DSM in an area consisting of residential, commercial and industrial load in smart grid

Abstract

With the latest introduction of the demand side management (DSM) in smart grids, the power distribution units are able to modify the load schedules of the consumers. This involves a co-operative interaction of the utility and the consumers so as to achieve customer load modifications in which the customer, utility and society all are benefited. The interaction is performed with the help of the devices known as the smart meter. This paper shows the use of game theory and logical mathematical expressions in order to achieve the objectives. The objectives are to minimize the peak to average ratio (PAR) and the energy cost. The outcome of the game between supplier and customers helps to shape the load profile. The design proposed in this paper is very user-friendly and is based on simple logarithmic programming computations. In this paper, residential, commercial and industrial types of loads are taken into account. A basic 24 h load schedule along with the fluctuating prices at each hour of the day is forecasted by the supplier of the various shiftable and non-shiftable loads and then that schedule is conveyed to the user. The users are encouraged to shift their high load devices to off-peak hours which will not only reduce their electricity costs but also substantially reduce the PAR in the load demand.

Keywords

demand side management (DSM) / smart grids / peak to average ratio (PAR) / smart meters and logarithmic price function

Introduction

A smart electrical grid uses information and communications technology. The grid is modernized to gather and act on information. It takes the information about the behaviors of suppliers and consumers in an automated fashion which is deployed to improve the efficiency, reliability, sustainability and economics of the production and distribution of electricity. A smart grid precisely limits the electrical power down to the residential level, network small-scale distributed energy generation and storage devices. It communicates information on needs and operating status, prices and grid conditions. Besides, it moves the grid beyond central control to a collaborative network. As customers can choose their electricity suppliers, depending on their different tariff methods, competition among supplier will increase. The reduction of maintenance and replacements costs stimulates more advanced control. The need of the hour is to use the available resources wisely and that is only possible if the ways of utilization are smartened up. DSM is one of such attempts in the area of power distribution. Earlier, only the production side was involved in contribution. Now, with the recent technological advances in the field of user friendly devices and intelligent software, even the consumers are able to contribute in the cause. With the rapid rate of dwindling fossil fuel and the increase of mercury due to the emission of global house gases, many countries around the globe are opting for smarter ways to meet their power demands. For the success of any idea, the contribution should come from both parties. In addition, it is natural that anyone who participates in any event wants to draw maximum benefit out of it. DSM is one of the important domains in a smart grid. It is customer response oriented. It helps the energy providers to reduce the peak load demand and reshape the load profile resulting in increased sustainability of the smart grid as well as reduced overall energy cost and carbon emission.

A lot of research has been conducted of smart grid in the past decade. Many evolutionary algorithms have been mapped into mathematical model in order to facilitate an efficient working of smart grids. One of the earliest works done in the area of DSM is using temporal pattern clustering algorithms [ 1] to study the fluctuation analysis of load curve pattern. What it did was that it did not compare the concrete data of power load but compared the curve shape. However, the advantage was that it could classify and cluster the main customers who created the peak and off-peak load of total consumption. These customers were the important management object to implement the DSM strategy. To schedule a significant number of varied interruptible loads over 16 h, binary particle swarm optimization (BPSO) [ 2] was used, which was capable of achieving near-optimal solutions in manageable computational time-frames for this relatively complex, nonlinear and non-continuous problem. It offered some potential advantages in scheduling significant numbers of widely varied and technically complex interruptible loads. Then came the genetic algorithm [ 3], the two main goals of which were minimizing instantaneous power demand and keeping the performance of the system good in store and retrieval times. It was observed that the initial load peak was reduced by 21%. Besides, the algorithm took 51 iterations to converge. The three step optimization [ 4] was able to reshape a demand profile full of peaks to a nicely smoothed demand profile. The ultimate goal was to reduce the CO₂ emission resulting from the generation of electricity. A less number of iterations were required. Only freezer was considered in this paper. It provided a better approach to determine the time to stop the planning in order to reduce the amount of iterations required during planning. One of the most important breakthroughs was achieved by the real-time pricing algorithm [ 5], which was used to find the optimal energy consumption levels for each subscriber to maximize the aggregate utility of all subscribers in the system in a fair and efficient fashion. Here not only the energy provider, but also the users benefited. Also, a system with multiple energy providers could be considered. The effect of malicious users could also be explored. The day-ahead load shifting technique was mathematically formulated as a minimization problem in Heuristic algorithm [ 6]. The simulations were conducted on a smart grid which contained a variety of loads in three service areas. The simulation results showed that the proposed DSM strategy achieved substantial savings and reduced the peak load demand. A real-time DSM strategy, together with this proposed day-ahead DSM strategy and DSM, could be developed for real-time operation of smart grid. The authors planned to carry out further work on real-time demand response on a distributed artificial intelligent platform provided by multi agent system. The home energy management (HEM) system [ 7] presented algorithms and architecture models for a HEM system. It made sure that supply and demand balance at real time was the main problem of smart grids. It was advantageous as it allowed integration of large scale renewable energy resources. The code generation made it possible to simulate the architecture in a future work and show its effective gain from the present system. The objective of game-theoretic energy schedule (GTES) method [ 8] was to reduce the peak to average power ratio (PAR) by optimizing the user’s energy schedules. It used a logarithmic price function. The program running time was shorter than conventional methods when the number of users increased. The GTES method was compared with the work in Ref. [ 9] and it was found that the running time and PAR decreased significantly.

