Intelligent Management of End Consumers Loads Including Electric Vehicles through a SCADA System Filipe Fernandes, Pedro Faria, Zita Vale, Hugo Morais, Carlos Ramos GECAD – Knowledge Engineering and Decision-Support Research Group of the Electrical Engineering Institute of Porto – Polytechnic Institute of Porto (ISEP/IPP), Rua Dr. António Bernardino de Almeida, 431, 4200-072 Porto, Portugal {fjgf, pnf, zav, hgvm, csr}@isep.ipp.pt Abstract. The large penetration of intermittent resources, such as solar and wind generation, involves the use of storage systems in order to improve power system operation. Electric Vehicles (EVs) with gridable capability (V2G) can operate as a means for storing energy. This paper proposes an algorithm to be included in a SCADA (Supervisory Control and Data Acquisition) system, which performs an intelligent management of three types of consumers: domes- tic, commercial and industrial, that includes the joint management of loads and the charge / discharge of EVs batteries. The proposed methodology has been implemented in a SCADA system developed by the authors of this paper – the SCADA House Intelligent Management (SHIM). Any event in the system, such as a Demand Response (DR) event, triggers the use of an optimization algo- rithm that performs the optimal energy resources scheduling (including loads and EVs), taking into account the priorities of each load defined by the installa- tion users. A case study considering a specific consumer with several loads and EVs is presented in this paper. Keywords: End consumers, demand response, intelligent management, electric vehicles, SCADA 1 Introduction The main goal of power systems is to guarantee that the generation meets the con- sumers demand, including the domestic, commercial, rural and industrial types of consumers. The power system should remain in a stable state, matching generation and demand values. The consideration of consumers’ behavior (participating in De- mand Response events) is one of the distributed energy resources, which have been of increasing importance [1]. In certain periods of the day, mainly when the generation is lower than the demand or the marginal cost of increasing generation is high, the use of Demand Response (DR) programs is interesting for reducing the demand level as an alternative to increase the generation. The demand level reduction in a real time horizon is one of the most common events of DR. In the opposite situation (when the generation is higher than demand), it is possible to store the excess energy in a storage system by optimizing the use of energy resources. The increase of small resources use at lower voltage levels of distribution networks leads to the context of Smart Grid (SG) operation [2, 3]. In this context, the Electric Vehicles (EVs) with gridable capability (V2G) can be used as a storage unit, storing energy when there is an excess of generation and, dis- charging energy when the load is higher than the generation [4]. Only one V2G has no impact on the grid; however, the large integration of V2Gs can have non- negligible positive or negative impacts. The different V2Gs user profiles, which con- sider the daily necessities of people, will cause changes in consumption daily dia- grams [5]. The optimization of load consumption and V2Gs use is able to improve the energy use efficiency, while allowing the system operator to control some loads of installa- tion. This makes possible the increase of energy resources management flexibility [6, 7]. A consumer endowed with an intelligent energy resources management system allows the interaction with the grid operator, improving the effectiveness of the con- sumer‘s participation in a DR event, by receiving and sending event-related infor- mation [8, 9]. A SCADA system must support a decentralized structure to control, monitor, supervise and optimize all consumer energy resources, even in a real time horizon [10]. The paper deals with the intelligent management of a Domestic Consumer (DC) that adds two V2Gs to the system beyond the