple approach for evaluating energy performance of various chiller sequencing control options which can assist operators to determine optimal operation patterns of chillers base on different operational criteria. The following items should be considered in this study:

This case study has not considered the factor of depreciation of chillers in the energy performance evaluation and SCOP calculation. It is recommended to use the instantaneous data of power consumption with the polynomial regression model and equation to predict the operating SCOP as the factor of depreciation can be reflected by SCOP from time to time.

As the instantaneous data of power consumption have not been logged by the SCADA system, it is not technically feasible to obtain a practical SCOP curve with different sea water temperatures. It is recommended to group the data points with same condensing temperature and set up a COP curve for that particular temperature in the future DCS development. It is also feasible to formulate an accurate and practical chiller sequencing strategy for different condensing temperatures through a regression curve of the operating condensing temperatures.

The DCS pumping systems at HZMP-HKP consists of a constant-primary pumping system and a variable-secondary pumping system. Primary chilled water pumps are required to maintain a stable chilled water circulation through evaporator to avoid freezing problem, it is considered as the W_Plant in the SCOP calculation. For DCS utilizing only one type of pumping system, such chilled water pumps should only be considered as W_Distribution for the SCOP calculation as its aim is to deliver chilled water to the terminal units.

The seawater cooling system at the condensing water loop is a once-through system. In case energy evaluation is to be conducted on a DCS adopting an indirect seawater cooling system for condensing water loop, the power consumption of the circulating pumps between condensers and seawater heat exchangers should be considered in the SCOP calculation.

Thermally driven cooling system (absorption cooling), utilizing waste heat or heat from renewable energy sources, is also commonly adopted for the DCS. The grades of energy indicate the order state of molecules stored in these forms of energy. The energy sources for an absorption cooling system are considered as low-grade energy, while energy sources for an electrically driven cooling system is a kind of high-grade energy. Because of different energy grades and sources for absorption and electric cooling system, the energy input cannot be directly adopted for SCOP calculation and determining the chiller sequencing control. Hence, the energy evaluation method in this case study cannot be adopted for the DCS utilizing absorption cooling system including combined cooling, heat and power application

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Tin Chung Kerrick Law

References

Audit Commission . The government of Hong Kong special administrative region (HKSAR), Report No. 53: Wider use of water-cooled air-conditioning systems, 2009.

Public Work Sub-committee . The government of the Hong Kong special administrative region (HKSAR), Item for public works subcommittee of finance committee (45CG) – district cooling system at the Kai Tak development, 2009.

Asian Development Bank . District cooling in the people’s republic of China status and development potential, 2017.

Environmental Bureau . The government of the Hong Kong special administrative region. Hong Kong’s Climate Action Plan, 2017.

Energy Efficiency Office, Electrical and Mechanical Services Department, The Government of the Hong Kong Special Administrative Region . Implementation of district cooling system at Kai Tak development, 2014.

Environmental Protection Department . The government of the Hong Kong special administrative region, “Guidelines to account for and report on greenhouse gas emissions and removals for buildings (Commercial, Residential or Institutional Purposes) in Hong Kong, 2010.

Legislative Council . The government of the Hong Kong special administrative region background brief on development of district cooling systems prepared by the Legislative council secretariat, 2020.

Kong Hong Observatory . The government of the Hong Kong special administrative region, definition of seasons. 2020, [Online]. Available: https://www.hko.gov.hk/en/education/climate/general-climatology/00545-definition-of-seasons.html (Accessed Mar. 19 2021).

Yan

Gang

Niu

, et al. Quantitative evaluation of the impact of building load characteristics on energy performance of district cooling systems. Appl Energy 2017; 205: 635–643. DOI: 10.1016/j.apenergy.2017.08.022.

10.

Gang

Wang

Gao

, et al. Performance assessment of district cooling systems for a new development district at planning stage. Appl Energy 2015; 140: 33–43. DOI: 10.1016/j.apenergy.2014.11.014.

11.

China Light & Power (CLP) Group . Sustainability Report, 2020.