基于免疫遗传算法的多目标厂级负荷优化分配研究

With the gradual implementation of the "separation of power plant and grid, competitive bidding ", the development of smart grid and the needs of energy conservation, although new energy has been actively developed, the central pillar of the status of thermal power in the grid has not changed, at the same time, the production and operation scheduling of thermal power systems has been required more higher. So there is an urgent need that load scheduling of power plant takes multi-objective optimization including safety, rapidity, economy, environmental protection and other, instead of the previous simple single or dual objective optimization. Plant-level load optimization problem is a high dimensionality, non-convex, discrete and nonlinear optimization problem, which is difficult to find the optimal solution in theory. The common optimization algorithms have some advantages and disadvantages, and their results are not satisfactory with a single method. So this subject proposes immune genetic algorithm which combines the advantages of the immune algorithm and genetic algorithm. This algorithm not only gives full play to the widely accepted advantages of genetic algorithm, but also takes advantages of immune algorithm, including immune memory, immune vaccines, antibodies promotion and inhibition. Immune genetic algorithm effectively improves the prematureness, poor local search ability, slow evolution and reduction of antibodies diversity of genetic algorithm. So this subject conforms to the theoretical and practical need, has important theorWith the gradual implementation of the "separation of power plant and grid, Competitive bidding ", the development of smart grid and the needs of energy conservation, the production and operation scheduling of thermal power systems has been required more higher. So there is an urgent need that load scheduling of power plant takes multi-objective optimization including safety, rapidity, economy, environmental protection and other. Plant-level load optimization problem is a high dimensionality, non-convex, discrete and nonlinear optimization problem, which is difficult to find the optimal solution in theory. The common optimization algorithms have some advantages and disadvantages, and their results are not satisfactory. So this subject proposes immune genetic algorithm which combines the advantages of the immune algorithm and genetic algorithm. This algorithm not only gives full play to the widely accepted advantages of genetic algorithm, but also uses advantages of immune algorithm, including immune memory, vaccines, antibodies promotion and inhibition. Immune genetic algorithm effectively improves the prematureness, poor local search ability, slow evolution and reduction of antibodies diversity of genetic algorithm. So this subject conforms to the theoretical and practical need, has important theoretical and practical significance, and solves the load optimization problem of plant-level under the new situation.

随着"厂网分开，竞价上网"的逐步实施，随着智能电网的发展，节能减排的需要，虽然新能源得到大力发展，但火力发电在电网中的核心支柱地位没变，同时对火力发电系统的生产、运行调度提出了新的更高要求，迫切需要厂级负荷优化调度是以安全性、快速性、经济性、环保性等多目标的优化而非从前的简单的单、双目标优化。厂级负荷优化问题是一个高位数、非凸的、离散的、非线性优化问题，很难找出理论上的最优解，用于厂级负荷优化比较有影响的算法各有利弊，用单一的方法解决起来效果不理想，本课题提出的免疫遗传算法融合了免疫算法和遗传算法的优点，充分发挥遗传算法被广泛接受、使用的优点，同时利用免疫算法具有免疫记忆、免疫疫苗、抗体促进抑制机制的优点有效改进遗传算法存在早熟、局部搜索能力差、进化缓慢、抗体多样保持欠缺等问题，因此本课题的研究顺应了理论和实际的需要，具有重要的理论意义和实际意义，解决了新形式下厂级负荷优化的难题。

随着“厂网分开，竞价上网”的逐步实施，随着智能电网的发展，节能减排的需要，虽然新能源得到大力发展，但火力发电在电网中的核心支柱地位没变，同时对火力发电系统的生产、运行调度提出了新的更高要求、迫切需要厂级负荷优化调度是以安全性、快速性、经济性、环保性等多目标的优化而非从前的简单的单、双目标优化。厂级负荷优化问题是一个高维数、非凸的、离散的非线性优化问题，很难找出理论上的最优解，用于厂级负荷优化比较有影响的算法各有利弊，用单一的方法解决起来效果不理想，本课题提出的免疫遗传算法融合了免疫算法和遗传算法的优点，充分发挥遗传算法被广泛接受和使用的优点，同时利用免疫算法具有免疫记忆、免疫疫苗、抗体促进抑制机制的优点有效改进遗传算法存在早熟、局部搜索能力差、进化缓慢、抗体多样保持欠缺等问题。因此基于免疫遗传算法的多目标厂级负荷优化问题的研究，不但顺应了算法发展的需要，也顺应了电力企业和国家可持续发展、节能降耗、减排的需要，具有重要的理论意义与实际应用价值。.课题组完成了厂级负荷优化分配基于安全性、经济性、快速性、环保性子目标的建模、总优化目标函数的建模，综合安全性和机组的实际情况合理考虑、选择了不同运行阶段的机组的众多约束，合理设计免疫遗传算法三个方面的工作，在算法设计上，采用理论证明以及利用matlab仿真软件进行了大量的函数验证仿真实验，有效证实算法具有很好的快速收敛性，最后结合实际电厂的的4台机组，利用matlab仿真软件实现基于安全性、经济性、快速性、环保性等多目标的优化，并与NSGA II优化结果以及电厂的出厂参数进行对比，都有效证实本课题在安全的基础上，基于免疫遗传算法的厂级负荷优化分配具有更好的经济性、快速性和环保性，圆满完成了课题研究计划的任务。与此同时，申请了国家专利1项（已由国家知识产权局受理）以及发表了11篇学术论文的丰厚研究成果，在此课题的支持下培养了8名硕士研究生。