Problem formulation

The users are given the forecasted feasible load schedule by the suppliers. Given the feasible energy scheduling set and the real time pricing (RTP) model [ 10, 11], the key question is: How should each user’s energy consumption be scheduled in response to time-varying prices? Before answering this question, it is argued that the user’s interest is twofold. First, each user wishes to minimize his/her payment. In fact, it is reasonable to assume that all users care about the amount on their electricity bills. Second, depending on the appliance, some users may also care about their comfort and getting the work done. The two main objectives are to minimize PAR and to minimize the energy cost. This can be done by minimizing the peak load demand (L_peak).

(1) \mathrm{PAR}={\displaystyle \frac{L_{\mathrm{peak}}}{L_{\mathrm{mean}}}}.

To minimize the cost function, a logarithmic function is used

(2) C_h\left(L_h\right)=\alpha L_h\log \left(L_h+\mathrm{1}\right),

where C_h\left(L_h\right) is the cost function, L_h is the total hourly load in an hour h, and \alpha is price parameter which is always greater than 0.

(3) \alpha ={\displaystyle \frac{p{\displaystyle \sum _{h=\mathrm{1}}^{\mathrm{24}}{L}_h}}{{\displaystyle \sum _{h=\mathrm{1}}^{\mathrm{24}}{L}_h^{\mathrm{2}}\log \left(L_h+\mathrm{1}\right)}}}.

There are certain constraints which have to be kept in mind. For instance, that the number of devices shifted cannot be a negative value, and the number of devices shifted away from a time step cannot be more than the number of devices available for control at the time step.

(4) {\displaystyle \sum _{t=\mathrm{1}}^N{X}_{kit}}\le N_{\mathrm{Ctrlable}\left(i\right)},

where X_{kit} is the number of devices of type k that are shifted from time step i to t and N_{\mathrm{Ctrlable}\left(i\right)} is the number of devices of type k available for control at time step i. The user’s optimized consumption in an hour h should be greater than that of the user’s non-controllable load schedule in the hour h. The total shifted load should be less than or equal to the shiftable load.

Flowchart of proposed algorithm

Flowchart of proposed algorithm is shown in Fig. 1:

The proposed algorithm is as follows:

Step 1: start.

Step 2: choose a random user.

Step 3: initialize the forecasted load. This is the 24 h load schedule which is provided by the consumers.

Step 4: initialize the variables i and j where i and j denote the number of iterations.

Step 5: calculate the value of α using Eq. (3).

Step 6: calculate the hourly price using Eq. (2).

Step 7: check if the hourly load price is greater than the average price. If greater, ask the user if they want to shift their load. If not greater, continue the iteration.

Step 8: get the new schedule from the user for 24 h. If they want to shift their load, repeat Steps 4 and 5. If they do not want to shift, continue with another random user.

Step 9: continue till all users are satisfied with their schedule.

Step 10: stop.

DSM in an area consisting of residential, commercial and industrial load

The forecasted load demand and wholesale energy prices for 24 h of three main sectors, i.e. residential, commercial and industrial load [ 6] were taken into account. With the proposed algorithm, the new wholesale prices and the hourly loads were calculated in order to meet the objective to reduce the peak load demand and the cost of energy.

