Design Of Piggery Environment Control System Based On Multi-Scale And Information Fusion

With the continuous improvement of China's breeding subsidy policy,there has been a marked increase in the number and range of livestock farms.However,there are many problems in the breeding industry,such as unscientific pig feeding mode,poor pig-house environment and waste of resources,Among these factors,the environment of pig house is very important to the healthy growth of pigs,the safety of pork and the outbreak of disease.In view of the uncertain change law of temperature and humidity in pigsties,the coupling of temperature and humidity control and the difficulty in monitoring environmental information in pigsties,This paper presents a piggery environment control system based on multi-scale information fusion.The system carries out multi-scale analysis on the historical temperature and humidity sequence of the pigsty to obtain the curve of temperature and humidity change.The temperature and humidity value in the pigsty at the next moment is predicted by the support vector machine regression.The predicted value is sent to the neural network controller,and then the fan and wet curtain are controlled to make the environment of the pig house suitable for the growth of pigs.At the same time,CFD software is used to simulate the temperature field and airflow field of the pig house,and the simulation results are fed back to the controller.The main work of this paper consists of the following five parts:(1)Overall design of the system scheme.According to the requirements of an animal husbandry company,the overall scheme of the system is designed,which is divided into the on-site environmental information collection part of the pigsty,the environmental control part,and the upper computer display part of the monitoring and control room.Each pig house contains seven environmental acquisition nodes,There are 14 control objects,including frequency converter,contactor and solenoid valve,multithreading technology is needed for environment information collection and transmission.Therefore,STM32 is selected as the microcontroller of this system.According to the characteristics oflarge communication data between microcontroller of 20 pig houses and monitoring room,Ethernet was selected for communication.(2)Multi-scale analysis of historical data and controller design.Due to outside weather conditions,the position of piggery,fan state will have an effect on the pigpen temperature humidity,so the piggery curves of temperature and humidity is difficult to determine,wavelet transform itself has the advantages of noise reduction,multi-scale analysis and time-frequency localization.Therefore,this paper uses the multi-scale property of wavelet transform to analyze the historical data of temperature and humidity of piggery,and obtains the change law of temperature and humidity.Then the regression prediction of temperature and humidity at the next moment of piggery is carried out and the results are sent to the neural network controller for environmental control.Step response is used to identify the environmental control system of the pig house,and then PID controller of neural network is designed.Decoupling device is designed for the coupling phenomenon of temperature and humidity control.Finally,the pigsty environment control system was built on Simulink to verify the feasibility of the system.(3)System hardware design.according to the actual environment of the pigsty and the requirements of this system,sensors are arranged.Two temperature and humidity sensors and one wind speed sensor are arranged at the wet curtain,two wind speed meters are arranged at the fan outlet,and one ammonia gas and one temperature and humidity sensor are arranged in the middle of the pigsty.According to the change range of temperature and humidity of the pig house-10?~50? and 20%RH~85%RH,and the accuracy of temperature and humidity measurement is within ±1.0? and ±5%RH.This system chooses the temperature and humidity sensor model STH10,and design its interface circuit.According to the resolution of wind speed acquisition 0.1m /s,measurement accuracy ±3% m/s and effective wind speed measurement range from 0 to 30 m/s,the model jl-fs2 of wind speed sensor was selected.According to the characteristics of large space,damp and inconvenient wired connection,ZigBee is selected to build the environmental information collection network in the pig house.(4)System software design.According to CFD simulation,boundaryconditions such as temperature and wind speed at wet curtain,wind speed at fan,thermal conductivity of pig body and convective heat transfer coefficient are required to collect environmental information of pig house.CFD technology is used to display the temperature field and air flow field of the pigsty in pictures,and the success of CFD pigsty modeling is verified by the simulation value and the measured value.The wind speed,temperature and humidity,ammonia concentration collection program is designed on the micro-controller.In order to make it easy for staff to monitor the pigsty environment,Delphi is used to design the upper computer interface,and SQL Server database is developed in order to query the historical data of pigsty.(5)The test.Two identical pigsties were selected in an animal husbandry company,and one of them was controlled by this system.The function of this system was tested by comparing the temperature and humidity data of the whole day.Results show that this system can stabilize the pigpen temperature around19 ?,humidity stable at about 60% RH.Temperature measured value and the maximum value of absolute error is 0.65 ?,maximum relative error is 3.4%,humidity measurement value and the value of the maximum absolute error is3.94% RH,maximum relative error is 6.6%,conform to the requirements of the breed.