Energy data anomaly detection based on association rule data mining

In recent years, the world has seen rapid advances in science and technology, and the level of automation in power plants has increased. Modern power plants generate huge amounts of data every hour, which are stored in real-time databases. Due to the complexity of modern power plants and the diversity of their operating characteristics, these data include very rich but difficult to be discovered knowledge, which makes the operation and management of power plants and fault diagnosis cannot be carried out in a timely and effective manner, seriously affecting the status monitoring and diagnosis of important auxiliary equipment such as power plant fans and pumps, and cannot meet the requirements of ensuring the safe and reliable operation of auxiliary equipment. In order to effectively monitor the auxiliary equipment of power plants and predict the occurrence of faults in a timely manner, this paper uses association rule data mining to establish a model for training based on obtaining a large amount of actual historical operation data in the database of a power plant, and uses the existing operation data for model testing, so as to determine whether the equipment is in the process of fault formation. The results show that using the historical operation data of the equipment and adopting association rules for analysis can effectively reflect the relationship between the measured values and thus achieve the purpose of fault early warning.