基于岭回归算法的压缩机性能曲线实时修正

Compressor Performance of Real-time Prediction Base on Ridge Regression

压缩机作为气体储能系统中的关键设备,其性能曲线的测定具有关键意义。在目前的实际工程应用中,压缩机的性能曲线通常是实测出几个工况点,然后通过简单的线性回归获得,这种方式存在一定问题。文章提出在压缩机运行过程中,实时通过机器学习的方式来处理数据,能够获得更加准确的性能预测模型和性能曲线,并且该曲线能够随着机组的运转得到连续的修正。

In current practical engineering applications,compressor characteristic curves are usually obtained by measuring a few operating points and then caculated with simple linear regression Model.This approach has a series of problems.In this paper,we propose to process the test data by ridge regression,which can obtain more accurate performance prediction models and characteristic curves.