Load-management applications for the industrial sector

S. Ashok*, R. Banerjee
Energy System Engineering, Indian Institute of Technology, Bombay, India

Abstract The goal of any load-management program is to maintain, as nearly as possible, a constant level of load, thereby allowing the system load factor to approach 100%. The important benets of load management are reduction in maximum demand, reduction in power loss, better equipment utilisation and saving through reduced maximum demand charges. Load shifting, one of the simplest methods of load management, is to reduce customer demand during the peak period by shifting the use of appliances and equipment to partial peak and o-peak periods. Here no loads are being switched o, but only shifted or rescheduled, and hence the total production is not aected. In this paper, a fully edged program is developed for load shifting and the same has been tried with the actual load data collected from a typical fertiliser and chemical industry plant.
Keywords: Load shifting; Peak demand; TOU rates

1. Introduction Electrical energy is an essential ingredient for the industrial and all round development of any country. Electrical energy today constitutes about 30% of the total annual energy consumption on a world wide basis. This gure is expected to rise as oil supply for industrial uses becomes more stringent. As far as India is concerned, about 60% of the total power produced is consumed by industrial sector. The electricity requirement of India has grown tremendously and the demand has been running ahead of supply. With the present installed power capacity of about 88 000 MW, India is facing an energy shortage of 12% with a peak decit of 18 to 20%. Within 10 years, a total peak demand of 200 000 MW is forecasted. To satisfy this demand, additional capacity is needed. But the cost involved for adding new generation

106

capacity is about Rs. 5 crores per MW, besides the cost of transmission and distribution of about Rs. 4 crores per MW. This huge investment for the capacity addition and corresponding networks, forces the authorities and utilities look into various energy conservation and energy management programmes; hence the importance of load management (LM) as an energy-management technique, particularly in the industrial sector, to reduce the peak demand and energy requirements [2]. 2. Load management Load management [1,6,7,12] is the process of scheduling the loads to reduce the electric energy consumption and or the maximum demand. It is basically optimizing the processes/loads to improve the system load factor. Load-management procedures involve changes to equipment and/or consumption patterns on the customer side. There are many methods of load management which can be followed by an industry or a utility, such as load shedding and restoring, load shifting, power wheeling, installing energy-ecient processes and equipment, energy storage devices, co-generation and non-conventional sources of energy, and reactive power-control [3]. Meeting the peak demand is one of the major problems now facing the electric utilities. With the existing generating capacity being unmanageable, authorities are forced to implement load shedding in various sectors during most of the seasons. Load shifting will be a better option for most industries. Load shifting basically means scheduling the load in such a way that loads are diverted from peak period to o-peak periods, thereby shaving the peak and lling the valley of the load curve, so improving the load factor. To encourage load shifting in industries, and thereby to reduce peak demand, many utilities have already implemented time of use rates (TOU) or have plans for introducing such rates [12]. 3. Load shifting This can be considered to be a better and simpler method of load management because it does not change the total production output and is easy to implement. In this technique, the total processes (loads) are rescheduled, and not switched o, in such a way to reduce the peak demand and improve the load factor. It is not possible to reschedule all the processes or shift certain types of loads. So the load-shifting technique has to select the loads which can be shifted or processes which can be rescheduled. Constraints like interlocked processes and storage-space limitations will have to be studied in detail before implementing the techniques. 3.1. Methodology of load shifting This is the use of economic incentives and disincentives oered through the electric rate structure to encourage voluntary changes in customer consumption-patterns.

107

Under the load-shifting method, electric power to all loads is always available, but the rate may vary with the time of day or use (TOU) and depend on the power demand as well as total energy consumed [11]. The aim of the load-shifting method is to reduce customer demands during peak periods by shifting the use of appliances and equipment to o-peak/partial peak periods. The rates used with the load-shifting method may also encourage new loads to occur during the o-peak period. No inconvenience or discomfort results from the load-shifting method, though there may be economic penalties resulting from excessive peak-period electricity use [8,9]. 3.2. Load or process constraints In industries, there will be a large number of processes (loads). Sometimes it is not possible to shift a particular load alone from one interval to another. For example, a particular process may be inter-locked with other processes. Or sometimes, certain loads or processes are continuous. Thus, loads having process interconnections or loads in a continuous process chain cannot be controlled independently. These loads will have to be controlled together. In certain cases, in between two process there may be a storage space. The limitations of the storage space will result in a storage constraint. 3.2.1. Interlocked processes In the example shown in Fig. 1, two loads, L1 the conveyer and L2 the heater and crusher processes of a fertilizer and chemical plant, are shown. The conveyor is loading the heater and crusher. Being continuous, these two loads are interlocked and cannot be controlled independently. For any LM program, both loads will have to be controlled together. We cannot shift process L1 or L2 independently. If required, we have to shift these two processes together. 3.2.2. Storage-space constraint Industries may have a number of storage spaces with loads associated with them (for example, see Fig. 2). The raw water pump L5 is pumping water from a river to the storage tank (S). After treatment of the water inside the tank, it is pumped out by the treated water pump L10.

Fig. 1. Interlocked process.

108

Fig. 2. Storage space constraint.

