基于多目标算法的企业负荷优化及变压器最优控制

发布时间：2018-07-02 21:08

本文选题：负荷优化分配 + 全局人工鱼群算法　；参考：《太原理工大学》2017年硕士论文

【摘要】：钢铁企业作为电力行业中最为典型的用电大户,其电能的消耗在生产成本中占有很大比重。随着经济下滑,钢铁企业负荷量严重缩水,因此导致企业配电变压器运行损耗剧增。本课题针对首钢长治钢铁有限公司中变压器普遍处于非经济运行状态的问题,通过结合该企业实际负荷情况构建负荷预测模型,并对长钢站下属钢南、钢西、钢北三站的负荷进行重新分配,之后根据负荷波动的情况提出变压器经济运行控制策略,用以实现变压器最优运行并达到降低该企业变压器运行损耗的目的。主要包括以下几个方面:(1)本文总结和分析现有负荷预测方法,比较传统最小二乘支持向量机与神经网络在负荷预测应用中的优劣,重点分析采用最小二乘支持向量机进行负荷预测的不足以及现有最优最小二乘支持向量机模型的特点,提出将对于二维空间寻优具有较高精度的全局人工鱼群算法与最小二乘支持向量机相结合,构建新的负荷预测模型,并结合该企业长钢站的负荷验证其预测模型的准确性。最后对长钢站未来五年负荷进行预测,该预测值同时为后续负荷分配提供了新的思路。(2)本文考察了当前首钢长治钢铁企业负荷现状,对变压器经济负载系数以及经济运行区间进行分析。考虑到由于负荷的地理因素在分配过程中将产生新架设线路的情况,这里通过考察各线路负荷情况,依据最新山西国家电网架空线以及杆塔选择条件对每条线路的架空线类型以及杆塔数量进行选择,并将其费用计入负荷分配优化过程中。为了同时满足变压器经济运行和负荷分配费用最低这两个分配因素,提出采用多目标NASG-II算法对钢南、钢北、钢西站负荷进行分配,得到Parteo前沿最优解集并对该解集进行深入分析,结果显示负荷分配后该站变压器的运行损耗明显减少。最后依据该站后五年负荷预测结果,为该企业制定一套切合实际的负荷分配方案。(3)当前钢铁企业为了保证自身运行的可靠性,企业中变电站长期保持两台变压器并列运行状态。然而钢铁企业负荷具有较大的波动性,其运行过程中难免出现非经济运行的状态。针对此问题,本文通过分析变压器经济运行方式,得出变压器在不同运行方式下的节电效益,基于全局人工鱼群算法确定变压器最优投切策略。考虑传统全局人工鱼群算法在寻优过程中可行域是连续的,然而实际上本文所研究问题的可行域是离散的,提出改进的全局人工鱼群算法用以实现在离散空间中确定变压器最优投切策略的目的。最后将该方法用于钢东站变压器运行中,验证了所提方法的有效性。
[Abstract]:Iron and steel enterprises as the most typical power users in the power industry, its electricity consumption in the production costs account for a large proportion. With the economic downturn, the load of iron and steel enterprises shrinks severely, resulting in a sharp increase in the operation loss of distribution transformers. Aiming at the problem that transformers in Changzhi Iron and Steel Co., Ltd of Shougang are generally in non-economical running state, the paper constructs a load forecasting model combining with the actual load situation of this enterprise, and makes a load forecasting model for the south and west of Changgang Station. According to the load fluctuation, the economic operation control strategy of transformer is put forward in order to realize the optimal operation of transformer and reduce the operation loss of transformer in this enterprise. The main contents are as follows: (1) this paper summarizes and analyzes the existing load forecasting methods and compares the advantages and disadvantages of traditional least squares support vector machines and neural networks in load forecasting applications. The deficiency of least square support vector machine (LS-SVM) for load forecasting and the characteristics of the existing optimal LS-SVM model are analyzed. A new load forecasting model is constructed by combining the global artificial fish swarm algorithm with the least square support vector machine, which has high precision for two-dimensional spatial optimization, and the accuracy of the forecasting model is verified by the load of the enterprise's Changgang station. Finally, the load of Changzhi Iron and Steel Station in the next five years is forecasted, which also provides a new idea for the subsequent load distribution. (2) this paper reviews the current load situation of Changzhi Iron and Steel Enterprise in Shougang. The economic load coefficient and economic operation interval of transformer are analyzed. Taking into account the fact that the geographical factors of the load will result in new lines being erected during the distribution process, here through an examination of the load on each line, According to the latest selection conditions of overhead lines and towers of Shanxi State Grid, the types of overhead lines and the number of towers of each line are selected, and the cost is taken into account in the optimization process of load distribution. In order to satisfy the two distribution factors of transformer economic operation and lowest load distribution cost simultaneously, a multi-objective NASG-II algorithm is proposed to distribute the load of steel south, north and west stations. The optimal solution set of Parteo front is obtained and analyzed in depth. The results show that the operation loss of transformer in this station is obviously reduced after load distribution. Finally, according to the result of load forecasting in the later five years of the station, a suit of practical load distribution scheme is worked out for the enterprise. (3) in order to ensure the reliability of its own operation, the substation in the enterprise keeps two transformers running side by side for a long time. However, the load of iron and steel enterprises is fluctuating, and it is inevitable that non-economic operation occurs in the process of operation. In order to solve this problem, by analyzing the economic operation mode of transformer, this paper obtains the power saving benefit of transformer under different operation modes, and determines the optimal switching strategy of transformer based on global artificial fish swarm algorithm. Considering that the feasible region of traditional global artificial fish swarm algorithm is continuous in the process of optimization, however, in fact, the feasible region of the problem studied in this paper is discrete. An improved global artificial fish swarm algorithm is proposed to determine the optimal switching strategy of transformer in discrete space. Finally, the method is applied to the transformer operation of East Steel Station, and the validity of the proposed method is verified.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM41

【参考文献】