摘要
提出了一种基于核主成分分析(KPCA)-高斯混合模型(GMM)的球磨机状态实时监测与评估新方法。通过选取不同工况下的球磨机过程参数,基于KPCA提取反映设备运行状态的特征量,基于GMM建立表征不同工况的球磨机状态模型,并引入高斯混模型概率距离,计算当前状态与正常工况的相似度,作为状态指标(SI)实时监控球磨机的运行状态。通过对某电厂球磨机实际运行过程的监控与评估,表明所提方法的有效性和实用性。
A new monitoring and evaluation method for ball mill system based on kernel principle component analysis(KPCA) and Gaussian mixture model(GMM) was proposed. After process parameters of the ball mill in various operation modes being chosen, feature parameters reflecting the operation state of the ball mill were extracted based on KPCA, and the model expressing the state of the ball mill was established based on GMM. In addition, the probability distance of GMM was introduced to calculate the similarity between current operation state and normal operation state to serve as the state index(SI) for realtime monitoring the operation state of the ball mill. Actual operation process monitoring and evaluation of the ball mill system in a thermal power demonstrated the effectiveness and practicability of the proposed method.
出处
《矿山机械》
北大核心
2014年第9期66-70,共5页
Mining & Processing Equipment
基金
国家自然科学基金资助项目(50775035)
江苏省自然科学基金资助项目(BK2011391)
