基于模糊神经网络的供热负荷预测的研究

发布时间：2018-01-04 05:03

本文关键词：基于模糊神经网络的供热负荷预测的研究　出处：《青岛理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着我国经济的发展，供热面积也在不断地扩大，供热的能耗也不断地增加。在我国三北地区，仅供热能耗这一项就占其总能耗的27.2%。此外，我国的能源的不断耗竭以及自动控制技术的不断发展，使得在供热管网方面的投资也逐渐增加，供热管网研究的重要性也在日益彰显。由于集中供热管网系统庞大性以及复杂性，精确的集中供热系统热负荷预测值对运行效率的提高，以及环保和节能等方面都具有非常重要的意义。本论文首先在对供热负荷预测的特点和现状总结的基础上，分析了可能影响负荷的主要因素，对影响因素采用模糊量化的方式进行研究处理，进而提出在供热负荷预测中应用模糊神经网络系统。本系统的设计核心是BP神经网络，即将模糊量化后的影响因素作为系统的输入值，去调整神经网络的权值，从而得到预测的网络模型。确定预测模型的运行参数后，运用MATLAB7.0进行预测仿真，，仿真结果表明，预测的结果能够满足要求，相对误差在合理的范围内，并且模糊神经网络算法比单纯神经网络算法具有更好的预测精度，泛化能力更强，从而能更好的提高了供热品质，节约能源。最后，在实现热负荷预测的基础上，对集中供热预测控制进行研究。从数学模型的角度来看，由于供热系统的非线性的特点，使得控制复杂性增加，本文对预测控制的必要性以及控制策略进行了探讨。
[Abstract]:With the development of our country ' s economy , the heating area is expanding constantly , and the energy consumption of heat supply is increasing continuously . In our country ' s three northern areas , only one item of heating energy consumption accounts for 27 . 2 % of its total energy consumption . In addition , our country ' s energy consumption and the continuous development of automatic control technology make the investment in the heating pipe network gradually increase . On the basis of summarizing the characteristics and present situation of heating load forecasting , this paper analyzes the main factors which may affect the load , carries out the research and treatment in the way of fuzzy quantification on the influencing factors , and puts forward the application of the fuzzy neural network system in the forecasting of heating load . The design core of the system is the BP neural network , and the influence factors after fuzzy quantization are used as input values of the system to adjust the weights of the neural networks to obtain the predicted network model . The simulation results show that the predicted results can meet the requirement and the relative error is within a reasonable range , and the fuzzy neural network algorithm has better prediction accuracy and generalization ability than the simple neural network algorithm , which can improve the heat supply quality and save energy . Finally , on the basis of realizing the thermal load forecasting , the central heating prediction control is studied . From the point of view of the mathematical model , the control complexity is increased due to the nonlinear characteristics of the heating system , and the necessity and control strategy of the predictive control are discussed .

【学位授予单位】：青岛理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183;TU995

【参考文献】