normal consumption loads. As these V2Gs are used by house users to travel to their respective workplaces, the effects of V2Gs on the intelligent management of a Commercial Consumer (CC) and an Indus- trial Consumer (IC) are analyzed. In this way, an optimization methodology is pro- posed and implemented in a SCADA system considering each type of consumer. This SCADA system considers all loads and V2Gs to perform an intelligent management. Given the differences between the load curves for each type of consumer, it is im- portant to analyze how the management systems will consider the V2G connection when applied to the three types of consumers at different times of the day. After the introductory section, Section 2 presents the proposed methodology and Section 3 describes the energy resources considered by the SCADA system developed by the authors of the paper. A case study is presented in Section 4. The final section includes some conclusions. 2 Methodology applied in SCADA management The intelligent management systems used by different consumers have the same basic structure, being some differences related to the characteristics of each consumer. Con- sidering the context of participating in DR programs, any of the systems used aims to optimize the total consumption of the installation, keeping it lower than the estab- lished limit consumption or cutting power indicated by the system operator or by the installation user [11]. The optimization is directly affected by the users’ consumption patterns and by the context of the day, which depends on several factors, such as the season, the temperature, the day of week, the time and the electricity price [12]. The SCADA House Intelligent Management (SHIM) has been described in previ- ous works, and it was only applied to the DC [13]. In the present work, the integration of V2Gs is also considered. The same SHIM methodology is used to implement intel- ligent management systems for the CC and the IC. These systems include the follow- ing features, management of the power consumption in an installation, while main- taining user comfort and loads operation continuity; adaptability of the system to several daily factors that may influence the consumption; and ability to interact with DR events in the SG context. The base algorithm, presented in [13], is able to manage the installation consump- tion whenever there are changes in the system operation conditions. The same algo- rithm has been improved in order to consider the CC and the IC. As mentioned be- fore, this algorithm considers the consumers’ loads and V2Gs, associated with a usage priority defined by the installation user. In the presence of a new event, the algorithm evaluates the current state of all installation equipment and, in accordance with the priority of each resource. It performs an optimal scheduling regarding the methodolo- gy presented in Figure 1. The Mixed Integer Non-Linear Programming (MINLP) optimization algorithm has been implemented in General Algebraic Modelling Sys- tem (GAMS). Fig. 1. SCADA End Consumer methodology 3 Energy Resources Description 3.1 End consumers with V2G Each consumer installation has its load. Otherwise, V2Gs are common resources for the three considered installations of the SCADA system (see Figure 2), i.e., these resources have different connection points (installations) in different periods of the day as V2Gs travels between them. The current state/position of each V2G depends on the period of the day. For example, on Monday at 9 a.m. (peak consumption in the IC), the V2Gs are connected to the IC network and the charge / discharge periods could be managed by the SCADA system focusing on the IC installation resources use optimization. On the other hand, at 8 p.m. (peak consumption for DC and CC) V2Gs are connected to the housing network. The SCADA system is able to manage two V2Gs at the same period in a house (DC). In the case of the CC or the IC, it is prepared to receive one of the V2Gs that belongs to a DC. This is due to house users who give different