【作者】 仲元昌；

【导师】 汤宝平；

【作者基本信息】 重庆大学 ， 机械电子工程， 2011， 博士

【摘要】 随着社会生活节奏的加快和现代工程技术的发展,促使大型运动机械不断提速,带来了系统工作效率的大幅度提升,但由此产生的振动危害已越来越严重。因此,大型装备的状态监测和故障诊断越来越受到重视。而振动参数是装备运行中重要的特征参数,振动监测是装备状态监测和故障诊断的重要手段。研究表明,对于大型风洞机组的振动监测,现有的振动监测技术已不适应多参数监测的需要,传统的振动监测技术主要存在监测点参数单一、监测系统的动态范围小、通信方式不灵活等问题,影响了装备状态监测的可靠性。目前,对大型、复杂装备的振动状况监测,已成为装备状态监测和故障诊断技术领域的前沿课题。本文在深入研究现有振动监测技术和通信理论的基础上,结合大型风洞机组振动监测领域的具体应用环境,提出一种适用于大型风洞机组振动监测的“多点-多参数”远程监测方法,构建了基于多传感器管理的测振传感器阵列模型;研究了有线与无线并存的远程监测通信模式;研究了适用于风洞机组振动“多点-多参数”远程监测的多源信息融合模型;探索了解决大型风洞机组振动远程监测问题的新方法。本文的研究工作,有望改变大型风洞机组振动监测现状,以提高风洞机组振动监测的实时性、准确性和可靠性,为大型风洞机组的状态监测与故障诊断、以及工程设计提供科学依据。这对提升我国振动监测的理论水平和应用水平具有重要的理论价值和重大的现实意义。主要研究工作如下:①为了全方位监测大型风洞机组的振动状态,研究了大型风洞机组的振动特征,提出一种“多点-多参数”监测方法。现有的振动监测技术,一般都选择振动体的某些具有代表性的部位作为监测点,安装测振传感器,以实施其振动监测。这对于简单装备的振动监测是可行的。但是,对于大型的、复杂的装备而言,不同部件、同一部件的不同部位,其振动特征都是不同的,其单点单参数监测结果不足以全面表征监测对象的振动状态。为此,提出一种在不同部件、同一部件的不同部位分别布置监测点的“多点-多参数监测”模式。②为了确定“多点监测”的多个目标监测点,实现振动的多点-多参数监测,研究了基于有限元分析的振动特征提取与监测点布置方法,构建了一种由多个测振传感器组成的测振传感器阵列模型。大型风洞机组的振动情况复杂多变,必须全面考虑影响装备工作状况的各种因素,特别是模拟在极端条件下可能出现的振动状况,这就需要根据监测对象的几何参数和工作载荷等参数进行有限元分析,了解其基本振动特征,由此确定监测点布置模型。采用微型敏感元件作阵列元素,把若干具有不同功能的微型敏感元件构成“阵列”分布,并与共享的阵列管理器、信号处理电路以及智能芯片集成在一起,形成一种具有初级智能的多功能传感器组件——智能传感器阵列。③为了尽可能提高系统的可靠传输问题,研究了一种基带传输与频带传输并存的冗余监测通信模式。基于有线连接的振动监测系统应用广泛,技术成熟,可靠性高,但可维护性和扩展性差;基于无线网络通信技术的振动监测系统,对于旋转机械,无线监测更具优越性,但这类系统的安全性、可靠性及通信距离和功率受限等问题都是有待克服的挑战。为此,采用了以无线传感器网络(Wireless Sensor Network, WSN)、通用分组无线服务技术(General Packet Radio Service, GPRS)、Internet、与电力线通信(Power Line Communication,PLC)融合的振动远程监测通信方法作为补充,使其监测系统既能采用基带传输,也可采用频带传输。④为了提高监测系统运行效率、减小网络开销(特别是能量开销),采用适用于振动监测系统的分布式数据融合算法。一方面,振动监测所要求的采样频率通常在1kHz~10kHz范围内,高频采样将产生大量的振动数据;另一方面,由于采用“多点-多参数”监测布点模式,一个监测点处将有多种测振敏感元件,致使监测数据成海量增长。为了尽量减小网络开销、避免网络拥塞,确保信息可靠传输,必须采用多级数据融合技术,确保振动多点-多参数远程监测网络的高效运行。⑤多点-多参数远程监测系统的设计。在完成上述研究工作的基础上,构建了多参数远程测振与预警系统模型,研制了基于该模型的振动远程监测实验平台,并对该实验平台的基本性能及实际应用进行了测试实验。实验结果表明了该振动监测系统的可行性和先进性。文章最后对研究工作进行了总结和未来的研究进行了展望。更多还原

