Data reconciliation based on a novel robust estimator with application to volatile kiln process

Accurate and reliable process data form the foundation for modeling and optimizing industrial processes. Nevertheless, actual process data always entails measurement errors, and certain key variables remain unmeasured. For this reason, this paper proposes a novel robust estimator founded on the complementary error function. First, the influence function is employed to assess the robustness of the proposed estimator. Then, numerical examples are deployed to validate the efficacy of the proposed robust estimator. Finally, the proposed robust estimator is applied to the production process of the volatile kiln, and a hierarchical data reconciliation model of the volatile kiln is established. Comparative analysis between the proposed robust estimator and the Luan robust estimator reveals that the former exhibits superior performance, leading to more accurate and stable reconciled results.