【作者】 夏澍；

【导师】 周明；

【作者基本信息】 华北电力大学 ， 电力系统及其自动化， 2014， 博士

【摘要】 发电调度是保证电网安全稳定运行,实现发电资源高效配置的重要环节。近年来,为了促进能源、环境可持续发展,风电作为一种成熟的可再生发电资源,得到了大力开发,风电场规模不断扩大,并网容量不断增加。然而风电具有随机性、间歇性和难以准确预测等特点,这些特点给电力系统发电调度和安全运行带来了一系列新的问题。同时,随着发电调度系统管理水平的提高,为了实现综合效益最大化,不仅应该考虑有功出力的分配,还应该考虑机组备用的选择、机组检修计划优化等问题,统筹协调发电资源。基于上述背景,本文围绕风电接入电力系统的中短期发电调度问题开展了以下研究：在间隔10-15min有功优化调度方面,为了应对风电功率波动对备用容量和线路潮流带来的不确定性问题,提出了考虑线路安全校核的含风电场电力系统有功和备用协调调度模型。首先,在建模过程中结合置信水平,分析了负荷和风电出力预测误差对备用的影响,建立了备用与风电出力之间的量化关系。然后针对风电功率不确定性引起线路潮流波动的特点,引入线路安全约束的多场景模型,同时为了处理该模型存在校核场景和校核线路数量大的问题,进一步结合调度时段机组的出力范围,基于线性规划定理,提出校核线路和校核场景的缩减方法,降低了求解难度。最后进行了算例验证,结果显示,根据所建模型给出的调度方案能同时满足经济性和安全性。在上述研究基础上,进一步考虑环境效益和备用不足引起的运行风险,目标函数中增加了污染物排放量,定义了备用风险指标,构建了多目标有功优化调度模型。为了更好的求解该模型,提出了一种改进多目标微分进化算法。该算法引入混沌搜索策略以提高初始种群利用率,采用控制参数调整策略以克服算法对控制参数依赖性强的缺点,利用动态拥挤距离排序策略使得帕累托最优解集分布更加均匀。以所提算法求得的帕累托最优解集为决策矩阵,使用基于熵的序数偏好方法对最优解集进行排序,得到最终决策方案。对比分析结果表明,所提算法能够为多目标有功优化调度问题提供更加优良的候选方案。在日前机组组合方面,为了更高效地求解含风电场电力系统安全约束机组组合的多场景随机模型,提出了一种区间优化结合点估计的方法。首先基于风电出力置信区间,利用最紧约束集及相应的极限场景描述风电出力波动,进而对随机安全约束机组组合进行区间优化,在保证调度方案满足安全约束的同时提高了计算速度。然后为了评价应对风电出力波动的调整能耗,进一步利用点估计法对区间优化方法进行改进,使得优化调度方案更为经济。最后采用混合整数线性规划法对模型进行求解。通过算例对比分析,结果表明,所提方法仅需要对少量场景进行计算就可以保证调度方案满足安全性和经济性的要求,能够有效地求解含风电场的随机安全约束机组组合问题。在月度机组组合方面,为了更好地接纳风电,同时兼顾检修计划优化,建立了含风电场电力系统月度机组组合和检修计划联合优化模型。首先基于历史数据,采用威布尔函数描述风电的概率分布,利用机会约束规划理论处理风电不确定性对约束条件的影响,构建了相关概率模型并给出了多个风电场情况下概率模型确定化方法。然后引入了检修费用和检修相关约束,并针对机组组合和检修计划时段间隔不同,分别设置了时段索引变量,建立了相互之间的关联矩阵,实现了发电和检修协调优化。最后利用混合整数线性规划法对模型进行优化。算例结果表明,所建模型可以有效地权衡成本费用和运行风险之间的关系,能够同时给出合理的机组组合方案和检修计划方案,为风电接入电力系统的中期调度运行提供了一种有效的手段。上述论文工作研究了风电出力随机性对不同时间尺度发电调度的影响,并给出了一套建模和求解方法,为规模化波动性电源并网后电力系统中短期安全经济运行提供了理论和方法支撑。更多还原

