节点文献

提升机危险载荷分析与预警系统研究

Analysis of Hoist Dangerous Load and Study on Its Prewarning System

【作者】 马驰

【导师】 肖兴明;

【作者基本信息】 中国矿业大学 , 机械电子工程, 2012, 博士

【摘要】 矿井提升作为煤矿生产的重要环节,直接关系到生产的安全和效率。然而,迄今为止,提升作业中发生的安全事故在矿山机电事故中仍占有不可忽视的比例,而且其中有相当一部分是由提升机危险载荷引起的。故此,本文在对提升机危险载荷进行动力学分析的基础上,开展危险载荷预警系统的研究,旨在通过危险载荷的识别与预警进一步提高矿井提升机的安全性。在危险载荷的动力学分析部分,建立了包括容器、钢丝绳、有效载荷等部分的纵向振动模型,采用基于Ritz级数的模态分析方法对过装载、未卸净、卡罐、松绳等危险载荷模式下钢丝绳的动态行为进行了研究,得出了不同类型、不同程度的危险载荷与钢丝绳振动特性之间的内在联系,为载荷监测参数的确定、危险载荷的识别提供了必要的依据;以实例验证了Rayleigh比对于钢丝绳、容器振动系统基频估计的有效性,以及Ritz级数法求解弹性杆件振动问题的收敛性。采用基于外点惩罚函数的遗传算法对提升机载荷多传感器融合系统进行了约束优化研究,将广义代价在系统有效度得以满足的前提下降至最低;搭建了提升机载荷信息融合试验装置,可模拟并记录多种提升机危险载荷;基于D-S证据理论建立了信息融合模型,抑制了多传感器信息的不确定性,并以过装、卡罐两种故障为例验证了该理论对于提升机危险载荷模式识别的有效性。将小波降噪、基于相关系数、能量周期与平均密度的EMD降噪、EEMD降噪三种方法分别应用于载荷信号处理,从实时性、自适应性、影响因素等方面对小波、EMD、EEMD降噪方法进行了对比,并指出EMD降噪方法具有综合优势,更适宜在危险载荷预警系统中优先采用;通过基于小波变换模极大值的Lipschitz指数法对提升机载荷进行奇异性分析,结果表明此方法可对松绳、卡罐、装卸载过程等奇异特征进行捕捉、定位,而且可反映冲击的强弱。建立了基于产生式规则的危险载荷预警系统;采用基于Web的浏览器/服务器(B/S)方式进行专家知识获取,实现了基于影响图表与极大似然估计法的专家知识融合;以基于DSTP协议的Data-socket技术实现了专家在线服务。研制了提升机载荷监护装置,对装置的机械、电气、软件等部分进行了设计分析,阐述了天线的输入阻抗、辐射类型、有效长度等参数对无线通信的影响,并从耐爆性能、隔爆性能两方面对有防爆要求的部件进行了隔爆设计,所研制的提升载荷监护装置已在煤矿现场得到了应用。

【Abstract】 Mine hoisting is an important link of coal production and directly related toproduction safety and efficiency. However, by so far safety incidents during hoistingrepresent a large proportion of the mechanical and electrical accidents in mines, andlots of them are caused by Hoist Dangerous Load. So study on dangerous loadwarning system was carried out on the basis of kinetic analysis on dangerous load,which was to upgrade hoist safety by identification and Prewarning of dangerous load.In kinetic analysis of dangerous loads, a vibration model was built, whichincluded a container, a wire rope, useful load, etc. With modal analysis method baseon Ritz Series, vibration of the wire rope while overloading, incomplete unloading,container jamming, and rope slack take place was researched. Relations betweendifferent types and varying degrees of dangerous load and rope vibrationcharacteristics was found, which could provide necessary foundation for parametersselection for load monitoring and the identification of dangerous loads. With actualexamples, Rayleigh ratio effectiveness for fundamental frequency estimation of therope and container was verified, Ritz Series convergence during the solving of elasticrod vibration problems was proved.Genetic algorithms based on outside point penalty function was adopt in thestudy on constrained optimization research of multi-sensor fusion system for hoistloads. Under the premise that the system validity had been achieved, generalized costof the system was minimized. An information fusion experiment device was built totest hoist loads, which could imitate and record kinds of dangerous load models.Based on D-S evidence theory, an information fusion model was built; uncertainty ofmulti-sensor information was reduced. Effectiveness of this theory for the modelidentification of Hoist Dangerous Loads was verified with the actual faults seriousoverloading and slight container jamming.Three methods which are wavelet de-noising, EMD de-noising based oncorrelation coefficient, energy period/average density, EEMD de-noising were appliedin the processing of hoist load signals. They were compared on the respects ofreal-time performance, adaptability and influencing factors, etc. It was pointed outthat EMD de-noising method has the comprehensive advantages and is more suitablefor the dangerous load Prewarning system. Lipchitz method based on wavelettransform modulus maxima was used in singularity analysis on hoist loads, which showed that it could catch and locate the singular feature during the process of ropeslack, container jamming, loading and unloading, and reflect impact strength.A dangerous load warning system was built based on production rules. The modeof browser/server based on web was adopted in expert knowledge acquisition.Knowledge fusion was realized with the methods of influence diagram and maximumlikelihood estimation. Data-socket technology based on DSTP protocol was used inthe realization of expert online service. For field hoist load monitoring and protectingdevice which is directly link to the dynamic library of the expert system, mechanical,electrical and software parts were designed. Sensors selection was achieved with leastsquare estimation and F-test methods. How the input impedance, radiation type andeffective length of an antenna affect wireless communication was illustrated. Forseveral parts of the system that works in the environment where there could beexplosive mixtures, flameproof explosion-proof design was conducted from therespects of explosion-resistance and explosion-proof performance. Hoist loadmonitoring devices developed have been applied in mine field.

节点文献中: