Research Of Distribution Network Low Voltage Govern Model Based On In-Memory Computing

The Low voltage is an important index of electrical power quality.T he causes of low voltage in distribution network are complicated,and th e govern methods are varied.It is an urgent problem that how to accur ately judge the cause of low voltage and select the reasonable treatment method.In this paper,firstly we study the causes of low voltage,and set u p a low voltage genetic diagnosis model of distribution network based o n bisecting k-means algorithm and distributed echo state network algorith m.Then,we process the distribution network low-voltage users and distr ibution transformer data,and implement the distributed echo state networ k algorithm which is not given by the Mllib Spark machine learning lib rary.On this basis,this paper studies the evaluation model of low volta ge govern process.For the stage before cons truction and after constructi on,this paper separately designed and implemented the low voltage gove rn scheme pre-evaluation model which is based on the data envelopment analysis and the low voltage govern effect evaluation model which is ba sed on the multiple attribute comprehensive evaluation method.After tha t,the low voltage govern model is formed.The model covers the wholeprocesses of low voltage govern,from low voltage causes,to the proje ct management plan optimization,and finally to project post evaluation.Low voltage govern model has been used in Quzhou,Zhejiang.The results show that compared with the traditional way of decision-making,the use of the model can save the low voltage causes analysis time,im prove the level of decision-making,enhance the investment benefit of lo w voltage,so it has high research signific ance and value.