| Operation optimization is the main research direction of energy saving and emission reduction of thermal power units.It is not only an important means to ensure the production safety of thermal power units,but also an important means to improve economic operation.At present,the key and difficulty of the optimal working condition of thermal power units lies in how to determine the target value,operation optimization operation guidance and so on.With the construction of intelligent power plant and the popularization and application of cloud storage technology,electric power big data arises at the historic moment.Big data is a research hotspot in the computer field in the 21 st century.It includes advanced technologies such as association analysis,statistical analysis and model prediction.data mining can extract hidden information containing knowledge rules.A large number of historical operation data of units are stored in the thermal power plant management information system(MIS).Data mining can find the optimal operating conditions,which has important guiding significance for improving the economic operation of thermal power units.However,with the intelligent units and the growth of power consumption,the data stored by MIS is getting larger and larger,which makes the traditional data mining algorithms cannot meet the performance requirements.The emergence of cloud computing is just to solve this problem.Based on the actual operation of the unit,this paper uses the method of the combination of cloud computing and data mining to analyze big data of thermal power plant to guide the economic operation of the unit.It analyzes and studies several aspects in the optimization of power plant operation,such as the loss of unit energy consumption,the determination of target value,economic index,the construction of cloud computing platform and so on,and adopts a parallel clustering algorithm based on cloud computing,which has higher computing efficiency.Firstly,the cloud computing platform Hadoop and its sub-framework Map Reduce(distributed computing framework)and HDFS(distributed file system)are analyzed and studied,and the experimental platform is built with three computers,which provides the basis for the underlying implementation of data mining.Then,according to the data mining technology,the principle and application of fuzzy C-means clustering(FCM)algorithm are introduced.A new algorithm is formed by combining rough set theory(Rough set)with FCM algorithm,which is called RFCM algorithm.The attribute reduction of rough set is used to reduce the dimension of the data and realize the data preprocessing in the process of data mining.Finally,the RFCM algorithm is run on the cloud computing platform,the historical operation data of a 1000 MW unit is applied to the cloud computing platform for cluster analysis,the optimal operating point is mined,and the high efficiency of the algorithm is verified,which provides a basis for the economic operation of thermal power units. |
