Research on Condition Monitoring and Fault Diagnosis Technology of Wind Turbine Generator Sets

1. The significance of wind turbine condition monitoring and fault diagnosis

Wind power generation can ease the tight domestic energy supply situation and improve the energy structure, which is of great significance to the national environmental protection and sustainable development of the power industry. With the rapid development of the domestic wind power industry, wind turbine failure has become an issue that cannot be ignored. By real-time monitoring of the operating status of wind turbines, it is possible to timely discover the hidden faults in the operation of the units: by extracting the unit fault information and analyzing and processing it, it is possible to diagnose the cause of the unit failure and formulate effective treatment measures. This has great practical significance for improving the operating reliability of wind turbines and promoting the healthy development of the wind power industry.

2 Wind turbine condition monitoring technology

2.1 Technical Characteristics Analysis

One of the commonly used state monitoring methods for wind turbines is to start with the various operating parameters of the wind turbines and determine whether the unit is abnormal by monitoring the changes in the operating parameters of the unit. There are various types of operating parameters in wind turbines, which can be roughly divided into two categories: measurable parameters and calculated parameters. Among them, calculated parameters refer to the need to read measurable parameters and use certain algorithms to calculate them, and the calculated results are used as operating parameters. According to the actual situation of the unit, selecting appropriate measuring devices or algorithms as needed is an important prerequisite for wind turbine state detection. If the measuring devices do not match, it will not be possible to accurately measure the changes in the energy state or numerical value of the relevant power parameters: if the algorithm is used improperly or written incorrectly, incorrect calculated parameters will be obtained. This will affect the normal operation of the unit. However, different types and manufacturers of measuring devices and various algorithms have their own advantages and disadvantages. How to choose needs to be determined according to the actual needs of the unit.

2.2 Abnormal monitoring of measurable parameters

When monitoring the measurable parameters of a wind turbine, it is necessary to master certain methods and principles. Generally, the following points should be noted: (1) The measurable parameters in a wind turbine generally include voltage, current, frequency, hydraulic pressure, temperature, etc. Different measuring devices should be selected according to the type of measurable parameters, and multiple measuring devices should cooperate with each other. (2) Analyze and determine the upper and lower limits of the measurable parameters, and select measuring devices with appropriate ranges according to the range. (3) Analyze and determine the normal/abnormal value range of the measurable parameters and the action triggering conditions.

2.3 Abnormal monitoring of calculation parameters

The calculation parameters in the wind turbine generator set need to select appropriate monitoring technology and detection methods. In the actual operation process, the following points should be noted:

（1）在检测计算参数的过程中选择合适的算法。风力发电机组有多种类型,每种类型的风力发电机组内多处需要用到各类不同算法,而每类算法中又有多种策略可供选择。这要根据风力发电机组实际情况和需求来选择合适的算法,因为不同的算法直接影响最终的计算结果,选取合适的算法能够显著提高计算参数监测的准确性和计算效率。（2）选择合适的设备运行算法。所谓合适的设备是指:1）算法稳定、长期运行需要执行设备本身的硬件条件支持:2）风力发电机组内需配置可靠、稳定的数据传输设备及测量设备,为算法的运行提供计算依据及输出渠道。

2.4 Abnormal monitoring of measuring equipment

The measuring equipment in the wind turbine generator set may fail, so a corresponding monitoring mechanism is needed to prevent it. Common methods are: (1) Some measuring equipment is equipped with detection contacts, which are normally open/normally closed under normal conditions and normally closed/normally open under abnormal conditions. The equipment status can be judged based on the change of the contact status. (2) A set of signals are connected to the input and output ends of the measuring equipment to enter the main control system. At the same time, the main control system pre-stores the algorithm corresponding to the measuring equipment, reads the input value and output value at all times and compares them. If the output value does not match the input value, the monitoring equipment is judged to be abnormal.

3 Wind turbine fault diagnosis technology

3.1 Fault diagnosis and analysis

When diagnosing faults of wind turbines, it is necessary to comprehensively analyze various factors based on the complexity of the unit's own structure and the particularity of the unit's operating environment to improve the accuracy of the fault diagnosis results. Wind turbines have many moving parts and complex structures, making fault diagnosis difficult. Therefore, it is necessary to update traditional diagnostic technologies, actively apply new technologies and new concepts, accurately diagnose various faults, and provide a basis for fault resolution. For the fault diagnosis of wind turbines, it is necessary to accurately grasp various fault phenomena, conduct a comprehensive analysis based on the unit's power parameters, vibration, pressure, deformation, wear, temperature and other performance characteristics, and complete the fault diagnosis.

