Study on an Intelligent System of Big Data Mining and Key Target Optimization for Power Plant

Abstract

This paper studies and establishes an intelligent system of data mining and key target optimization for thermal power plant. Through the process of stable operating condition division， operation parameter clustering and key target benchmarking optimization， valuable information in historical operating conditions is mined to find the target value and operation parameters similar to the current real-time operating conditions to provide open-loop suggestions for optimal operation. In the practical engineering application of benchmarking optimization of key targets of boiler efficiency， the system found that under similar MV （manipulation variable） and DV （disturbance variable） parameters， the boiler efficiency under real-time condition is 92.13% and that under historical benchmark condition is 93.34%. It is proposed that the setting value of oxygen in flue gas can be reduced from 5.5% to 4% under this condition. As a result， the boiler operation efficiency is improved and the coal consumption of power generation is saved.

Keywords

• coal-fired power plant
• big data mining
• operating optimization
• benchmarking