随机鲁棒区间-生物质电厂选址风险分析模型

优化生物质发电厂厂址对于中国生物质发电行业意义重大。然而，生物质发电厂选址系统中包含多重不确定性和复杂性。忽略这些，将给生物质发电厂带来风险。因此，充分考虑生物质发电厂系统中的多重不确定性，将鲁棒随机规划（robust optimization，RO）与区间规划（interval parameter programming，IPP）融于两阶段规划（two-stage programming，TSP）框架中，建立基于随机鲁棒区间风险分析模型（stochastic robust interval model， SRIM）的生物质发电厂选址模型。该模型可以处理表现为离散区间和随机性的不确定变量。并且对于随机过程产生的风险进行追索，增强生物质发电系统的安全性。通过调节不同风险等级，可以对系统进行风险分析，利于决策者对系统安全性和经济性做出衡量。该文以装机容量为15 MW的生物质发电厂为案例。模型结果显示：该规划区域拟建设生物质发电厂数量为1，优化厂址介于（245,242）km至（250,247）km；各个燃料收储站优化配送方案；以及不同鲁棒等级下的系统风险和系统成本。通过模型得出的结果合理可行，可以为生物质电长选址提供科学的依据及决策支持。

Model of risk analysis on site selection of biomass power plant based on stochastic robust interval method

Abstract: The transport cost of biomass fuels accounts for a large proportion of the total cost of the operation of biomass power plant. Optimizing biomass power plant site can largely mitigate the transport cost and reduce the pollutant emissions from the transportation process of biomass fuels. Therefore, it is significant to optimize the biomass power plant sit. However, the biomass power system contains many uncertainties, because that many parameters can hardly be acquired as deterministic values but expressed as interval and/or stochastic formats. For example, the supply demand of biomass fuels can be expressed as probability distributions; also, interval values can describe the uncertain parameters such as the biomass fuels price, which fluctuates between lower and upper bounds. Energy systems would become insecurity and with a high risk without considering these uncertainties. Security is a priority in the operation of biomass power plant. In this study, a stochastic robust interval model (SRIM) was developed for the biomass power plant site selection under uncertainties, through incorporating interval-parameter programming (IPP) and robust optimization (RO) within two-stage programming (TSP) framework. In SRIM, decision variables were divided into two subsets: those that must be determined before the realizations of random variables were known, and those that were determined after the realized random variables were available. The SRIM can deal with the uncertainties described in the terms of the interval values and probability distributions, moreover, it can also reflect economic penalties as corrective measures or recourse against any infeasibilities arising due to a particular realization of an uncertain event. In the SRIM modeling formulation, penalties were exercised with the recourse against any infeasibility, and robustness measures were introduced to examine the variability of the second stage costs that were above the expected levels. The SRIM was generally suitable for risk-aversive planners under high-variability conditions. The SRIM method was applied to a hypothetical case of planning biomass power plant (with installed capacity of 15 MW) site selection with considering the uncertainties. A number of solutions under different robustness levels have been generated. The obtained results can help generate desired decision alternatives that will be able to enhance the safety of biomass power system with a low system-failure risk level and particularly useful for risk-aversive decision makers in handling high-variability conditions. The result are beneficial for managers analyzing the results to gain insights into the tradeoff between system's safety and economic, and analyzing the risk of the system. The results of SRIM shows: 1) The construction number of biomass power plant is one; 2) The optimum biomass power plant is from (245, 242) km to (250, 247) km; 3) The optimum allocation scheme for each fuel purchase and storage station; 4) The system costs under different robust levels; 5) The notion of risk in stochastic programming under different robust levels. The modeling results from the RISO can help generate desired decision alternatives that will be able to not only enhance the safety of planning biomass power plant site selection with a low system-failure risk level, but also mitigate pollutant emissions from the transportation process of biomass fuels. The results are reasonable, and could provide a reference for the selection of the biomass power plant site.