Research On On-line Monitoirnrg Analysis System Of Mine Hydrology Based On Probabilistic Neural Network

As one of the main energy sources that is urgently needed in theprocess of rapid development of industrialization in China, coal hasbecoming a mineral energy which is indispensable in people’s dailyproduction and life. The geology and hydrogeology condition of coalmine in our country is complicated, and water inrush accident happensfrequently in the process of coal mining. Type of coal mine water inrushis various, and the formation mechanism and control method is graduallycomplicated. Whether the area or the severity of damage under the threatof water disaster, it is rare in the world. Under the premise of normalsafety production of coal mine, to design and research a set of minehydrology monitoring system with the function of forecasting the waterinflow, which can master the mine hydrogeological information inreal-time has becoming a significant issue to solve in coal industry.On account of the current shortage in mine hydrology monitoringtechnology and technology that neural network is applicated in waterinflow prediction, this pepper designs a set of mine hydrological on-linemonitoring analysis system based on probabilistic neural network. Firstly, design system architecture, and research plans to adopt the pattern ofsubstation networking and master station centralized controlling tocomplete the real-time online monitoring of the mine hydrologicalinformation and remote automatic drainage control. STM32F103ZET6ofSTM32series is selected as the micro-controller of substation to designthe data collection and communication hardware circuit, and installmultiple parameter measurement sensors in water inrush prone areas ofthe coal mine to acquire hydrological information including the level,pressure, flow and temperature of mine water in water inrush sensitiveareas under the complex environment, such as coal mine main workingface, roadway, open channel and pipe in real-time. S7-200CPU226XPPLC of Siemens series is selected as the main controller of master station,and master station adopts the pattern of iterating through to realize datadistributed collection and running state centralized control of substationthrough the RS485bus, and then transmit information of each substationto ground monitoring center through Ethernet.Secondly, design the man-machine interface of host computer inground monitoring center using configuration software of KingView torealize real-time display and historical record of coal mine field datathrough animation display, trend curve analysis, report output, databasestorage and query, alarm processing and query, data printing and otherforms, and to realize several remote operations such as network distribution of mine hydrological data, system operation mode switching,equipment parameters modification and pump start-stop control throughDTU SMS module and WEB publishing technology.Lastly, establish the mine water inflow prediction model based onprobabilistic neural network on the basis of determining input parameters,training sample and network structure, and train and simulate model andanalyze results in MATLAB. Adopt OPC technology to realize dataexchange between the KingView and MATLAB software, and read coalmine hydrological data acquired from multiple parameter measurementsensors from the host computer in ground monitoring center, and thenimport thes data into MATLAB, adopting probabilistic neural networkalgorithm to classified predict water inrush risk, and then the groundmonitoring center transfers corresponding instruction to master stationafter identification and analysis to achieve the function of mine waterinrush risk early warning and alarming.The project practice and test results show that mine hydrologymonitoring system designed with high reliability and real-timeperformance, can realize real-time online monitoring of mine hydrologyand remote automatic drainage control. Water inflow forecast modelbased on probabilistic neural network in this paper can realize real-timeprediction of water inrush danger and achieve the desired accuracy, andsatisfies the design requirement of system.