Research On Data Preprocessing Technology In Supervisory Information System Of Power Plant

Supervisory Information System of thermal power plant is an important direction of development in recent years. Data preprocessing technology, which providing data support for the entire system is one of the key technologies to guarantee the system making reliable, scientific running. At present, the study of the steady-state process data preprocessing correction is almost consummate, but the actual production preprocess to the measure data for the dynamic measurement of thermal power plants and other dynamic data preprocessing techniques is still not able to form a reasonable and feasible solutions. Therefore, the study of dynamic data preprocessing techniques, not only contain theoretical significance of innovation, but also will conducive to the further development of many actual production process monitoring and optimization techniques.Dynamic data correction techniques include data pre-processing and data coordination. The main content of this paper is to research the identity and compensation methods for data preprocessing techniques in missing values and singular values. This paper used the real-time data collected from a power plant as the original data source, and focused on the missing data values and singular value problems, after researching of a variety of data preprocessing methods, considering the approach of error correction rates and compensation effect and thermal power plants Supervisory Information System requirements for timeliness combined 3σ-Criterion and Dynamic Measurement Uncertainty to identify singular value, and then take advantage of Autoregressive Integrated Moving Average Model data to predict the value of the missing values to obtain more accurate data without missing values and singular values. To improve the accuracy of the data further, and toke the timeliness of the whole system into account, the moving average filter is adapted to reduce the influence of random errors on data accuracy, and provided high-quality data for the next data coordination.This paper did some research about dynamic data correction and formed a primary method of data preprocessing technology in the real-time monitoring of thermal power plant information system, which includes recognizing singular value, compensating missing value and correcting singular value after the integrated study of data correction in data preprocessing technology.