We cannot control these processes (pump loads) independently without considering the storage capacity of the storage space (S units). Pump L5 has to stop whenever the water upper limit is reached in the tank. Similarly, there are several constraints for the operation of the pumps. Hence we have to control or shift the processes L1 and L2 by considering the capacity of the storage space (S units) as well as the pump ow (discharge and suction) rates. 4. Implemantation and results The load-shifting technique is applied to a typical fertilizer and chemical industry plant in Cochin, Kerala, India [11]. This plant is selected because (1) the processes are energy-intensive and (2) the plant is supplied by Kerala Grid, where TOU rates have been implemented. Load data have been collected from the industry on a normal shift day. For simplicity, the load data are taken only for every hour. The total connected load of the plant is 24.7 MW, and is composed of 12 independent process-plant loads, with details shown below. The contract maximum demand (CMD) is 21.5 MW:
Primary treatment plant Heater and crusher Raw water pumping station Oxide plant Chemical inspection and lab Lighting L1=2000 kW, L3=500 kW, L5=4250 kW, L7=1300 kW, L9=750 kW, L11=250 kW, Conveyor L2=100 kW, Nitrogen plant L4=1925 kW, Ammonia processor L6=2100 kW, Acid plant L8=1000 kW, Treated water pump load L10=5000 kW, Urea plant L12=4525 kW.

For simplicity, we have not considered the individual loads of the equipment in each plant separately. Most of the individual loads of each plant are interconnected, so giving less chance for load-shifting independently. The costs of the total electrical energy before shifting and after shifting are calculated with the TOU tari rate applying by the Kerala grid [10]. The tari for the industrial sector is as follows:

109

Flat rates:

Demand charge: Rs. 133/kVA Energy charge: Rs. 1.22 per unit (Note: present transaction rate 1 US$Rs. 42.0) TOU rates: Energy charges: partial peak 6 am6 pm: at rate (normal) peak period 6 pm10 pm: double O peak 10 pm6 am: quarter Steps in the load shifting algorithm made are as follows: 1. Read the load data at dierent time instants; contract maximum-demand (CMD) or any preferred maximum-demand limit (PMD), TOU rates, `not possible' loads for shifting with their time intervals and constraints. 2. Calculate the maximum demand (MD) at each time instant and cost of electricity for the day. 3. IF MDCMD or PMD, nd all the loads (excluding those `not possible' for the instant for shifting) `on' and select the smallest. 4. Check for constraints: 4.1. if interlocked, select all the interlocked loads together 4.2. if storage constrained, see the constraints and follow. 5. Shift the loads to other intervals where MD<PMD. Calculate new MD and cost of electricity for the day with new load data. 6. Do the steps for all possible loads and for all the time instants in a repetitive manner such that (cost of electrical energy for the day) is a minimum.

Here, `not possible' loads for a time period indicate loads which cannot be shifted from that period to any other period, e.g. lighting load. As already said, the loads L2 and L3 are interlocked. The conveyor is feeding the raw material to the heater crusher. We have to shift these two processes simultaneously only if required. L5 and L10 are considered to be storage space constraints with control as follows: start L5 stop L5 operate L10 (1) (2) (1) whenever water level =Lmin provided L10 is on when water level is < Lmax and period is not peak whenever water level is =Lmax whenever water level is > Lmin and the period is not a peak, for a total time period of T h/day sucient for lling the treated water tank for the day's requirement.

Results of the load-shifting implementation are given in Table 1.

Table 1 Results for load-shifting technique [4,5] Before shifting Total electricity cost Maximum Total demand energy 23.7 MW With at rate (Rs.) With TOU rate (Rs.) Maximum demand Total energy After shifting Total electricity cost With at rate (Rs.) With TOU rate (Rs.)

517.4 MWh 3, 7 83, 328 3, 8 41, 354 22.85 MW 517.4 MWh 3, 670,2 78 3, 734, 068

110

Fig. 3. Load curves.

The daily load curves for the industry before and after load shifting are shown in Fig. 3. By applying the LM program, the load factor is improved by 4.5%. The reduction in peak demand is 0.85 MW, thereby saving electricity costs by Rs. 107 286 for 1 day. 5. Conclusion It is shown that by implementing the load-shifting technique, the maximum demand can be reduced and hence the load factor can be improved. Reducing the

111

peak demand will reduce the electricity cost for the industries especially with the TOU taris. Utilities will get a relief as the peak of the load curve is attened. Loadshifting techniques will be a better method for load management programs in industries, because they do not change the total production output and are easier to implement. The application of the load-shifting technique to a typical energy intensive industry reveals the scope of the program for the industrial sector.

References
[1] Bjork CO. Industrial load management theory practice & simulation. New York: Elsevier, 1989. [2] Parikh JK, Reddy BS, Banerjee R. Planning for demand side management in the electricity sector. New Delhi: Tata McGrawHill, 1994. [3] Cochen AI, Wang CC. An optimisation method for load management scheduling. IEEE Transactions on Power Systems 1988;3(2):6128. [4] Bhagwan Das D, Patvardhan C. Solution of economic load dispatch using real coded hybrid stochastic search. Electrical Power and Energy Systems 1998;21(3):16570. [5] Chirattananon S et al. A model for the performance evaluation of the operation of EAF. Energy Coserv Mgmt 1996;37:1616. [6] Vogt LJ, Conner DA. Electrical energy management. Toronto: Lexington Books, 1977. [7] Isaksen, L, Ma, FS, and Simons, NW. Bibliography and load management. IEEE Transactions on Power Apparatus and Systems 1981: PAS-100(5):1981. [8] Nagarath, IJ, Kothari, DP. Modern power system engineering. Tata McGrawHill, 1995. [9] Wood J, Wollenberg BF. Power generation, operation and control. New York: John Wiley and Sons, 1983. [10] KSEB Engineers Journal, HYDEL 1996:42(1). [11] Bharanam V. Objectivity linked microlevel budget of Kerala State Electricity Board (India). Trivandrum, Kerala: KSEB, 1996. [12] Sheen JN et al. TOU pricing of electricity for load management in Taiwan power company. IEEE Trans on Power Systems 1994;9:38896.