Comparison of Semirigorous and Empirical Models Derived Using Data Quality Assessment Methods

For the first time, I am featuring a paper by me and colleagues from the Universities of Ilmenau, Pretoria and Anglo American. This paper extends work that Professor Shardt and I have been doing in the general area of using historical data for system identification, thus avoiding the step testing that is normally used. The example in this paper is a primary mill, and linear and non-linear models are built using the DQA methods.

For the case of a single input-output system, the method has been coded in a Matlab App. Further work will be done to extend this app to use heuristics to cut unsuitable data, as well as the extension to MIMO systems.

The paper is open access and can be found here https://www.mdpi.com/2075-163X/11/9/954

Brooks, K., le Roux, D., Shardt, Y. A., & Steyn, C. (2021). Comparison of Semirigorous and Empirical Models Derived Using Data Quality Assessment Methods. Minerals11(9), 954.