基于运行数据的风电场风速传感器运行品质评价方法及应用

An Operation Quality Evaluation Method and Its Applications for Wind Speed Sensors of Wind Farms Based on Operation Data

风电场风速传感器运行的可靠性和稳定性是机组高效发电和安全运行的关键,精准识别风速传感器的失稳和失准等缺陷模式,客观地描述其运行品质对于风电场提寿增效意义重大。该文对风电场风速传感器运行品质评价展开研究,围绕运行、测量、抗扰动等维度,建立风速传感器运行品质标签架构、量化规则和评价模型,提出基于数据驱动的多维标签应用体系,挖掘风电场风速传感器的扰动敏感度、韧性、性能演化趋势、集群特征,为辨识设备的运行态势提供一套精细化数学建模方法。以某风电场运行数据为研究对象,对所提的运行品质评价体系及应用策略进行阐述,验证所提方法在风电场主动运检和精益化管理方面的工程应用价值。

Wind speed sensors of wind turbines,especially offshore wind turbines,are prone to suffer from defects due to long-term exposure to complex and harsh environments,which negatively affects the quality of wind speed data.No wind signal or wrong wind signal fed into the unit control system tends to cause false start-stop and pitching actions,endangering the power generation efficiency and operation safety of wind farms.Existing methods focus on the fault diagnosis of sensors,serving as real-time operation control of units.Considering that the sensor may fail due to natural weather events(NWEs),the sensor is likely to recover smooth operation after NWEs.Extracting the operational quality(measurement accuracy,operational stability,disturbance sensitivity,and performance evolution trend)is difficult.Thus,refinement modeling for sensor operating quality helps to analyze its instability,misalignment,and other defects for the full-life cycle healthy management and proactive inspection of wind farms.Firstly,the operational quality characterization indicators and quantification rules for wind speed sensors are constructed from the dimensions of operation reliability,measurement accuracy,anti-disturbance ability,environment features,and inherent property.Then,a comprehensive evaluation model is employed using data envelope analysis(DEA)to describe the general attributes of sensors.An application framework of the constructed indicators is developed to deeply mine the risks and operating situations for the wind speed sensor,including a sensitivity model for the weakness of NWEs,a resilience model for the recovery ability against NWEs,a performance evolution trend model for defective modes,and a clustering model for the general features of the sensor community in the wind farm.Experimental results on an actual wind farm dataset,including the operational data of 9 turbines for one year,show that the proposed method can characterize the differences in operating quality for wind speed sensors.Specifically,the proposed model can detect the sensitive sensors prone to defects by NWEs and identify the defective modes,including misalignment and destabilization.Wind speed sensors in the studied wind farms,sensitive to cold waves and strong convective NWEs,are likely subjected to performance degradation.The rate of sensitive sensors to disturbances in the studied wind farm reaches 33.3%.Most sensors can recover to smooth operation within 1 hour after NWEs.According to operation quality,the sensor community in the wind farm is classified into fluctuating,robust,and degraded sensors.The following conclusions can be drawn from the case analysis.(1)The proposed method can evaluate the operation behavior of wind speed sensors.The application system can further excavate the attributes of the sensor,such as perturbation sensitivity,resilience,and clustering characteristics,effectively identifying its weaknesses and shortcomings.(2)Typical defects of wind speed sensors and their trend features are sorted out.The proposed application model can characterize the operational situations of the sensors and describe the occurrence probability of defective patterns.(3)The proposed post-assessment model and application system can describe sensor operation behaviors,which guide the whole-life-cycle management,active inspection,and lean maintenance of wind farms.