【摘要】 随着电能替代的不断深化,由"电采暖"引起的用电高峰时段电网负荷过高的问题逐步突显出来,能源系统用电负荷的预测显得尤为重要。同时,为了解决采用多阶段规划方法求解能源系统运行策略时耗时较多的问题,提出了基于负荷预测的园区供热系统运行优化技术。首先采用太阳辐射、室外温度与历史负荷作为输入参数的人工神经网络构建了热负荷预测模型;然后,通过将多阶段动态规划问题转化为线性规划问题,开发了供热系统运行优化方案快速求解方法;最后,使用该方法对北方某园区进行仿真验证,得到最优运行方案。结果显示,预测期内电采暖日运行费用可节约1.15万元,用电峰谷负荷差减小5 489 k W,所提出的优化方法可在短时间、较少的监测参数下得到可行的运行方案,适用于工程实际。更多还原

【Abstract】 With the deepening of the electrical energy substitution,the electric heating tends to highlight that the peak load could be too high during the peak period of electricity consumption,which makes it especially important of power system load forecasting. At the same time,in order to solve the problem of time consuming by using multi-stage planning method to solve the energy system operation strategy,this paper presents operation optimization technology of regional heating system for park based on load prediction. Firstly,we build a thermal load prediction model by using artificial neural network,which takes solar radiation,ambient temperature and historical load as input parameters. And then,we develop a quick solution method for engineering application face to regional heating system operation optimization,by converting multi-stage dynamic programming problem into linear programming problem. Finally,we validate this method through the simulation of a park in the North,to obtain the optimal running scheme. The results show that the daily operation cost of electric heating can save 11 500 yuan during forecast period,and the peak valley load difference can reduce 5 489 k W. The proposed optimization method can obtain the feasible running scheme in a short time and less monitoring parameters,which is suitable for engineering practice.更多还原