The air conditioning loads of a typical underground railway station comprise the following components:
- Solar loads negligible
- Transmission loads 20 percent
- Electrical loads 20 percent
- People and fresh air loads 60 percent.
Usually the air conditioning equipment is sized to meet the peak-hour cooling load at the optimum efficiency. A typical station cooling load profile with peak-hour cooling capacity of 1,000TR is shown in Figure 1. At part load operation, the air conditioning equipment will not be operated at its full capacity and, hence, the efficiency of air conditioning equipment may not be at its optimum. Variable speed drives (VSDs) can save energy by tuning down the system characteristics to meet loading requirements whilst keeping the optimum equipment efficiency. Equipment that is commonly operated by VSDs includes fans and pumps.

Figure 1: Cooling Load Profile and Non-Storage System |

Figure 2: Typical Fan Curve |
Fundamentals of Variable Speed Drives
To determine how VSDs will save energy, we must revisit the fundamentals of fan curves, fan laws and motor selection.
Application of VSD to Fan System. A typical fan curve is shown in Figure 2. The system resistance is proportional to the square of the flow. The duty point is the intersection of the fan curve and the system resistance curve. Usually, optimum fan efficiency is selected at the duty point and it is not uncommon that 70 percent to 85 percent fan efficiency could be achieved.
Motor Selection. In sizing motors, a 15 percent service factor or margin is adopted in the industry. Firstly, it provides a margin to increase the fan flow rate without changing the switch-gear. Secondly, it caters for the ageing of the motor and hence degrade in performance in service.
The motor electrical input power is determined by dividing motor output power by the motor efficiency: The motor efficiency varies slightly with load. For full load, it could be 94 percent, whereas for half load it could drop to 91 percent. Motor efficiency of more than 98 percent is not uncommon. In practice, the selected motor seldom runs at its full capacity. The actual loading required will usually be less than the full load capacity due to the following reasons:
- Operation at off-peak hours
- Station opening year is not at the full capacity of the design year
- Safety factors and margins applied to the fan pressure and the fan motor
- Oversized motor.
Constant Speed Motors. The delivered flow will be the same irrespective of cooling load requirements.
Variable Speed Motors. If the flow rate required at part load is 60 percent of the peak hour capacity, the variable speed drive will tune down the fan speed to match with the new duty. According to fan law, the pressure is proportional to the square of the flow rate and the flow rate is proportional to the fan speed. Power consumption is the product of pressure and flow and, hence, the power reduction at reduced speed will be in the order of a cubic function.
For a 40 percent reduction in flow rate, the corresponding savings in power could be above 70 percent. The motor energy consumption is then determined by the sum of motor input power multiplied by operating hours.
Application to Pumps. The above discussion pertains to fans in air handling units. In theory, VSDs can also be applied to chilled water pumps and condenser water pumps, depending on system design requirements. Condenser water circuit is usually constant flow design to prevent high pressure cut-out of the chiller compressor. Capacity control is achieved by cycling cooling tower fans and step control of condenser water pumps/chillers. Chilled water circuit is also a constant flow design for primary circuit to prevent freezing of evaporator tubes. Capacity control is achieved by step control of chilled water pumps/chillers via the pressure differential- by-pass valve.
Technical Considerations
VSDs for industrial applications are nowadays more and more popular. Most of the VSD technology adopted adjustable frequency, and are also commonly known as inverters. There are some limitations of VSD systems. In practice, the minimum turn-down capacity of the system will be in the order of 30 percent to 40 percent. There are inherent requirements such as minimum fresh air and minimum air circulation / air change rate, which preclude the system from operating below the minimum turn down capacity.
Frequency inverters also tend to produce noise and harmonics, which may affect other electrical/ electronic systems. However, the noise and harmonics can be filtered and controlled. VSD will also induce a loss in the system but its efficiency is usually quite high as 97 percent.
Thermal Storage Systems
A thermal storage system is attractive when one or more of the following conditions apply:
- Peak cooling loads are of short duration.
