【作者】 吴新忠；

【导师】 马小平；

【作者基本信息】 中国矿业大学 ， 电力电子与电力传动， 2010， 博士

【摘要】 作为煤矿通风系统安全可靠的重要指标,煤矿主通风机需要运行稳定、故障少、无喘振、工况合理;有效的煤矿主通风机通风失稳控制对整个通风系统的安全乃至煤矿的安全生产都具有重要的意义。论文引入了“通风机侧通风系统”的概念,并通过对原通风系统的地面风道进行改造建立了该“通风机侧通风系统”的实际模型,利用通风机的等效变位理论,该物理模型可以等效成一台变位通风机,通过对影响其性能曲线的关键参数的分析,进一步提出从“通风机侧通风系统”的角度来进行“通风机运行异常通风失稳的防范”和“通风机侧通风系统稳定的控制”以保证整个矿井通风系统的稳定;相对于传统的针对单一风机控制来实现通风稳定的不足,该方法具有更高的可靠性。在防范通风机运行异常通风失稳方面,论文将通风机的常见故障重新分类为“致命性故障和非致命性故障”并给出其定义;从防范通风失稳的角度,提出了一种将故障树、人工免疫和神经网络相结合的通风机综合故障预警和诊断方法。其中,故障树推理用于实现对通风机的致命性和非致命性故障的快速分类;对于非致命性故障,则通过人工免疫和神经网络相结合来进一步实现对故障异常度的检测和简单分类;人工免疫的反面选择算法解决了目前在故障样本普遍缺少的情况下进行通风机故障诊断困难的问题。神经网络通过根据经验设置的人工免疫检测器进行训练,实现了对非致命性故障的不同异常度和故障种类的多分类,为通风机在故障情况下的通风失稳控制提供及时、准确的决策依据。在通风机倒机通风失稳控制方面,论文分析了备用通风机能否正常启动具有一定的不确定性,提出了煤矿通风机倒机前热备用的策略并予以实践;备用风机提前启动进入热备用,可以回避风险,消除因其启动失败对通风系统造成的影响。鉴于在传统的“停机倒机”模式下,倒机过程中通风动力的缺失是造成煤矿主通风机通风失稳、瓦斯积聚超限的根本原因,论文提出了一种“不停风倒机”的控制策略,该控制策略可以保证在整个倒机过程中通风动力可以得到持续可靠地供给。鉴于轴流式通风机的特性曲线存在不稳定工作区的问题,在主通风机并联运行、利用风门进行风路切换的不停风倒机过程中,为了防止通风机工况点落入喘振区,基于风阻等效的理论,论文研究了在主通风机切换过程中等效到通风机入口的通风网络的风阻的动态变化情况,给出了保证通风设备安全运行的边界条件。针对主通风机不停风倒机过程的特点,论文研究提出了基于顺序控制的不停风倒机控制方案。该方案基于风门匀速开、闭进行风路切换,并通过对倒机过程中等效到风机入口的风阻和风量波动情况与延时时间关系的数值模拟,得到了风路切换过程中的风门的配合方法和最佳延时。由于顺序控制在通风机倒机过程中应用仍然存在一定程度的通风失稳,文中进一步研究了一种基于模糊控制的通风机倒机通风失稳控制方法;由于模糊控制系统不仅要实现风量的稳定,更重要的是完成风门的规定的切换动作和保证风机的安全运行,该方法首先将通风机倒机过程中风机安全运行的边界条件,作为模糊控制器设计的约束条件;并研究了在有约束条件下模糊控制器规则的设计方法;为了完成规定的风路切换任务,让四个风门之中的任意三个风门执行预定开、闭动作,将其风阻变化和由其造成的两台通风机的风量波动均作为扰动处理,而模糊控制器通过对余下的一个风门的开度控制实现了在倒机过程中保持系统风量稳定的目标。文中通过现场测量积累的样本数据训练神经网络对通风机侧通风系统进行模型辨识;以神经网络辨识的系统模型作为控制对象,对模糊控制器在通风机倒机过程中通风失稳的控制作用进行仿真,得到被控风门角度和风量输出的仿真曲线,验证了模糊控制方案的可行性。鉴于通风机喘振不仅是通风系统失稳的重要原因,同时是对通风设备自身安全的严重威胁,结合通风机喘振通风失稳控制中对保证通风系统稳定和通风设备安全的双重需求,文中提出了一种基于喘振边界线(SLL)和轴向位移或式组合的煤矿主通风机喘振预测方法;进一步地,作为通风机发生喘振的有效控制,文中提出一种通风机喘振分级消除的控制策略,该策略可以在通风机发生喘振时,实现将其工作点迅速控制在喘振报警线(SAL)以下的就近区域的目标。最后,以平煤股份的科研项目为依托,将本文的研究成果应用到矿井主通风机监控系统中,实现了主通风机在定期自动倒机,和故障状态下的自动倒机期间保持通风系统稳定的目标,并消除了高瓦斯矿井在传统“停机倒机”方式下的倒机期间因为通风失稳造成的瓦斯积聚超限的安全隐患,倒机过程中风速和瓦斯浓度的现场运行数据进一步证明了通风机通风失稳控制的有效性。更多还原