functions to each V2G presented in Figure 2: 1. Move one user from the house to the industry (30 km) and return (30 km); 2. Move one user from the house to the commerce (15 km) and return (15 km). The charge/discharge rate considered for both vehicles is 2.3 kW/h and a V2G bat- tery at full charge has 16 kWh. The V2G have 160 km autonomy [14]. In this way, the V2G consumes 1500 Wh to travel 15 km (go to CC) and 3000 Wh to travel 30 km (go to IC). Fig. 2. End consumers with V2G at different times of the day 3.2 Description of End Consumers Loads It has been considered different loads to perform the developed SCADA system, re- garding the three consumers of different types. The GECAD’s Intelligent Energy Systems Laboratory (LASIE) [11-13] loads were considered to represent the DC. The LASIE loads considered for the CC (coffee shop type) and for the IC (textile factory type) are presented in Table 1. Table 1. Loads characteristics of end consumers Domestic Consumer Commercial Consumer Industrial Consumer Type* Max Max Max Load Power Load Power Load Power (W) (W) (W) V1 Induction Motor 1 90 Induction Motor 1 180 Induction Motor 1 450 V2 Induction Motor 2 200 Induction Motor 2 400 Induction Motor 2 1 450 V3 Induction Motor 3 300 Induction Motor 3 600 Induction Motor 3 1 700 V4 Fluorescent Lamp 70 Fluorescent Lamp 700 Fluorescent Lamp 2 800 F1 Incandescent Lamp 1 30 Incandescent Lamp 1 300 Incandescent Lamp 1 300 F2 Incandescent Lamp 2 30 Incandescent Lamp 2 300 Incandescent Lamp 2 300 F3 Heat Accumulator 1 1 600 Coffee Maker 1 500 Heat Accumulator 1 1 800 F4 Heat Accumulator 2 1 000 Heat Accumulator 1 3 200 Heat Accumulator 2 1 800 F5 Halogen Lamp 500 Dishwasher 1 500 Halogen Lamp 2 500 F6 Exhauster 138 Plasma TV 1 276 Sewing Machine 1 690 F7 Refrigerator 1 300 Desktop Computer 600 Iron 1 3 000 F8 Washing Machine 550 Refrigerator 1 1 100 Clothes Dryer 5 500 F9 Television 1 138 Security System 138 Sewing Machine 2 690 F10 Refrigerator 2 300 Dehumidifier 600 Iron 2 1 500 F11 Microwave 550 Plasma TV 2 300 Clothes Washer 2 750 F12 Television 2 138 Sound System 138 Vacuum upright 1 690 F13 Kettle 300 Refrigerator 2 300 Vacuum upright 2 1 500 F14 Dishwasher 550 Air Conditioner 2 750 Cutting Machine 2 750 Total Maximum Power 6 784 Maximum Power 14 882 Maximum Power 32 170 4 Case Study In the case study, the SCADA system is applied for each type of end consumers (do- mestic, commercial and industrial) to perform an intelligent management of their energy resources (loads and V2Gs). The optimal energy resource scheduling is solved by the methodology presented in Section 2. Some scenarios are simulated in order to analyze the scheduling results for both consumers in different periods of the day. All end consumers presented in this work have a database that contains a load pri- ority, and the minimum and maximum power of each load and each V2G, depending on the context of the day. The SCADA system optimization analyzes the database and, according to the context, performs an optimal scheduling. The first development in previous work is the inclusion of two V2Gs in the optimization of the DC installa- tion. After that, the methodology was improved in order to consider the CC and the IC, regarding the different characteristics of these consumers. The IC is the building with higher consumption. In this way, only one V2G should have low impact in the optimization process. The present case study considers a Monday in the winter season, with an external temperature of 10ºC. At 9 a.m. and 8 p.m. the three consumers receive the simulated DR event respectively, according to the results presented in Table 2. 