Adaptive Weighted Hybrid Modeling of Hydrocracking Process and Its Operational Optimization

Abstract

Hybrid modeling, aiming to integrate the advantages of both first-principles models and data-driven models, is an important technology for refinery process simulation and optimization. The commonly used hybrid models include series models, parallel models, and series−parallel models. Many studies have reported the operational optimization results based on these models. However, it is unknown whether the results obtained based on these models are consistent with the actual plant optimal operation. Moreover, the hybrid models that have been used in operational optimization are of traditional structures and cannot ensure competent performance. To fill these gaps, we investigate the modeling and operational optimization of a typical refinery unit hydrocracking unit. First, we establish a first-principles model, a data-driven model, and three hybrid models and analyze the strengths and weaknesses of these models. Next, we propose the concepts of "mechanism-dominated models" and "data-dominated models" to classify the existing models according to the prediction abilities instead of the structural features. Furthermore, a novel type of hybrid model termed adaptive weighted hybrid model (AWHM) is proposed to gain a better balance between extrapolation and interpolation capabilities. Then, the performance of the operational optimization based on these models is compared under four optimization scenarios. Experimental results demonstrate that the two variants of the proposed AWHM have the best optimization performance and their optimization results are the most consistent with the actual process.

A hybrid science‐guided machine learning approach for modeling chemical processes: A review

“…Song et al 81 also apply the direct hybrid model configurations of Figures 3–5 to an industrial hydrocracking process and analyze the strengths and weaknesses of these configurations. They call a model a mechanism‐dominated model if the accuracy of its outputs is mainly dominated by the available theoretical knowledge used to develop the model; and they also call a model a data‐dominated model if the accuracy of its outputs is mainly dominated by the quality of the training data and the performance of the resulting data‐based model.…”

“…In their work, Song et al 81 combine a mechanism‐dominated model with a data‐dominated model as a hybrid direct model of Figure 2, with the weighting factors for the outputs of two individual models being determined in an adaptive fashion. For their application, Song et al work with a mechanism‐dominated model of an industrial hydrocracking process based on kinetic lumping, 80,81 and with a data‐dominated model based on a self‐organizing map (SOM) followed by a convolutional neural network (CNN), with both being trained by simulated process data‐based on Aspen HYSYS 81 . They evaluate the performance of the hybrid model for operational optimization of the hydrocracking producing different product scenarios.…”

Research on Product Yield Prediction and Benefit of Tuning Diesel Hydrogenation Conversion Device Based on Data-Driven System

“…The first is the mechanism-driven model (MDM), also known as the summation model, which is based on a number of assumptions and reaction mechanisms to model the actual reaction process [1]. The three main types of MDMs are wide fraction summation, narrow fraction summation, and continuous summation.…”

“…They are also limited by the coupling effects between operational conditions, leading to a relatively lower accuracy. Therefore, their usage is subject to limitations [1,6].…”

“…The second category is the data-driven model (DDM), which is also known as a black box model [1]. The main characteristic of the DDM model is that it is based on existing operational data to establish the mapping relationship from inputs to outputs.…”