【Abstract】 With the development of pace of life and modern mechanical engineering technology, speed of large-scale moving machinery become faster and faster, and the efficiency of the mechanical system is improved a lot. But as a result, vibration hazards become a serious problem. Therefore, condition monitoring and fault diagnosis technology of large-scale mechanical equipment have gotten more and more attention. The mechanical vibration is an important characteristic parameter of the operation of mechanical equipment. Mechanical vibration monitoring is an important means of mechanical equipment condition monitoring and fault diagnosis. Studies have shown that: In some area of vibration monitoring to large-scale machinery and equipment (such as a large wind tunnel unit), existing mechanical vibration monitoring methods cannot meet the needs of large dynamic range monitoring of large-scale mechanical equipment. It mainly has the following shortcomings: single monitoring point, small dynamic range of the monitoring system, rigid communication mode and so on. These seriously affect the reliability of mechanical equipment condition monitoring and fault diagnosis. At present, vibration monitoring of large-scale and complex mechanical equipment has become the frontier of the field of mechanical equipment condition monitoring and fault diagnosis.Based on intensive study of existing mechanical vibration monitoring technology and communication theory, combining with the specific application environments in the field of large-scale and complex mechanical vibration monitoring area, a“multi-point-multi-parameter”remote vibration monitoring program for large-scale mechanical has been proposed, and the vibration sensor array model based on multi-sensor management has been built in this paper. We studied on the communication model, which is combined wireless sensor network with mobile communication network, internet and low voltage power line communication network for mechanical vibration remote monitoring. And also we do the research on the application of suitable multi-source information fusion model in“multi-point-multi-parameter”remote monitoring for mechanical vibration. Then, a new method of vibration remote monitoring of large-scale and complex mechanical equipment has been explored. Based on our research works, it is expected to completely change the status of mechanical vibration monitoring, the real-time、accuracy and reliability of mechanical vibration monitoring can be improved and the provide the scientific basis for the mechanical engineering design and mechanical equipment fault diagnosis. It has important theoretical value and practical significance to the improvement of the theoretical level and application level of mechanical vibration monitoring in China.Main research works are listed as follows:(1)For the all-round vibration status monitoring of large-scale mechanical equipment, a“multi-point-multi-parameter”monitoring program has been proposed.Generally in the existing mechanical vibration monitoring technology, a representative part of the vibration body is chosen as a monitoring point that is installed with the vibration sensors to implement the vibration monitoring. But this is only suitable for the simple mechanical equipment. For the large-scale and complex mechanical equipment, the vibration status of different parts and different positions of the same part is different. And the result of single point monitoring cannot fully reflect the vibration status of the objects. So, this“single point of monitoring”mode is far from meeting the needs of vibration monitoring and fault diagnosis for large-scale and complex mechanical equipment. Hence, the“multi-point-multi-parameter”mode has been proposed in which the monitoring points are arranged respectively at different parts and different positions of the same part. Through the vibration sensors from multiple monitoring points, it will all-round monitor the vibration characteristics of mechanical equipment simultaneously, then multi-sensor management technology and information fusion technology have been used to integrate the measured information, and at last, the statistical vibration characteristics of monitoring objects can be obtained. It provides reliable statistical data for vibration monitoring and fault diagnosis for large-scale mechanical equipment.(2)To determine the multiple monitoring points of“multi-point monitoring”, a program of mechanical vibration features extraction and monitoring points layout based on finite element analysis has been proposed.Because of the complexity of vibration status of large-scale mechanical equipment, the determination of monitoring points is so important that it will brings serious consequence if there is something wrong with it. Therefore, the selection and arrangement of monitoring points is the key point of the issue. All kinds of factors that will affect working conditions of equipment should be considered, and especially in the case that simulate the vibration status under extreme conditions. The finite element analysis need to be applied based on the geometric parameters and working load parameters of monitoring objects to get the basic vibration characteristics and then determine the layout points.The performance of vibration sensor is the key issue should be considered firstly in mechanical vibration monitoring. Generally, traditional sensors can only fulfill the conversion between non-electricity and electricity, and there are many shortcomings and inconvenience, such as weak signals, low sensitivity and nonlinear error. Especially when we monitor multiple parameters simultaneously, multiple sensors will be needed, and the traditional collection system is too complex. For this, we used mini-sensors as the array elements, and made the lots of mini-sensors of different function to construct an“array”. Then we integrated the array with the shared array managers, signal processing circuit and intelligent chip to form the multifunctional sensor component with primary intelligent, and it is called intelligent sensor array.(3)To achieve the reliability transimission of the system, a redaunt monitoring communication model based on both the baseband transimission and the band transmission has been proposed.Wired mechanical vibration monitoring system has been widely used, but it has shortcomings such as: complex wiring, higher deployment costs, poor maintainability and scalability. Therefore, mechanical vibration monitoring system based on wireless sensor network has emerged, and it brings new vitality to the wireless monitoring mode for mechanical vibration. However, the security、reliability、communication distance and the power limited issues of these systems are challenging.For this, after we deeply studied on the existing communication technique and made full use of its network resource, a remote monitoring communication program integrated of Wireless Sensor Network, General Packet Radio Service, Internet and Power Line Communication has been proposed. It is used as a supplement so that both the baseband transimission and the band transimission can be used in the monitoring system.(4)In order to improve the monitoring efficiency and reduce network overhead (especially energy costs), the distributed data fusion algorithm for mechanical vibration network has been proposed.On the one hand, the sampling frequency required by mechanical vibration monitoring is usually among 1 kHz~10 kHz, and lots of vibration data will be generated by the high-frequency sampling. On the other hand, because of using“multi-point-multi-parameter”mode there will be multiple sensors at one monitoring point, and as a result, the monitoring data will be into a massive growth. To minimize network overhead, avoid congestion and ensure reliable transmission of information, the multi-grade data fusion technology need to be used, so that the efficient operation of multi-point-multi-parameter remote monitoring network for mechanical vibration can be achieved.(5)Design and implementation of the multi-point-multi-parameter remote monitoring system for mechanical vibration.Based on completion of above research works, the "multi-point-multi-parameter remote monitoring system" model has been constructed and the remote monitoring experimental platform based on this model for mechanical vibration has been developed. Also, a lot experiments has been carried out to demonstrate the basic performance and practical application of the system. Experimental results have demonstrated the feasibility and advancement of the mechanical vibration monitoring system. The system can be widely used in the mechanical vibration monitoring of kinds of large-scale and complex machinery and equipment, and also can be applied to the area of monitoring health status of large-scale equipments.The end of the paper is the summary of research works and prospects for future research.更多还原