- Loads occur infrequently.
- Loads are cyclical in nature.
- Loads are not well matched to the availability of the energy source.
- Energy costs are time dependent, e.g., off peak tariff.
- There are utility rebates, tax credits or other economical incentives for load shifting.
- Energy supply from utility is limited, thus limiting the use of full size non-storage system.
The cooling load characteristic of a railway station does not necessarily favour the use of a full thermal storage system because cooling loads are required for a prolonged period and the cooling load profile is relatively flat. A typical station may operate from 6:00 am to 12:00 am (midnight) everyday. The refrigeration plant thus has only 6 hours a day to recharge the thermal storage, which is insufficient for the charging cycle. Full thermal storage would not be economical in this case and partial storage may be considered.
Ice Storage Tank and Ice Making Chiller
Thermal energy can be stored in the form of latent heat of fusion of ice. Water has the highest latent heat of fusion amongst all common materials, 335 kJ/kg. In practice, only one-half of the water in an ice storage tank is actually frozen when it is fully charged. The effective cool storage capacity of ice storage is taken as 209 kJ/kg. The reason is when the thickness of ice is increasing on the exterior surfaces of the tubes, the rate of heat transfer between the brine solution and the water in tank decreases. There is little or no ice forming toward the end of the charging cycle.
Ice storage systems operate with distribution water temperature differentials of up to 13º C (54° F) compared with 5.5º C (42° F) for conventional chilled water system. The compressor for the ice making chiller suffers a penalty because of the much lower evaporative temperature than that in air conditioning. The energy consumption during the ice making cycle would be higher than the air conditioning cycle. The efficiency of the chiller plant could drop from 0.91 kW/TR during the air conditioning cycle to 1.43 kW / TR during the ice making/charging cycle.
There are a lot of available thermal storage methods, such as chilled water, ice or eutectic salt. Among all, ice storage is comparatively more economical when capital investment, spatial acquisition, operation and maintenance costs are appraised.
Cooling Load Profile
The cooling load profile of a typical railway station having a peak summer cooling load of 1,000TR is adopted for this case study and shown in Figure 1. Chiller combinations to match the 1,000TR will be provided for a conventional chiller plant without thermal storage.
When full storage is considered, the required storage capacity is 14,600 TR-hours in a typical summer day. For a 6-hour charging period, the chiller plant capacity required is 2,440TR, which is significantly higher than the no-storage system design.
The capacity of the chiller plant will be smaller than the conventional chilled water plant when partial storage is considered for load leveling during peak hours. At peak hours, the chillers operate in air conditioning mode that cannot generate sufficient cooling capacity to meet the peak demand. The shortage in cooling capacity is supplemented with ice discharging. In ice making mode, the chillers will switch to the lower leaving water temperature set point to produce a low-temperature brine solution. The combined performance of the chillers and storage system is critical to the success of the thermal storage system. For a load leveling partial storage system, the chiller plant capacity is calculated to be 610TR and the storage capacity is 4,000TR-hours.
Recommendations
When the charging period is short, a full storage system is not worth considering because the chiller plant will be significantly larger than the non-storage system and the partial ice storage plant. Although the annual operating costs are the lowest, usually after considerations of initial equipment costs, plant space costs and annual interest rates, no cost savings are anticipated.
The partial storage system reduces the initial chiller plant capacity by about 40 percent but an ice storage plant capacity of 4,000TR-hour is required. The chillers are required to run 24 hours a day continuously. When the cooling load required is lower than the refrigeration capacity, ice will be made and stored. When the cooling load required is higher than the refrigeration capacity, cooling is made up from the ice storage plant. Although the chiller plant capacity is smaller than the non-storage system, the demand for additional plant space is higher because of the need for housing the ice storage plant. The simple payback period is estimated to be in the order of 10 years taking the off-peak energy cost to be about one-third of the peak-hour energy cost.
|