Study On Operation Path Optimization Based On DCS Human-Machine Interface Feature Recognition

Fewer people on duty,reducing staff to increase efficiency is the goal of the development of power generation enterprises.One of the problems to be solved is to identify the DCS interface of the thermal power plant and the feasible path from the current screen to the operation target in order to control the DCS interface and simulate the specific operation of the operator through the software.First of all,interface recognition is the basis of the current man-machine interface positioning.Aiming at the composing features of the DCS man-machine interface of a power plant simulator,the image features are highlighted by image pre-processing,and the recognition region is determined by the method of feature region segmentation.The text features of the DCS man-machine interface are recognized based on BP neural network.For the recognized text features,through TF-IDF(Term Frequency-Inverse Document Frequency)algorithm different interface text feature words weighting,establish DCS inverse file frequency(IDF)text weight database,calculate the feature vector of the sample text,compare the text vectors of the detected image and the sample database image by the cosine similarity method,judge and identify the current test interface image according to the calculated value of the text similarity,and determine the current interface.Secondly,based on the in-depth study of the relationship between DCS man-machine interface recognition and interface in thermal power plant,the relationship modeling of image information layer and node location layer is completed.The static interface information on the image is better separated from the dynamic information of the actual screen switching,and enough prior information can be stored for the switching and operation of DCS nodes,which is convenient to generate the operation path.An interface description adjacency table is proposed to describe the information of the DCS interface in detail.Finally,aiming at the feasible path problem from the current picture to the operation target,an improved Dijkstra algorithm is proposed by establishing the node position directed graph and according to the characteristics of less vertex multi-path.The results show that the application of image recognition and path optimization technology can promote the intelligence and automation of DCS operating station system,improve the work efficiency of substation staff,and will have a great application prospect.