CAJ | 학술논문

针对粮食颗粒管道气力输送系统中风速、压损、料气比等参数之间存在复杂非线性关系，难以建立准确的数学模型以实现闭环控制的问题，以CXLD50吸压混合输送移动式吸粮机为研究平台，设计了用BP神经网络的物料流量预测为反馈环节的模糊控制系统：使用神经网络工具建立物料流量预测模型，以快速方便地进行两相流流量在线测量；模型输出的预测流量与期望流量进行比较后，输入到模糊控制器进行判断推理并输出．仿真结果表明，系统响应迅速，可在50 s内达到理想输出；抗干扰能力强，其误差量稳定在±0．5 kg／s左右，有效改善了离线测量方法的信号反馈滞后现象，提高了输送系统的稳定性．
침대양식과립관도기력수송계통중풍속、압손、료기비등삼수지간존재복잡비선성관계，난이건립준학적수학모형이실현폐배공제적문제，이CXLD50흡압혼합수송이동식흡량궤위연구평태，설계료용BP신경망락적물료류량예측위반궤배절적모호공제계통：사용신경망락공구건립물료류량예측모형，이쾌속방편지진행량상류류량재선측량；모형수출적예측류량여기망류량진행비교후，수입도모호공제기진행판단추리병수출．방진결과표명，계통향응신속，가재50 s내체도이상수출；항간우능력강，기오차량은정재±0．5 kg／s좌우，유효개선료리선측량방법적신호반궤체후현상，제고료수송계통적은정성．
It is difficult to establishe accurate mathematical model to realize the closed-loop control,because of the complex nonlinear relationship among the wind speed,pressure loss,feed-gas ratio and other parame-ters in grain partieles pneumatic transport.To solve this problem,with CXLD50 suction pressure mixed conve-ying mobile grain sucking machine as the research platform,fuzzy control strategy was put forward with the material flow prediction of BP neural network as feedback loop.The system uses the material flow prediction model based on neural network tools which could quickly and conveniently online measured two-phase flow. After comparing the model output flow of prediction and expectation,it was input to the fuzzy controller for judgement and output.Simulation showed that the system has rapid response,achieving ideal output in 50 s and strong anti-interference ability,keeping deviation stable in ±0.5 kg/s so that the fuzzy control system could improve the signal of off-line measurement feedback lag and raise the stability of transport system.