|
1. A control arrangement having a controller, which is arranged to drive an actuator in an industrial process, the control arrangement comprising a setpoint controller arranged to utilize deviations between explanation values of machine learning and normal explanation values of machine learning, the deviations being input data to the setpoint controller, the setpoint controller forming a setpoint value for the controller, wherein the explanation values of machine learning and normal explanation values of machine learning are SHAP values, values from a LIME method, values from a DeepLIFT method or any other possible explanation values, wherein the setpoint controller is also arranged to utilize deviations between machine learning values and normal machine learning values, and/or deviations between measurement data and normal measurement data, and wherein said normal values are values that have been derived from good running periods of a process, which is controlled by the control arrangement, wherein the setpoint controller comprises at least one P module, an I module, or a D module, or any combination of these modules, said deviations being input data into said modules, the setpoint controller comprising also input mapping module/s for each output of said module/s, a summation module to sum output/s of the input mapping module/s, an output scaling module to scale an output of the summation module, an output mapping module in order to provide a normalized output, and a setpoint adjusting module utilizing the normalized output in order to change the setpoint value.
|
