Generic causal determination of thermal model and heat input by using backward simulator

摘要

Generally, it is difficult to estimate model parameters for time varying output system especially when the input is unknown. One way to estimate them is to repeat simulations to search for the case where the outputs match by changing model parameters and input signals. However, such method is time consuming. Our previous studies show that the backward simulation with an appropriate model can estimate the input signal. Therefore, it is efficient only to change model parameters in the backward simulation if we can properly assume an appropriate model for the system. This paper describes a method using a generic causal model for determination of a thermal model to match the measured temperature change. Such simulator is implemented by using Node-RED. The performance of the simulator is shown for the cases of ideal resolution data, limited resolution data, and real measured data.