【Abstract】 Generation scheduling is vital to ensure the security operation of power system and to rearlize the high efficient allocation of generation resources. In recent years, in order to promote energy and environmental sustainability, wind power, which is a mature renewable generation resource, has been vigorously developed. The wind farm scale has been expanded and the grid-connected wind power capacity has been increased. However, due to the randomness, intermittence and unpredictability of wind power, a series of problems are brought to security operation and generation dispatching in the power system. In addition, with the improvement of systematic management level for generation scheduling, all of the active output, reserve allocating and maintenance scheduling should be considered in the optimal scheduling model to coordination generation resources and maximize the overall benefit. Based on such backgrounds, this dissertation focuses on the medium-short term generation scheduling for power system with wind power integration. The main research work involves:In aspect of active optimal scheduling with10-15min interval, to deal with the uncertainties of reserve determination and line power flow calculation brought by wind power, the coordinated active power and reserve dispatching model for wind power integrated power systems considering line security verification is proposed. Firstly, in the modeling process, combining the confidence level, the influence of load and wind power forecast errors on the reserve is analyzed, the relationship of reserve and wind output power is established, and the demand of reserve is accurately quantified. Secondly, considering the fluctuant features of line power flow arouse by wind power uncertainty, the multi-scenario model of line power constraints is introduced. However, the multi-scenario model takes a large amount of computation. To improve the efficiency, the checked scenario reduction method and checked lines reduction strategy are developed, which are based on linear programming theorem, and consider the output range of units in scheduling period. Finally, numerical examples show that the dispatch schemes given by the proposed model are provided to be more feasible and safety with the satisfaction of economy.On the basis of the above research, a multi-objective optimization scheduling model is formulated. In the proposed model, the environmental benefits and the operational risk caused by lacking reserve are taken to account, pollutant emissions and reserve risk indicator are introduced to the objective function, and the reserve risk indicators are developed. To solve this model, an improved differential evolution algorithm for multiobjective optimization (IDEMO) is proposed. In the IDEMO, the chaotic searching strategy is introduced to improve the availability rate of initial population, the adjustment strategies for control parameters are used to strengthen global optimal searching capability, and dynamic crowding distance is employed to keep the diverity of population. After that, the decision matrix is given by pareto solution set of the IDEMO, and the optimal solution set is odered by similarity to ideal solution (TOPSIS) based on the entropy, so the optimal solution can be decided. Numerical examples show that the proposed algorithm is able to provide excellent candidate plans for multi-objective optimization scheduling problem.In aspect of day-ahead unit commitment, an interval optimization combined with point estimation method (IO-PEM) is proposed to solve stochastic security constrained unit commitment (SCUC) problem. Firstly, accordance with the confidence interval of the wind power, the fluctuation of wind power is described by the most compact constraints set and the corresponding extreme scenarios. And the stochastic security constrained unit commitment is solved by interval optimization method. Such method can accelerate the solution speed in the premise to ensure that the scheduling results meet security constraints. Secondly, in order to evaluate the adjusting energy consumption with the consideration of wind power output fluctuation, point estimation method is introduced to improve the interval optimization approach. As a result, the optimal scheduling plan can be more economical. Finally, the model is optimized by mix-integer linear programming. And simulation results show that, the proposed method only need to caculate a few scenarios for keeping security and economy of the dispatching scheme, so the proposed method can effectively solve the stochastic SCUC problem.In aspect of monthly unit commitment, the joint optimization model of monthly unit commitment and maintenance scheduling for wind power integrated power system is proposed to handle the randomness of wind power and optimize maintenance scheduling. Firstly, based on historical data, the stochastic behavior of wind power is represented by weibull distribution, and the impact of wind power uncertainty on constraints is coped with chance-constrained programming theory. The relating probability model is formulated, and the methods of probability model transferring to determined model with wind farms is proposed. Secondly, maintenance expenses and the constraints of maintenance are introduced. Since unit commitment and maintenance scheduling have different scheduling intervals, the index variables of them are setted respectively, incidence matrix is created, generation and maintenance scheduling are coordinative optimized as well. Finally, the proposed model is solved by mixed integer linear programming. Simulation results show that proposed model can effectively balance the relationship between the costs and risks, the results of unit commiment and maintenance scheduling are more reasonable, and the optimal scheduling can provide an effective mean of mid term generation scheduling for power systems with wind power.As previously mentioned, in this dissertation the influence of wind power randomness on generation scheduling under different time scales is studied, the relevant approaches are furnished, and this work can provide a theoretical basis and method support for medium-short term economic and security operation in the environment of large-scale wind power integration.更多还原