Development Of Fan Centralized Control System With Predictive Maintenance

Under the background of "Made in China 2025" and "Industry 4.0",the livestock and poultry industry is gradually developing in a large-scale and intensive direction.In recent years,the rise of building breeding is to concentrate livestock and poultry and various breeding equipment in the building,which improves the breeding density and reduces the breeding cost.The increase in breeding density has put forward higher requirements for the stability and reliability of equipment work,while the awareness of equipment use and maintenance is still in the traditional way: regular manual inspection is carried out to check the daily running state of equipment,and only when the equipment is in serious failure,the shutdown maintenance is not realized predictive maintenance.Because of the actual needs of current building breeding and the problems existing in equipment maintenance,this paper takes the ventilation fan(hereinafter referred to as the fan),the key equipment of the breeding farm,as the research object,and carries out the research and design of the fan centralized control system with predictive maintenance function.The main research contents are as follows:(1)To reduce system maintenance costs and realize more flexible data acquisition,a master-many-slave data acquisition device is designed.The slave machine is responsible for collecting the status data of the fan,while the host machine summarizes the data of several slave machines.The design of one master and many slaves avoids the complex system wiring,and the data acquisition function of the bottom layer is stripped to the slave machine,thus reducing the system coupling,reducing the system maintenance cost,and increasing the system flexibility.The core of the hardware device is an embedded microcontroller,which consists of a fan vibration signal acquisition module,a fan current acquisition module,and a data communication module.(2)To facilitate the staff to view the system data and remotely control the fan start and stop,this paper completed the design of the upper computer software based on Qt.The software can display the current status of the equipment in the system,support RS485 wired serial communication and 4G wireless communication,can remotely control the fan start and stop,and has the function of historical data recording.To ensure the stability and reliability of the communication between the master and slave computer and the communication between the host computer,this paper designs the related communication protocols.(3)To realize the predictive maintenance function of fans,a fan fault warning method based on WDCNN-LSTM is proposed and the embedded model is completed.In this paper,the one-dimensional data feature extraction capability of WDCNN(the first layer wide convolutional kernel deep convolutional neural network)is combined with the sequence feature extraction capability of LSTM(short and long time memory neural network)to generate the WDCNn-LSTM network model,which reduces the complexity of each model,avoids the stacking of network layers,and reduces the total number of parameters of the model.To complete the embedded model,this paper converts the format of the model parameter file into C language.(4)Build a system test environment and verify system functions.Electromagnetic interference experiments,high and low-temperature experiments,and humidity experiments are carried out on the hardware part of the system to verify the reliability and stability of hardware functions.Meanwhile,the implementation logic of software such as system data communication,state evaluation,and equipment control is tested to ensure the normal operation of software functions.The test shows that the fan centralized control system with a predictive maintenance function can meet the relevant design requirements.