【Abstract】 As an important index of mine ventilation system safety and reliability, main fan is required to operate stably, with little fault, no surge and a reasonable working point. Therefore, effective control to prevent mine main fan from ventilation instability has great significance on the safety of ventilation system as well as mine.The concept of“main fan side ventilation system”as a whole including two main fans and air doors is introduced in the dissertation and its actual model is established through reconstruction of original ground air duct ventilation system. With ventilation equal theory, the model can be equivalent to a special displacement fan at the outlet of air shaft. Based on analysis of all key parameters which may influence its working point, a novel ideal to keep mine ventilation stability through preventing main fan from abnormal operation and controlling fan side of ventilation system stability at the same time is put forward. Compared with the shortage of conventional controlling single-fan for ventilation stability, the strategy of controlling main fan side ventilation system has higher reliability in achieving the stability of ventilation system.In the prevention of ventilation instability due to main fan running abnormally, based on the requirement analysis of main fan fault diagnosis for ventilation stability, main fan common faults are reclassified as“fatal fault and non-fatal fault”and their definitions are given. Then a novel method with the integration of Fault Tree Analysis (FTA.), Artificial Immune System (AIS.) and Artificial Neural Network (ANN.) for main fan fault early-warning and diagnosis is proposed, in which fault tree inference can identify fatal fault from non-fatal fault quickly. For the non-fatal fault, further fault classification and abnormal degree detection will be achieved by AIS and ANN. Based on Negative-Selection Algorithm, AIS fault diagnosis doesn’t need field fault samples, therefore, difficulty on main fan fault diagnosis due to lack of fault samples is solved successfully. Then many immune detectors on behalf of all kinds of fault with different abnormal degree are generated randomly and selected with rules for the training of a BP ANN, which will be used to realize main fan multi-fault classification.On another side of ventilation instability due to main fan switchover, considering the uncertainty of whether main fan standby can start successfully, main fan warm-standby before switchover is proposed, in which the risk of main fan starting failure during main fan switchover originally can be avoided. Considering the root cause of gas concentration exceeding limit during main fan switchover in traditional way is lack of ventilation power, a novel strategy for main fan switchover is proposed to realize ventilation power supplying continually and reliably during the whole process of switchover. As axial flow fan has an unstable working area, and in order to prevent main fan working point from falling to surge area during switchover, the equivalent resistance at the inlet of main fan is calculated, and the limit for main fan safety operating during main fan switchover has been deduced.Considering the characteristics of main fan switchover aiming at ventilation unceasing, a kind of sequential control scheme is put forward for main fan automatic switchover. Based on numerical simulation of equivalent resistance and flow rate during main fan switchover with four air doors in uniform motion, the cooperating way and the optimal delay time in sequential control have been gotten. For certain extent ventilation instability still exists during main fan switchover with sequential control, a kind of control based on fuzzy control theory is put forward in the dissertation, in which, the limit of main fan safety operating during main fan switchover is treated as a restriction in fuzzy controller designing. In order to accomplish the main fan switchover operation, three of four air doors actuate in each scheduled way, and their resistance changes and corresponding flow rate fluctuation of two main fans in parallel are considered as disturbance, and the objective of ventilation stability is achieved through adjusting the rest one of four air doors. In order to check the validity of fuzzy controller, field data during main fan switchover are collected to train a BP ANN, which is used to replace real ventilation system in control system simulation. And the simulation indicates that fuzzy controller is well designed to realize the objective of ventilation stability during main fan switchover.Considering that main fan surge is not only an important reason for ventilation instability but also a serious threat to main fan security, with the requirement analysis of main fan surge ventilation instability control, a way for main fan surge forecast with the combination of either axial displacement exceeding limits or working point across surge limit line is given. Fatherly, a level-based surge eliminating strategy is put forward too, which can control the working point nearly under the surge alarm line while main fan surge occurs.Finally, with a scientific research project of Ping Coal Mine Group, the achievement in the dissertation has been applied into a main fan monitoring system in a high gas mine. Field operation shows that ventilation stability during main fan switchover has been realized, and gas concentration exceeding limits due to main fan ventilation instability has been eliminated. The data of flow rate and gas concentration proves the feasibility of main fan ventilation instability control.更多还原