4.1 Timeline Results of the V2G State In order to validate the case study, the results were analyzed. At 9 a.m. and 8 p.m. all end consumers receive the information regarding the DR event to cut or reduce loads consumption or to sell the energy stored in the V2G. The optimization of each end consumer installation depends on the availability of the V2G to be considered by the SCADA system. The timeline of the V2G state is presented in Table 2 according the DR event. The blue color means that the V2G is connected to the SCADA system respective, the red color means that the V2G is travelling and the green color repre- sents the DR event participation at 9 a.m. (100 W to DC, 1450 W to CC and 5235 W to IC) and 8 p.m. (400 W to DC, 600 W to CC and 850 W to IC). Table 2. V2G state and participation in a DR event Domestic Consumer Commercial Consumer Industrial Consumer Hour V2G Charge/ Charge/ Charge/ Energy Energy Energy State Discharge State Discharge State Discharge (Wh) (Wh) (Wh) Rate (W) Rate (W) Rate (W) Connected 16 000 0 Out None None Out None None 1 7 a.m. Connected 16 000 0 Out None None Out None None 2 Travel None 1 500 Out None None Out None None 1 8 a.m. Travel None 3 000 Out None None Out None None 2 Out None None Connected 14 500 1 450 Out None None 1 9 a.m. Out None None Out None None Connected 13 000 2 300 2 Out None None Connected 13 050 0 Out None None 10 a.m. 1 Out None None Out None None Connected 10 700 0 2 Out None None Travel None 1 500 Out None None 1 6 p.m. Out None None Out None None Travel None 3 000 2 Connected 11 550 400 Out None None Out None None 1 8 p.m. Connected 7 700 0 Out None None Out None None 2 Connected 11 150 0 Out None None Out None None 1 9 p.m. Connected 7 700 0 Out None None Out None None 2 At 7 a.m. both V2Gs are connected to the DC SCADA system (with 16000 Wh of energy stored) but consumers do not receive any DR event. At 8 a.m., DC users begin the journey to their workplaces and at 9 a.m. a DR event is announced. At this time, V2G1 is connected to the CC and V2G2 is connected to the IC. This means that the DC will only be able to meet the DR event by cutting or reducing the loads consump- tion. Other end consumers can also use the discharge capacity of the V2G in the op- timization process, according to the priority of the SCADA database. At 9 a.m. all consumers receive the DR event order to reduce 100W, 1450W and 5235W corresponding to the DC, CC and IC respectively. The V2G1 have 14500 Wh of energy due to the journey of 15 km to the CC and V2G 2 have 13000 Wh due to the journey of 30 km to the IC. At 10 a.m. one can verify that the V2G 1 storage energy is of 13050 Wh. This means that V2G1 discharges 1450 W over one hour after the DR event announcement. The discharge value corresponds to the reduce power of the DR event and the loads that were being used by the CC were not changed, ensuring con- sumers’ priorities. Regarding the V2G2, the storage energy is of 10700 Wh. This means that V2G2 discharges 2300W over one hour after the DR event announcement. The discharge value corresponds to a portion of the total reduce power of the DR event (5235 W). In this case, the IC SCADA system also needed to reduce the con- sumption in the lower loads’ priority in order to fully meet the DR event require- ments. At 6 p.m. one can verify that any consumer of the SCADA system have V2Gs connected, because the V2G users begin the return journey, from the workplace to their houses, which will reduce 1500 Wh in V2G1, and 3000 Wh in V2G2. At 8 p.m. all consumers receive the DR event order to reduce 400W, 600W and 850W corre- sponding to the DC, CC and IC respectively. V2G1 has 11550 Wh of energy and V2G2 has 7700 Wh of energy. At 9 p.m. one can verify that V2G1 energy is of 11150 Wh. Thus, V2G1 discharges 400W over one hour after the DR event announcement. The discharge value corre- sponds to the reduce power of DR event and the loads that were used by DC were not changed, ensuring consumers’ priorities. Regarding V2G2, the energy is 7700 Wh, maintaining therefore the initial energy. In this case, the CC and IC see the consump- tion reduced in the lower loads’ priorities in order to fully meet the DR event, 600W and 850W respectively. 