Design and Analysis of Distributed Energy Management Systems

This chapter introduces an activity-based approach to model home and transport energy demand. In this method, in-home and out-of-home activities are simulated and home and transport energy demands are quantified accordingly. This method enables quantification of energy demand with realistic temporal variation. An important application of this method is the analysis of an energy management system that integrates a home and an electric vehicle, which can be achieved when in-home and out-of-home activity is consistently modeled. If this method is applied to a group of households, electricity demand flexibility that electric vehicles and various home appliances, such as washing machines and electric water heaters, provide can be quantified. To model home energy demand, occupancy and in-home activities such as sleeping, cooking, and laundry are stochastically simulated. Appliance operation and energy consumption are determined based on simulated occupancy and activity. To model transport energy demand, out-of-home activity is modeled as trips considering travel time, duration, destination, and mode (e.g., car, walk, and train). Energy consumption derived from the trip is then quantified accordingly. To develop models, data describing people’s activity, which many countries conduct surveys to collect, are used. After introducing the modeling approach, this chapter presents a case study to demonstrate how data provided by time use and person-trip surveys are used in model development and how home and transport energy demands are integrated through activity modeling.

Electric vehicles (EVs) are expected to work, not only as transportation means, but also as power storage units in energy management systems (EMS) given their high capacity batteries. To utilize the in-vehicle battery in an EMS, considering the acceptance by the users, the EMS should be able to identify when the vehicle is being driven and when it is parked, which represents the profile of departure and travel time. This chapter presents a method to predict the most probable car use profile over 1 day, based on statistics of the customer’s daily car use. The prediction problem is formulated as a maximum-likelihood estimation problem and the usefulness of the proposed method is evaluated by numerical experiments.

This chapter proposes a home energy management system (HEMS) which simultaneously controls the charge/discharge process of an in-vehicle battery of an electric vehicle (EV) and the operation plan for a heat pump water heater (HPWH). We propose a model predictive control method, which iteratively calculates the plan of charging and discharging of the in-vehicle battery and the operation schedule of the HPWH, to minimize the electricity bill with consideration of both reverse surplus power as a penalty and vehicle usage as a constraint. Notably, a piecewise linear model for the HPWH is introduced in the optimization problem in order to realize a more realistic computation time. As a result, both suppressing the surplus power reversing to the grid, and decreasing electricity bills of the residents, are realized. The effectiveness of the proposed system is verified in simulations using real data of a household.

Electric vehicles (EVs) have been recognized as a practical solution to several environmental and energy conservation problems. The mileage per charge of EVs, however, is lower than the mileage of internal combustion engine vehicles (ICEVs). In this study, a range extension autonomous driving (READ) system that considers the traffic signal information is proposed. The proposed system optimizes the velocity profile of autonomous driving based on the precise loss models of vehicles. We conducted simulations and experiments that prove the effectiveness of the proposal in terms of mileage per charge.

This chapter investigates a decentralized energy management strategy for the community consisting of households equipped with EV storage. We consider the power flow balancing between a prediction-based power supply and the actual demand of the community. Each agent, i.e., a household, is allowed to determine its own storage charging or discharging set-point. Meanwhile, the utility, which is an independent entity and corresponds to the management office of the community, attempts to realize a socially optimal solution that enables steady-state power flow balance. To align the individual decision-making of each agent with the socially optimal solution, the utility is allowed to provide an additional price, which conceptually represents a tax or subsidy for the agents. The proposed real-time pricing strategy does not depend on the number of agents, and it allows plug-and-play type operations. This conforms to the EV storage management problem, since unpredictable connections or disconnections of the vehicles may occur. The effectiveness of the proposed pricing-based decentralized management strategy is evaluated by numerical experiments.

At the University of Delaware, we are providing regulation services by controlling bidirectional power transfer between a fleet of electric vehicles (EVs) and the power grid. As EVs become more popular, thus increasing the size of our EV fleet, large-scale control will become an important challenge. Power transfer of thousands of EVs may need to be controlled in a way that considers driver requirements while providing regulation services. To cope with this challenge, we propose a grid-integrated vehicle (GIV) control approach that classifies EVs with similar charging characteristics into bins. Instead of calculating a charging schedule for each EV, this approach calculates a schedule for each bin, significantly reducing the scheduling problem computational complexity. In simulations, the proposed GIV control approach is compared against two alternative solutions: a centralized GIV approach and a distributed GIV approach. Our results show that this GIV approach can combine the scheduling qualities of a centralized GIV approach with the scalability of a distributed GIV approach.

Assessing the impact of coordinated use of electric vehicles (EVs) on power distribution grids is of basic importance for designing integrated transportation-energy management. In this chapter, we address the assessment problem on how charging operations of a large population of EVs impact the spatial voltage profile of a power distribution grid. Unlike the conventional space-discretization approach (also known as power-flow equations), we introduce an alternative approach for the assessment based on a continuum representation of the distribution voltage profile. Continuum representation explicitly keeps spatial (or geographical) information on a distribution grid and thus enables us to quantify the spatial impact of EV charging, such as how far charging of an EV at a particular location affects the voltage profile. Here we demonstrate the approach using numerical simulations of a simple distribution grid that are incorporated with operational data on EVs in an EV-sharing system demonstration project.

In this chapter, we discuss cyber security issues that can arise in the voltage regulation of distribution systems. As the amount of distributed generation rapidly increases, voltage regulation along the distribution lines becomes more complicated. This change in the system has prompted the use of more voltage measurements collected over networks for accurate monitoring and control. Here, we provide an overview of our recent research on false data injection attacks on voltage measurements transmitted by sensors to a centralized controller. Our approach is to first introduce an attack detection algorithm and then enhance the security level by a resilient controller that is capable of performing the regulation in the presence of attacks by utilizing the detection results. We illustrate our methods by simulation studies on a realistic, small-scale distribution system.

The importance of cybersecurity has increased with the networked and highly complex structure of computer systems, and the increased value of information. Traditionally, control systems did not use networked communication systems. So, the cybersecurity was not important for the control systems. The networked control systems such as an intelligent distribution network system and so on are appearing and the cybersecurity will become very important for the control systems in the near future. However, we have few actual cyberattacks against the control systems. The intrusion detection should be developed by using only normal control system communication. This chapter consists of two parts which are intrusion detections for the control system communication without sequence patterns and for the control system communication with sequence patterns. The first part is an intrusion detection for the control system communication without sequence patterns. In the first part, we compare supervised machine learning based intrusion detection methods with unsupervised machine learning based intrusion detection methods. The supervised learning are C4.5 and support vector machine. And the unsupervised machine learning are local outlier factor, one-class support vector machine, and support vector domain description. We applied these intrusion detection methods to the water storage tank control system communication data and the gas pipeline control system communication data, and compared the differences in the performance. The second part is an intrusion detection for the control system communication with sequence patterns. In the second part, we compare conditional random field based intrusion detection with the other probabilistic models based intrusion detection. These methods use the sequence characteristics of network traffic in the control system communication. The learning only utilizes normal network traffic data, assuming that there is no prior knowledge on attacks in the system. We applied these two probabilistic models to intrusion detection in DARPA data and an experimental control system network, and compared the differences in the performance.