3.2 Thermal parameter analysis

Thermal parameter analysis of wind turbines is to determine the operating status of wind turbines by analyzing the changes in temperature and humidity during their operation.

The temperature inside the wind turbine generator set mainly includes: the internal temperature of major components (such as generators, gear boxes, generators, motion motors, converters, etc.), the temperature inside the cabin, the temperature inside the control cabinet, and the temperature of various hydraulic oils/lubricating oils. The humidity inside the wind turbine generator set mainly includes: the humidity inside the cabin, the humidity inside the control cabinet, etc.

By monitoring the thermal parameters in the wind turbine generator set, the operating status of the unit can be effectively monitored. At the same time, according to the change trend and feedback results of the thermal parameters, the location of the equipment causing the fault in the unit can be accurately determined, which can provide a sufficient and detailed basis for analyzing the cause of the fault.

3.3 Vibration analysis

The application principle of vibration analysis is to install vibration sensors on major components in the unit (such as gearbox bracket, generator bracket, main shaft bracket, unit frame). These sensors can accurately measure the vibration of major components in the unit. By processing and analyzing the vibration signals fed back by the sensors, the vibration status and operating trend of each component in the unit can be quickly and accurately determined, and the source and cause of the vibration can be analyzed to determine whether there is a fault in the unit operation.

Monitoring and fault diagnosis of wind turbines under the background of four major data

The traditional wind turbine condition monitoring and fault diagnosis mainly use various monitoring equipment to monitor, analyze and diagnose the relevant parameters in the unit. However, with the continuous development of big data technology, various big data-based analysis methods are constantly being applied to the condition monitoring and fault diagnosis of wind turbines.

4.1 Big Data Technology

大数据技术是指从大量的、不完全的、有噪声的、模糊的、随机的工业生产数据中,通过算法提取数据背后隐含的具有价值的规律。大数据技术在风力发电行业中的应用有:收集并汇总一定区域内或一定类型风力发电机组中的海量运行数据,通过大数据挖掘和分析技术,总结数据背后存在的显著的统计因素,利用这些统计因素进一步判断风力发电机组可能存在的故障类型以及故障发生的最大概率,建立相关区域或相关类型的风力发电机组故障模型,并根据故障模型提前制定处理策略。大数据挖掘和分析的技术手段有很多,常见的手段有基于最小二乘法或利用多元回归模型来构建大数据模型,然后进行回归分析得到影响变量的主要因素,这些因素可以广泛应用于风力发电机组状态检测和故障诊断。

4.2 Advantages of wind turbine status detection and fault diagnosis based on big data technology

The traditional method of wind turbine status detection and fault diagnosis is based on a small amount of data, mainly relying on the engineer's own technical experience and related inference assumptions to complete the detection and diagnosis of the unit status. However, this method has certain risks. Insufficient analytical data or defects in the engineer's own experience will lead to deviations in the detection and diagnosis results. On the other hand, the data obtained by traditional technology will be affected by sensor noise, data transmission medium and external interference during the collection and transmission process. The detection and diagnosis results obtained after analysis of such data are often inconsistent with the actual situation.

Big data analysis technology can effectively make up for the above shortcomings. By collecting, analyzing and processing a large amount of data, using data mining technology to remove potential interference data, the conclusions obtained are often more reliable. In addition, big data technology can also detect some subtle changes in the state of wind turbines during operation, and can timely discover and correct some potential faults in the units in advance, effectively reducing the probability of failure of wind turbines and improving the safety and reliability of unit operation.

5 Conclusion

In summary, due to the limitations of the operating environment and its own structure, wind turbines have a higher probability of failure compared to traditional power generation equipment, and the causes of failure are complex and diverse. In the face of this situation, it is necessary to conduct real-time, comprehensive and systematic monitoring of wind turbines, and adopt a variety of analysis and diagnosis methods to promptly discover and solve the failures during the operation of the unit to avoid huge economic losses.