4.2 Energy Resources Scheduling Results Tables 3 and 4 present the results of the optimization process to validate the meth- odology proposed in Section 2. Table 3 shows the priority of each load and of each V2G to charge (Ch) or discharge (Dch), when end consumers meet the first DR event at 9 a.m.. Table 4 summarizes the optimization results at 8 p.m. (second DR event). The coloured cells represent the resources which were subjected to changes. The SCADA management system selects the loads or V2G mode (charge or dis- charge) according each priority. For example in first DR event to CC, the resource with low priority is 20 (V2G1 discharge) and the resource with higher priority is 1 (Induction motor #2). The priorities are predefined by the installation users. Table 3. Optimization results at 9 a.m. to end consumers Domestic Consumer Commercial Consumer Industrial Consumer Power Power Power Power Power Power Type Priority before after Priority before after Priority before after (W) (W) (W) (W) (W) (W) V1 12 0 0 9 180 180 12 450 450 V2 17 200 130 1 400 400 17 1 450 615 V3 11 0 0 2 600 600 11 1 700 1 700 V4 10 0 0 10 700 700 10 2 800 2 800 F1 9 0 0 14 0 0 19 300 0 F2 18 30 0 15 0 0 5 300 300 F3 15 1 600 1 600 3 1 500 1 500 3 1 800 1 800 F4 16 1 000 1 000 7 3 200 3 200 18 1 800 0 F5 8 0 0 6 1 500 1 500 2 2 500 2 500 F6 6 0 0 8 276 276 9 690 690 F7 14 300 300 11 600 600 4 3 000 3 000 F8 7 0 0 4 1 100 1 100 15 5 500 5 500 F9 5 0 0 17 138 138 8 690 690 F10 13 300 300 16 0 0 16 1 500 1 500 F11 1 0 0 12 300 300 1 2 750 2 750 F12 3 0 0 13 138 138 13 690 690 F13 4 0 0 5 300 300 7 1 500 1 500 F14 2 0 0 18 0 0 6 2 750 2 750 Total Power 3 430 3 330 - 10 932 10 932 - 32 170 29 235 Ch – 19 V2G1 None None None 14 500 13 050 None None None Dch – 20 Ch – 14 V2G2 None None None None None None 13 000 10 700 Dch – 20 Reduce Power 100 1 450 5 235 Table 4. Optimization results at 8 p.m. to end consumers Domestic Consumer Commercial Consumer Industrial Consumer Type Power Power Power Power Power Power Before (W) After (W) Before (W) After (W) Before (W) After (W) Loads Power 6 784 6 784 12 132 11 532 15 140 14 290 V2G1 11 550 11 150 None None None None V2G2 7 700 7 700 None None None None Reduce Power 400 600 850 In the first DR event, at 9 a.m., the DC fulfilled the reduce power (100W) by turn- ing off the incandescent lamp #2 and reducing the consumption of the induction mo- tor #2 (loads with lower priority). The CC fulfilled the DR event requirement (1450W) through the V2G1 discharge, keeping the same load consumption. The IC beyond the V2G2 discharge, also turned off the incandescent lamp#1 and the heat accumulator #2, and reduced the consumption of the induction motor #2 to guarantee the required reduced power of DR event (5235W). In the second DR event at 8 p.m., the DC fulfilled the reduce power (400W) with V2G1 discharge, keeping the same load consumption. The CC fulfilled the DR event requirement (600W) through the loads with lower priority; the same happened in the IC (850W). 5 Conclusions This paper presents a case study considering a SCADA system to manage and opti- mize the consumption of all energy resources of the DC, CC and IC consumers. The case study is discussed and analyzed applying the methodology to the CC and IC in a particular context, and verifying the usefulness of V2Gs in their management systems. The results of using the proposed methodology regarding working days or weekends specificities, which require some distinct characteristics, have been addressed and will be reported in near future work. In the present work, one can verify that V2G has direct participation and impact in the consumption optimization. The SCADA system of any consumer is provided with the priorities defined by the installation users to each load and each V2G according to the operation context. The SCADA database allows to know the user’s needs in real time in order to guarantee the fulfilment of the DR event requirements. One V2G may have more influence in the DC optimization than in the IC optimi- zation, as the IC is a consumer with higher energy requirements. In this way, one V2G may have little impact in IC resources use optimization, but if we are dealing with a considerable number of V2Gs, this impact must be adequately analyzed. This means that the optimization decisions depend