论文部分内容阅读
对机组中长期运行趋势的提取方法进行了研究，提出了基于小波包时频掩模技术去除机组运行趋势噪声的方法，用该方法去除了短期的随机波动，得到了均值意义上的趋势。对该趋势提前进行时间序列预报，可以得到机组未来均值意义上的趋势，显著提高机组运行状态的预报效果，且该方法具有对数据的广泛自适应性。
The extraction method of mid-term and long-term operation trend of the unit is studied, and a method of eliminating the trend of unit operation noise based on wavelet packet time-frequency mask is proposed. By using this method, short-term stochastic fluctuation is removed and the mean value trend is obtained. Forecasting the trend ahead of time series can get the trend in the future average of the unit and significantly improve the forecast effect of the operation status of the unit. The method has a wide range of adaptability to the data.