基于遗传改进神经网络的煤矿井下传感器非线性校正方法  

A nonlinear correction method for coal mine underground sensors based on genetic improved neural network

作　　者：王凯雄 WANG Kaixiong(CHN Energy,Zhuneng Group Coal Preparation Plant,Inner Mongolia Ordos 010300,China)

机构地区：[1]国家能源集团准能集团选煤厂,内蒙古鄂尔多斯010300

出　　处：《工业仪表与自动化装置》2024年第4期114-119,共6页Industrial Instrumentation & Automation

摘　　要：煤矿井下的工作环境十分恶劣,包括高湿度、高温、持续振动和大量粉尘等因素,这些因素都可能对传感器造成非线性干扰,使其输出结果与实际值产生显著偏差。这种偏差如果不加以校正,会严重影响煤矿的生产安全和效率。为此,该研究提出一种基于遗传算法优化的神经网络方法,旨在校正煤矿井下传感器非线性误差。根据已获取的传感器非线性误差数据,确定校正模型的输入与输出。以神经网络为核心构建校正模型,并利用遗传算法对校正模型的权值和阈值参数进行优化,以提高模型的校正性能。在模型训练完成后,将其应用于实际煤矿井下传感器非线性校正中。实验结果表明,经过该研究方法校正后,瓦斯浓度值更接近标准瓦斯浓度值,误差显著降低,这充分证明了该研究方法在实际应用中的有效性和可靠性。The working environment underground in coal mines is extremely harsh,including high humidity,high temperature,continuous vibration,and a large amount of dust.These factors may cause nonlinear interference to sensors,resulting in significant deviations between their output results and actual values.If this deviation is not corrected,it will seriously affect the production safety and efficiency of coal mines.Therefore,this study proposes a neural network method based on genetic algorithm optimization,aiming to correct the nonlinear error of underground sensors in coal mines.Determine the input and output of the calibration model based on the obtained sensor nonlinear error data.Construct a calibration model with neural networks as the core,and use genetic algorithms to optimize the weight and threshold parameters of the calibration model to improve its calibration performance.After the model training is completed,it is applied to the nonlinear correction of sensors in actual coal mine underground.The experimental results show that after calibration by this research method,the gas concentration value is closer to the standard gas concentration value,and the error is significantly reduced.This fully demonstrates the effectiveness and reliability of this research method in practical applications.

关 键 词：遗传算法 神经网络 煤矿井下传感器 非线性校正方法 

分 类 号：TP13.66[自动化与计算机技术—控制理论与控制工程]