directly on the consumers’ energy needs, on the number of V2Gs considered and on the type of end consumer. The ben- efits of using the proposed SCADA system can be summarized as follows:  Creation of a system with its own capacity for decision in real time to support the grid operator with energy management capability;  The methodology development can be adapted for any type of end consumer and amount of energy resources;  Each SCADA system is able to be adapted to the current system conditions over the day with or without of V2G;  The inclusion of V2G in the SCADA system makes possible to ensure the end consumer comfort through the V2G batteries energy discharge. Acknowledgment This work is supported by FEDER Funds through COMPETE program and by Na- tional Funds through FCT under the projects FCOMP-01-0124-FEDER: PEst- OE/EEI/UI0760/2011, PTDC/EEA-EEL/099832/2008, PTDC/SEN- ENR/099844/2008, and PTDC/SEN-ENR/122174/2010. References 1. P. Faria, Z. Vale, “Demand response in electrical energy supply: An optimal real time pric- ing approach”, Energy, Volume 36, Issue 8, August 2011 2. Z. Vale, H. Morais, H. Kohdr, “Intelligent Multi-Player Smart Grid Management Consider- ing Distributed Energy Resources and Demand Response”, IEEE PES General Meeting, 2010, Minneapolis, MN US, 25 - 29 July, 2010 3. T. Hammerschmidt, A. Gaul, J. Schneider, “Smart Grids are the efficient base for future en- ergy applications”, CIRED Workshop 2010: Sustainable Distribution Asset Management & Financing, Lyon, France, 7-8 June, 2010 4. T. Sousa, H. Morais, Z. Vale; P. Faria, J. Soares, "Intelligent Energy Resource Management Considering Vehicle-to-Grid: A Simulated Annealing Approach," IEEE Transactions on Smart Grid, vol. 3, no. 1, pp. 535 -542, March 2012 5. K. Clement-Nyns , E. Haesen, J. Driesen, "The impact of charging plug-in hybrid electric vehicles on a residential distribution grid", IEEE Transactions on Power System, vol. 25, no. 1, pp. 371 -380 2010 6. S. Fernandes, N. Silva, M. Oleskovicz, "Identification of residential load profile in the Smart Grid context," Power and Energy Society General Meeting, 2010 IEEE, pp. 1-6, 25-29 July 2010 7. S. Tiptipakorn, L. Wei-Jen, "A Residential Consumer-Centered Load Control Strategy in Real-Time Electricity Pricing Environment," Power Symposium, 2007. NAPS '07. 39th North American , pp. 505 -510, Sept. 30 2007 -Oct. 2 2007 8. S. Shao, T. Zhang, M. Pipattanasomporn, S. Rahman, "Impact of TOU rates on distribution load shapes in a smart grid with PHEV penetration," Transmission and Distribution Confer- ence and Exposition, 2010 IEEE PES , pp. 1-6, 19-22 April 2010 9. K. Kok, S. Karnouskos, D. Nestle, A. Dimeas, A. Weidlich, C. Warmer, P. Strauss, B. Buchholz, S. Drenkard, N. Hatziargyriou, V. Liolioum “Smart Houses for a Smart Grid”, 20th International Conference on Electricity Distribution CIRED, Prague, June 2009 10. D. Choi, H. Kim, D. Won, S. Kim, “Advanced key-management architecture for secure SCADA communications”, IEEE Transactions on Power Delivery, vol. 24, no. 3 , pp. 1154- 1163, July 2009 11. F. Fernandes, T. Sousa, P. Faria, M. Silva, H. Morais, Z. Vale, “Intelligent SCADA for Load Control”, IEEE International Conference on Systems, Man and Cybernetics - SMC 2010, Istanbul, Turkey, 12-15 October, 2010 12. F. Fernandes, T. Sousa, M. Silva, H. Morais, Z. Vale, P. Faria, “Genetic Algorithm Method- ology applied to Intelligent House Control”, Symposium on Computational Intelligence Ap- plications in Smart Grid (CIASG), IEEE SSCI 2011 (IEEE Symposium Series on Computa- tional Intelligence), Paris, France, April 11-15, 2011 13. L. Gomes, F. Fernandes, T. Sousa, M. Silva, H. Morais, Z. Vale, C. Ramos, “Contextual Intelligent Load Management with ANN Adaptive Learning Module”, International Confer- ence on Intelligent System Applications to Power Systems - ISAP 2011, Hersonissos, Crete, Greece, 25-28 September, 2011 14. Mitsubishi, "Mitsubishi i-MiEV Technical Specifications", Consulted: May 2012, Available: http://www.mitsubishi-motors.com/special/ev/whatis/index.html.