| With the development of the economy and the advancement of society,electric energy has long been an indispensable source of energy in production and life.Improving the operational efficiency of the power equipment and enhancing the continuous operation of power equipment are the key concerns of the power industry.While improving the industrial system,it is also necessary to improve the operational efficiency,reliability,safety and economy of electrical equipment.Whether it is a thermal power plant or a nuclear power plant,the steam turbine generator set is a very important key equipment.80%of the electricity used by humans is provided by steam turbine generator sets.To ensure the normal operation of the steam turbine generator set and reduce its downtime due to abnormal reasons,it has a positive and farreaching impact on promoting the development of the power industry and even the national economy.Therefore,it is of great significance to monitor the operation status of the steam turbine generator set.As an important power equipment in industrial production and life,the concept of reliability must be put in the first place during the operation and maintenance of equipment.To evaluate the reliability of a steam turbine generator set by traditional reliability,it is necessary to rely on historical sample data of a large number of equipment components,however,in engineering practice,not all sample data can be completely preserved or have a sufficient number of historical sample data,which is also a deficiency of traditional reliability.In view of this situation,this paper will apply machine learning technology to equipment condition monitoring of steam turbines based on equipment operation data,and design a system to monitor the operating status of steam turbine components to understand their health level.The health rate is displayed in the form of a percentage to achieve the purpose of mastering the operation reliability of the equipment;at the same time,from the perspective of reliability,the reliability evaluation method based on Weibull proportional hazards model is used to calculate the reliability related index of steam turbine equipment and evaluate the reliability.Relevant indicator calculation and reliability assessment are compared with the results obtained by the equipment condition monitoring system.A evaluation method based on equipment operating state characteristics is used to evaluate the reliability of key components of steam turbine shafting.Based on the proportional hazards model and the two-parameter Weibull distribution model,the method establishes a relationship between equipment operating state characteristics and equipment reliability,and a reliability evaluation method based on the Weibull proportional hazards model is formed.The signal characteristic indicators that can reflect the running state of the equipment are selected,and the parameters in the model are estimated by the maximum likelihood parameter estimation method,and then the reliability of the equipment is evaluated and the reliability index is obtained.The machine learning technology is applied to the condition monitoring of the Steam turbine generator set.The support vector machine related theory is combined with the data to construct the support vector machine model.The digital steam turbine 3D model is used to establish the visual interface,Based on the health index system of the equipment,a steam turbine condition monitoring system based on support vector machine is designed.The content includes system requirements analysis,system function analysis and system realization process design.The support vector machine is applied to the state monitoring of steam turbine generator sets.Based on the results of the monitoring,the maintenance strategy and the maintenance management strategy are optimized to better guide the operation and maintenance of the equipment. |
