Research On Wireless Electric Signal Acquisition And Strength Prediction Of Joints In Arc Welding Process

Welding monitoring refers to the collection and transmission of sound,light,electricity and other physical information during the welding process through the relevant physical sensors.By monitoring the welding signal in the welding process,welding problems can be found and data basis can be provided for improving the welding procedure specification.In this paper,a data monitoring system of welding workshop based on wireless sensor network is proposed,which realizes the low-speed collection and wireless transmission of welding parameters.And an artificial neural network model is built based on the monitoring data to predict the mechanical properties of the welded joint.The main work is as the following:First of all,the program of the lower computer is compiled to realize the collection of the monitoring data and the summary in the coordinator.KY41 Hall current sensor and CHV-25P/50 Hall voltage sensor are used to collect welding voltage and current.In CC2530 chip,the lower computer program is written to process the data collected by the sensor.The output of Hall voltage sensor and Hall current sensor is 0-5V and 0-3v respectively.In order to get the actual welding voltage and current value,A / D conversion is needed to convert the analog signal into digital signal.This system uses the ADC function of CC2530 chip to collect the analog output of the sensor through the specific IO port,and completes the signal conversion.After the acquisition and conversion of welding parameter signals,ZigBee wireless sensor network is constructed with CC2530 as the network node,and the monitoring data is sent to the coordinator through ZigBee network for summary.Secondly,the upper computer interface program and welding datase are designed,and the upper computer interface is programmed with C# programming language,which realizes the data analysis and visualization function.The upper computer communicates with the coordinator through the serial port,and the lower computer coordinator summarizes the monitoring field detection data and sends it through the serial port.The upper computer analyzes the data sent by the lower computer through the serial port programming technology,and realizes the real-time digital and graphical display of the monitoring data.The system database is established by using My SQL database software,which can store and manage welding process parameters,welding monitoring data,welding quality,alarm information and system personnel information.The connection between C# and database is established by mysql.data.dll,and the upper computer calls the database.At the same time,XML configuration document is designed in the monitoring system.Users can configure key data such as database,sensor and network information by modifying XML related parameters,which enhances the flexibility of the system.Thirdly,the aluminum alloy MIG welding test was carried out to test the monitoring system Firstly,the accuracy of the monitoring data of the monitoring system is verified.The data monitoring of the aluminum alloy welding process is carried out by using the wired communication module and the monitoring system.The data monitoring function,data management function and user management module of the system are tested.Compared with the wired monitoring module,the system can realize the monitoring of welding parameters with higher sampling frequency,accurately reflect the different welding process,and realize the network function of monitoring data under the local area network,which meets the needs of wireless remote monitoring of welding process.Then,the welding parameters of aluminum alloy MIG welding are designed,and the monitoring system is used to monitor and collect the welding process.In the end,the yield and tensile strength of the welded sample were obtained by tensile test,which laid a data foundation for the construction of the mechanical property prediction model of the aluminum alloy MIG welding joint.The mechanical properties prediction model of aluminum alloy welding head based on artificial neural network is designed.Based on the data collected from welding test,30 sets of training data and 6 sets of testing data are used to train the model respectively by using monitoring data and preset parameter information.The training results show that,at the same time of network structure,the model trained by monitoring data has faster convergence speed,smaller prediction error and higher network accuracy than the model trained by preset parameters.It is better to train and predict the mechanical properties of welded joints by using the monitoring data.