Risk Analysis Of The Power Company’s Power Purchasing In Electricity Markets

The power industry restructuring and deregulation conduct in the world. They guidemarket operators in the electric power industry, break monopoly, encourage competition,improve efficiency, and also make market participants confront with hitherto unknownrisk. Under the traditional power control system, electricity price is determined byregulators according to the cost. The electricity price is stable and predictable in a longperiod. However, the power market reform changes the rules. In the competitiveelectricity market, the power cannot be effectively stored. It is the biggest feature of thepower as a commodity. This defect leads to the low elasticity of power demand, the highprice volatility influenced by the electric power supply and demand relationship. In orderto avoid price risk, the forward, futures and options of the power financial derivativesappear in succession which play a role in reducing or transferring price risk. In thecompetitive electricity market, the power supply company as an independent marketparticipant owns distribution network and provides power purchased from power tradecenter and power generation company to end-users. It is important to study andinvestigate the power company’s trading strategy because the strategy plays an importantrole in optimizing power distribution of different markets and reducing market risk. Inaddition, it has importment significance on stabilizing power market.The paper first introduces the features of power financial derivatives and thefinancial risk measure methods in electricity market. Then, we apply the finance portfoliotheory to the electric power market and research the power supply company and the largeconsumer’s power purchase from different markets. Later, the correlation among powerassets is analysed. And the Copula function is adopted to measure the nonlinearcharacteristic. At last, we make deep analysis about the multi-phase power purchasestrategy of power supply company. The main work and innovation are as following:①We usually analysis the power supply company’s power purchase problem fromthe perspective of expected return and risk. In the existing literatures, multi objectiveoptimization problem is mostly transformed to single-objective problem by the weightedmethod. The single-objective optimization problem can be solved easily. In this paper, wemake the objective function fuzzy and solve the model by fuzzy multi-objectiveoptimization algorithm in order to seek out the relatively optimum solution as close aspossible to each target. Some existing literatures on the large consumer’s power purchase problem risk problem mainly use VaR and CVaR measure methods without consideringthe large consumer’s subjective degree of risk aversion. In this paper, a new method is putforward. The spectral risk function measuring risk is used to extablish the largeconsumer’s power purchase model. Compared with VaR and ES, the spectral risk methodis more flexible and practical without setting the confidence level previously. The largeconsumers can choose spectral risk freely according to the degree of risk aversion.②For the joint distribution of price sequences in different markets, some of theprevious literatures make the hypothesis that price sequence follows a normal distributionand measure the correlation among electric power assets by using linear correlationcoefficient. However, it is studyed that price return series have obvious thick tail andheteroscedasticity phenomenon. The strong nonlinear characteristic between power assetsmakes it hard for linear correlation coefficient to describe their relations properly.Considering the correlation of the profits between the electricity real-time market andfutures market and the statistical characteristics of the profit series, in this paper weestablish the dependence portfolio model-Gumbel Copula-（GARCH-GED，GARCH-t）based on the advantages of the copula function and the GARCH model. We analysis thepower supply company’s purchasing risk based on the model. The empirical results showthat the risk measure results under the copula model is more accurate than the resultsunder binary normal joint distribution.③For the power supply company’s multi-phase power purchase problem, most ofprevious literatures mainly measure multi-phase portfolio risk by simple sum of singleperiod risk or use single-phase succession strategy. In fact, investors mainly consider theminimum risk at the end of the investment instead of making all the single-phase riskvalues as small as possible. Based on the above analysis, we establish a multi-phasepower purchasing model which takes the maximum expected accumulated return and themimimum risk measured by dynamic coherent risk function as objectives. Beside this,time-variation SJC Copula-（GARCH-T，GARCH-GED）model is adopted to fit andsimulate power price series, considering the correlation of the return between theelectricity real-time market and futures market and the statistical characteristics of theprices series. The empirical results show that multi-phase strategy is superior tosingle-phase succession strategy, also better than the entirety strategy.