【作者】 马鑫；

【导师】 徐殿国； 张东来；

【作者基本信息】 哈尔滨工业大学 ， 电力电子与电力传动， 2009， 博士

【摘要】 地下综采工作面顶板支护是煤炭生产安全、产量和效率的重要问题,其安全性、可靠性及自动化程度直接影响煤炭生产的安全性及生产效率。液压支架电液控制系统是液压支架控制的核心,我国现使用的液压支架电液控制系统主要依赖国外进口,国外产品价格高、交货期长、备件昂贵不能完全适应我国煤矿的条件。因此,参考国外先进液压支架电液控制系统,并结合我国的煤矿条件实现液压支架电液控制系统国产化,可摆脱对国外产品的依赖、积累相关技术、提高我国采煤自动化水平、减少煤矿事故发生率,具有重要的社会效益和经济效益。本文在此技术背景下,研制了液压支架电液控制系统;为实现煤机联动综采自动化、故障快速定位及预测,提高煤炭生产安全性及生产效率,研究煤机定位及关键设备故障预测方法。分析了地下综采系统架构、采煤进程及国际先进综采工作面顶板支护系统液压支架电液控制系统。对液压支架电液控制系统的供电、组网、隔离方式、控制功能和控制过程,以及单台液压支架控制器与其配套设备的连接组成进行研究,为后续研发设计支架控制系统提供有效参考。研制了液压支架电液控制系统。液压支架电液控制系统由若干液压支架控制器串行连接组成,液压支架控制器包含主控制器及电磁阀驱动控制器。在硬件上提出上电时序化设计,通过确保电磁阀驱动控制主回路各管脚状态确定后释放阀体开启控制权限,解决因电磁阀驱动控制器在上电或意外复位时的管脚不定态造成的工作面失控问题。设计了同步积分型红外广角收发装置,研究了基于红外信号强度的采煤机定位方法。分析信号调理及红外接收二级管实际模型及红外广角定位接收装置I/V转换电路的不稳定性,采用超前校正的方法向反馈系数中引入零点,补偿了实际模型产生的极点及引入相移,降低了环路增益穿越斜率;提出了基于通讯的红外信号同步积分方法,解决了多接收装置信号的同步问题,确保接收的模拟信号强度仅与收发装置位置相关;分析了井下采煤机红外发射点投影位置邻近若干接收点处接收红外光强的分布情况,提出了在接收装置被遮挡及未被遮挡情况下的采煤机发射点定位方法,及基于卡尔曼滤波算法的采煤机速度跟踪预测方法,提高煤机定位准确性。分析了电磁阀机械特性及电参数模型,设计了电磁阀实时故障诊断及驱动端线圈电流缓变数据获取平台,研究了基于电磁阀驱动线圈电流的缓变失效特征提取方法。分析了液压支架电液信号转换关键装置电磁阀的机械结构及执行机构机械运动特性,在建立电磁阀电特性模型的基础上,分析了电磁阀开启时线圈电流暂态过程的正常及故障状态特征;通过电磁阀实时故障诊断及驱动端线圈电流缓变数据获取平台,在获取电磁阀从正常到故障过程的电磁阀线圈电流数据基础上,结合电磁阀电参数模型分析了电磁阀线圈电流动态过程的缓变特征,采用模板匹配、K-L变换、小波变换、云理论等方法提取了电磁阀驱动端电流的缓变失效特征。基于神经网络的电磁阀缓变失效分类与预测研究。提出了基于受限卡尔曼滤波算法的缓变特征增强算法,解决了电磁阀在缓变失效过程中因突发故障引入特征量突变而引起的神经网络分类及预测分辨率降低问题;以趋势增强后的特征量作为网络输入量,使用LVQ网络、GRNN网络和PNN网络等五种常用神经网络对电磁阀缓变失效状态属性进行分类与预测,对各网络在相同平台及样本下的检验结果及网络参数进行对比。更多还原

【Abstract】 Roof support is the most important thing for mine produtcion safety, yield and efficiency in underground longwall coal mining. The safety, reliability and automation of the roof support system directly relate to the safety and production efficiency of coal production. As the core of the hydraulic support system, the hydraulic support electro-hydraulic control systems used in China mainly depend on foreign imports. The high price, long delivery time and expensive spare parts of the imported foreign products can not meet the needs of coal mine conditions of China. Therefore, consulting the advanced electro-hydraulic control system, and considering the mine environment of China to realize the domestic of electro-hydraulic control system have important economic and social significance for getting rid of the dependence on foreign products, accumulating the relevant technologies, improving the mining automation level, and reducing coal mine accidents. In this technology context, the hydraulic support electro-hydraulic control system is developed in this dissertation. To realize shearer coal linkage, rapid fault localization and prediction, and improve the safety and procuction efficiency of coal production, the mehtods of shearer localization and fault prediction for key equipent are researched.The mining system architecture, the mining process, and the advanced hydraulic support electro-hydraulic control system are analized. The power supply pattern, networking, isolation pattern, control functions, control process of electro-hydraulic control system for roof support system in underground longwall coal mining, and a single hydraulic support controller with its matching equipments are researched, which provide effective reference for following reseach.The hydraulic support electro-hydraulic control system is developed. The electro-hydraulic control system consists of a number of serial connected hydraulic support controllers. A single hydraulic support controller is mainly consisted of a main controller and an electromagnetic valve controller. The sequential power on method is proposed. By releasing the control right of valve control after ensuring that the pin states of the electromagnetic valve control loop are convinced, the longwall face collapse caused by unconvined pin states during reseting is avoid.Integral synchronization type infrared wide-angle transmitter and receiver are designed, and the localization methods based on received infared signal strength are researched. The actual models of signal conditioner and photodiode are analized. The instability of I/V conversion circuit is analized as well. Through lead compensation, the zero is introduced to the feedback coefficient to compensate the pole and the phase shift caused by the real model of the circuit, and reduce the slope of the loop gain for the purpose of making the system stable. The method of communication based infrared signal synchronization integration is proposed. The problem of signal synchroniaztion is solved, which ensure that the received analog signal strength only relates to the relative positions of transmitter and receiver. The infrared signal strength distributions of sevral neighbored receivers near the launch position of transmitter locatted on the shearer are analized. The localization methods at the cases of the reciever being shielded and not shielded, and the methods of tracking and prediction for the runing rate of shearer based on Kalman filtering algorithm are proposed to improve the accuracy of localization.The mechanism charateristics and the electrical parameter model of the electromagnetic valve is analized. The platform for real-time fault diagnosis of electromagnetic valve and coil current degradation data acquisition is designed. The degradation feature extraction methods are researched. As the key device that converts the electrical control signal to hydraulic control signal, the mechanical structure and character of mechanical movement of the electromagnetic valve are analized. The electrical model of electromagnetic valve is established. The characteristics of transient process of the coil current in normal and fault states are analized. Based on the platform, the degradation coil current data of the electromagnetic valve runing until fault is acquired, and the degradation characteristic of the dynamic process of coil current is analyed considering the electrical parameter model. The methods of template matching, K-L transform, wavelet transform, the cloud theory etc. are used to extract the degradation features of electromagnetic valve.The neural network based degradation qualitative classification and prediction method for electromagnetic valve is researched. The degradation trend feature enhance method based on limited Kalman filtering algorithm is proposed to reduce the influnence of mutation information in the extracted features caused by transient fault which reduces the performance of neural network. The LVQ network, GRNN network, PNN network etc. are used for degradation qualitative classification and prediction with the trend enhanced features as inputs of the networks. The prediction results of the five common used networks on the same conditions of simulation platform and samples are compared.更多还原