An Improved BP Neural Network Algorithm For The Application Of The Central Heating System

In Recent years, the central heating system is a universal application, this is because thecentral heating to improve energy efficiency, energy conservation. Therefore, this articleimproved BP neural network algorithm and in accordance with the characteristics of the BPalgorithm give a centralized heating applications.Firstly, improve the learning rate of the BP algorithm, the learning rate can changeaccording to the error function of spontaneous adjustment, thereby making the learning ratewithout losing stability in every step of the BP algorithm is able to maintain a relatively largevalue; then in order to overcome the lack of features of the robustness of the error function,using the strategy of "outliers" in the sample is ignored, the same time, the basis for thevariance of the error and the average to get the error estimates, to accelerate the convergencespeed of BP algorithm; last ultimately fall into local minimum in order to avoid the BPalgorithm, to amend the error function, the criterion is the BP algorithm is stable aftermultiple learning convergence in a relatively small threshold.This article is to study the various improvements of the BP algorithm, the integrateddesign algorithm implementation program. And basis for the algorithm to achieve the resultswere compared with the traditional BP algorithm, prove that the improved algorithm hascertain advantages in accuracy.This article based on improved BP neural network algorithm designed and implementeda central heating applications. The system first input of the input data, as well as idealexpectations; Then, the input data, after learning of BP neural network training process, thefinal heating program can be a relatively close to expectations; the end, will be Theapplication of the results output. It design a Central heating application systems and give arelatively user-friendly interface.Based on improved BP neural network algorithm to design the central heating systemand tested to prove that the application can be an ideal heating solution. After several roundsof comparative analysis, prove to have certain advantages in comparison with a similar centralheating system accuracy.Finally, this paper analyzes the advantages and disadvantages of central heatingapplication systems based on improved BP neural network algorithm, the shortcomings and make the outlook will be refined in subsequent work.