Stochastic Information Management in Microgrid Operations

By Hao Liang and Weihua Zhuang

Designed to supply electrical and heat loads in local areas based on distributed energy resources, microgrids can significantly reduce power transmission and distribution losses and improve the utilization of renewable energy. Given the randomness in renewable power generation, and the buffering effect of energy storage devices and various customer behavior patterns, stochastic information management—employment of stochastic modeling and optimization techniques for information processing and decision making—is critical for microgrid operations to ensure system efficiency, reliability and sound economics.

In the foreseeable future, energy will continue to be the backbone of the global economy. But because of fast-rising energy prices, climate change and advances in technology, the energy industry is being radically reshaped worldwide. A critical step involves utilization of distributed energy resources for economic and environment-friendly energy production.

According to an International Energy Agency forecast, electric power generation from renewable energy sources will nearly triple from 2010 to 2035, reaching 31 percent of the world’s total power generation, with wind and solar providing large shares. Meanwhile, combined heat and power (CHP) will be harnessed to improve the overall energy efficiency of thermal power plants and reduce the thermal pollution in water systems. The utilization of waste heat can be further improved by using the heat as a source of energy to drive a cooling system such as an absorption refrigerator. Accordingly, the United States Department of Energy aims to have CHP make up 20 percent of the country’s generation capacity by 2030.

Some distributed energy resources—whether wind, solar, CHP or others—can be efficiently integrated in local areas such as a small community, a university or school, or a commercial area, leading to the formation of local, small-scale and self-contained grids, typically referred to as microgrids. Taking advantage of the proximity between power generators and loads, microgrids can significantly reduce transmission and distribution losses and improve the utilization of renewable energy.

The intermittent and weather-dependent output of renewable energy sources such as wind and solar represent a serious complication, which can jeopardize microgrid reliability and cause load curtailment. Energy storage systems such as batteries and flywheels can be integrated in a microgrid to smooth out the intermittent power supply. Also, demand-side management (DSM) programs can help electricity consumers adjust their electricity consumption according to power supply to ensure microgrid reliability.

All such devices and technologies are managed by a microgrid operator. The operator aims to balance power generation and demand while achieving certain objectives such as minimizing operation cost and minimizing system losses, and satisfying microgrid operation constraints in terms of system loading, line flow and voltage constraints.

Over the last decades, the development of information and communication technologies under the umbrella of the smart grid has enabled advanced microgrid monitoring and control to facilitate its operations. Based on the two-way communications throughout the microgrid, the information system can collect microgrid status information, process the information and make decisions on microgrid operation. But with the integration of renewable energy sources, energy storage devices and DSM programs, new technical challenges arise in the information management for microgrid operations. Specifically, the randomness of renewable power generation must be taken into account in microgrid operations, as the predefined microgrid operations schedules may be violated due to power generation fluctuations.

The buffering effect of energy storage devices introduces more state variables to microgrid operations, and requires the modeling of inter-period buffer state transitions over the entire time frame of microgrid operations (in practice, one day in daily microgrid operations for generation scheduling), which leads to high computational complexity. Customer energy demands are especially dynamic and difficult to predict in the presence of DSM programs, as demand can be shifted over time as a result of customer responses to electricity prices.

In order to address the above challenges, stochastic information management can be employed in microgrid operations. Specifically, stochastic modeling and optimization techniques can be used to process microgrid status information obtained from the communication system, such that optimal decisions on microgrid operations can be made. There is a large body of research on stochastic information management in the smart grid. Some standard tools can be applied to address the technical challenges in microgrid operations. A few of the more important approaches follow:

• Monte Carlo simulation. Scenarios are generated according to certain distributions of the random variables in the microgrid such as the power generation from renewable energy sources. A deterministic microgrid operations problem is solved for each scenario based on power flow analysis. The performance metrics of microgrid operations such as power generation costs and system losses can be evaluated based on the solutions of the deterministic problem and the probability of each scenario occurring.
• Queuing theory. It is typically used to analyze the performance of waiting lines or queues of customers. In an islanded microgrid, each energy storage device can be modeled as a queue based on an analogy between the energy stored in the device and the number of customers in a queue. The performance metrics of microgrid operations can be calculated based on the stationary distribution of queue length, which corresponds to the statistics of the amount of energy in the energy storage devices that can be used to supply electrical loads when power generation is insufficient from renewable energy sources.
• Stochastic inventory theory. It addresses optimal design of an inventory (or storage) system to minimize its operation cost. In contrast to queueing models, the ordering (or arrival) process of an inventory can be regulated. In a grid-connected microgrid, energy supply from the main grid is available. The stochastic inventory theory can be applied to optimize the amount of energy drawn from the main grid to recharge the energy storage devices, based on an analogy between energy storage and inventory level.
• Stochastic game. It represents a class of dynamic games with one or more players in terms of probabilistic state transitions. In a microgrid, the randomness of electricity customer behaviors in the presence of DSM can be modeled by probabilistic state transitions. Moreover, due to the competitive nature of the players in the game, the interactions among multiple electricity customers in a dynamically changing system can be characterized, such as in a real-time electricity market.

When applying the stochastic information management tools for microgrid operations, extra care should be taken with respect to the unique features of microgrids, as follows:

• Microgrids have two different operation modes—islanded and grid-connected—which correspond to different microgrid operations objectives. In an islanded mode, each microgrid aims at minimizing its own operation costs or losses. But in the grid-connected mode, which allows energy transactions between the microgrid and main grid, minimizing costs and losses may conflict with the main grid’s cost and loss minimization. Therefore, different stochastic information management schemes need to be developed for different operation modes of a microgrid, and the tradeoffs between main grid and microgrid operations objectives need to be investigated.
• Microgrids are designed to supply the electrical and heat loads in a small geographical area. Power generation and load demands may exhibit substantial spatial correlations due to the similarity of weather conditions in the small area. On one hand, the spatial correlation should be investigated for accurate microgrid modeling. On the other hand, exploiting the spatial correlation may facilitate microgrid operations, specifically for computational complexity reduction by using fewer variables for microgrid representation and decision making.
• Both electrical and heat loads can exist in microgrids with CHP. A two-dimensional model needs to be developed to characterize the output of a CHP plant, as its heat to power ratio varies with system loading. The model is further complicated if the CHP plant is combined with electricity and heat buffers, which have significantly diverse storage and charging/discharging characteristics.

In the application of stochastic information management schemes in microgrid operations, there are two major research issues. The first and foremost one is computational complexity. In comparison with deterministic information management, the computational complexity of stochastic information management is significantly higher. Reducing the computational complexity is a critical step to bridge the gap between research and implementation. The second issue is the requirement of statistics of power generation from renewable energy sources and customers’ responses to DSM programs. Such information may not be available in real-time microgrid operations. How to achieve joint learning and optimization in microgrid operations needs further investigation.

Despite all the technical challenges, stochastic information management is a major avenue for microgrid operations in order to harness renewable energy sources, energy storage devices and DSM programs, such that the economical and environmental benefits of microgrids can be fully realized. The related research is interdisciplinary in nature and calls for a close collaboration between the researchers in power/energy system discipline and information/communication system discipline.

Contributors