It can be observed from Tables 1 and 2 that the actual wholesale price is not charged according to the load schedule in the input data from Ref. [ 6]. But the proposed wholesale price in the paper is based on the hourly load.

Results and inference

The PAR reduction comparison is shown in Table 3. It can be observed that the percentage reduction in peak load demand is more in the proposed algorithm for residential area than commercial and industrial load. The variation is comparatively less in industrial load.

From Fig. 2 it can be seen that the peak has been reduced which also results in reducing PAR. It is also observed that in load curve before implementation of DSM there were a sudden addition and subtraction of load, but in the proposed case it has been reduced significantly.

The comparison of cost reduction is shown in Table 4. It can be observed that the cost of load without DSM is much greater than that of the proposed algorithm without DSM. Also, the cost with DSM has been decreased in the proposed algorithm.

It can be seen from Figs. 3 and 4 that the quadratic cost function gives a nonlinear curve whereas the proposed logarithmic cost function gives a much more linear curve, indicating that users should not pay extensively for using load at peak times.

Conclusions

Both users and power companies can benefit from the proposed algorithm. For users, the energy bill has been reduced. For power companies, the installation of peaking generation capacity has been reduced. With the reduced peak load, the energy cost associated with it was avoided. Also the transmission and distribution capacity were reduced.

Simulations were conducted on a smart grid which contained three different kinds of customer areas. The simulation results showed that the proposed algorithm was able to handle controllable devices of several types. It was able to achieve substantial savings while reducing the peak load demand of the smart grid.

When the deviation in the load schedule is great, it is better to use the proposed algorithm. The more number of controllable loads and the more reduction in peak are possible. Price function plays a major role in reducing the energy bill of the customers according to the results obtained from the proposed algorithm.

References

[1] Wen L. The application of temporal pattern clustering algorithms in DSM. In: 2006 6th International Conference on Intelligent Systems Design and Applications. Jinan, China, 2006, 569–573

[2] Pedrasa A A, Spooner T D, MacGill I F. Scheduling of demand side resources using binary particle swarm optimization. IEEE Transactions on Power Systems, 2009, 24(3): 1173–1181

[3] Cárdenas J J, Garcia A, Romeral J L, Andrade F. A genetic algorithm approach to optimization of power peaks in an automated warehouse. In: Proceedings of the 35th IEEE Industrial Electronics Society Congress. IEEE Press, 2009, 3297–3302

[4] Bakker V, Bosman M G C, Molderink A, Hurink J L, Smit G J M. Demand side load management using a three step optimization methodology. In: 1st IEEE International Conference on Smart Grid Communications. Gaithersburg, USA, 2010, 431–436

[5] Samadi P, Mohsenian-Rad A H, Schober R, Wong V W S, Jatskevich J. Optimal real-time pricing algorithm based on utility maximization for smart grid. In: IEEE International Conference on Smart Grid Communications, Gaithersburg, USA, 2010, 415–420

[6] Logenthiran T, Srinivasan D, Shun T Z. Demand side management in smart grid using heuristic optimization. IEEE Transactions on Smart Grid, 2012, 3(3): 1244–1252

[7] Et-Tolba El H, Maaroufiand M, Ouassaid M. Demand side management algorithms and modeling in smart grids. In: 2013 International Renewable and Sustainable Energy Conference. Ouarzazate, Morocco, 2013, 531–536

[8] Fadlullah Z, Quan D M, Kato N, Stojmenovic I. GTES: an optimized game-theoretic demand side management scheme for smart grid. IEEE Systems Journal, 2014, 8(2): 588–597

[9] Mohsenian-Rad A H, Wong V W S, Jatskevich J, Schober R, Leon-Garcia A. Autonomous demand-side management based on game-theoretic energy consumption scheduling for the future smart grid. IEEE Transactions on Smart Grid, 2010, 1(3): 320–331

[10] Mohsenian-Rad A H, Leon-Garcia A. Optimal residential load control with price prediction in real-time electricity pricing environments. IEEE Transactions on Smart Grid, 2010, 1(2): 120–133

[11] Chuang A S, Gellings C W. Demand-side integration in a restructured electric power industry. In: 42nd International Conference on Large High Voltage Electric Systems. Paris, France, 2008

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg