CAJ | 학술논문

电煤供给和库存是燃煤电厂十分关注的问题。为避免由于电煤供应中断而影响正常发电，燃煤电厂必须具备一定量的电煤储备。在电力市场环境下，电煤价格和发电上网电价都具有不确定性。为了在保证一定的收益率水平上最小化与收益相关的风险，燃煤电厂就需要合理确定电煤库存量。在此背景下，借鉴金融领域发展起来的风险管理理论，以条件风险价值(CVaR)作为风险计量指标，建立了燃煤电厂电煤库存优化的均值-CVaR 非线性规划模型；所构造的模型综合考虑了煤价和发电上网电价的不确定性、电厂煤耗量以及合同煤、市场煤兑现率的波动，以电厂年度收益CVaR最小为目标函数。之后，将所构造的模型转化为线性规划问题进行求解。最后，以某电厂的电煤库存量优化问题为例进行了仿真计算，算例结果表明了所建立模型的合理性和求解方法的有效性。
전매공급화고존시연매전엄십분관주적문제。위피면유우전매공응중단이영향정상발전，연매전엄필수구비일정량적전매저비。재전력시장배경하，전매개격화발전상망전개도구유불학정성。위료재보증일정적수익솔수평상최소화여수익상관적풍험，연매전엄취수요합리학정전매고존량。재차배경하，차감금융영역발전기래적풍험관리이론，이조건풍험개치(CVaR)작위풍험계량지표，건립료연매전엄전매고존우화적균치-CVaR 비선성규화모형；소구조적모형종합고필료매개화발전상망전개적불학정성、전엄매모량이급합동매、시장매태현솔적파동，이전엄년도수익CVaR최소위목표함수。지후，장소구조적모형전화위선성규화문제진행구해。최후，이모전엄적전매고존량우화문제위례진행료방진계산，산례결과표명료소건립모형적합이성화구해방법적유효성。
The coal inventory and supply are two major issues to concern for coal-fired power plants.It is necessary to have a certain amount of coal reserves so as to maintain the stable power generation of a coal-fired power plant.In the electricity market environment,fluctuations of electricity prices and power outputs of power plants concerned are inevitable as the results of market competition and load fluctuations.Similarly,the coal price in the coal market also exhibits fluctuating behaviours. To achieve the expected profits at the minimum conditional value at risk (CVaR),the power plant needs to make an appropriate planning for coal inventory.Given this background and based on the risk management theory developed in the financial industry,a nonlinear programming mean-CVaR model for optimizing the coal inventory is developed by taking the CVaR as a risk measuring index.In the developed optimization model,the objective is to minimize the CVaR under a given expected profit,and some uncertainties are taken into account including the coal price,electricity price,coal consumption,the execution rates of the contracted coal and the coal purchased from the coal market.Then,the developed optimization model is transformed into a linear programming problem so as to improve the solving efficiency.Finally,a sample example is employed to demonstrate the feasibility and efficiency of the model and algorithm developed.