Model Migration with Inclusive Similarity for Development of a New Process Model

Abstract

In the processing industries, operating conditions change to meet the requirements of the market and customers. Under different operating conditions, data-based process modeling must be repeated for the development of a new process model. Obviously, this is inefficient and uneconomical. Effective use and adaptation of the existing process model can reduce the number of experiments in the development of a new process model, resulting in savings of time, cost, and effort. In this paper, a particular process similarity, inclusive similarity, is discussed in detail. A model migration strategy for processes with this type of similarity is developed to model a new process by taking advantage of existing models and data from the new process. The new model is built by aggregating the existing models using a bagging algorithm. As an illustrated example, the development of a new soft-sensor model for online prediction of melt-flow length for new mold geometry for an injection molding process is demonstrated by taking advantage of existing models for different molds.

Bayesian migration of Gaussian process regression for rapid process modeling and optimization

“…A straightforward approach to allocating initial experiments for the new process is to apply a space-filling DoE method, similar to the design of the old process, with the aim to obtain a fair coverage of the factors' space [12,14,11]. Later, this approach was recognized to ignore important information that could lead to better initial design [13].…”

“…The scale and bias parameters may be estimated by minimizing the squared prediction errors. More recently, other migration strategies have been explored, including the local modeling approach [13] and ensemble modeling for dynamic models [14,11]. In these previous studies, either quadratic polynomials [12,13] or ANNs [14,11] were chosen for empirical modeling.…”

“…In addition, the current model migration studies [11,12,13,14] are primarily focused on improving the prediction accuracy of the new model across the entire range of the factors. If the objective of modeling is to identify the process factors that give the optimal process response, then the new model will be subsequently used for optimization purpose.…”

Aging trajectory prediction for lithium-ion batteries via model migration and Bayesian Monte Carlo method

“…The model migration method was initially developed by Lu and Gao in 2008 to reduce experimental efforts when modeling similar injection molding processes [26]. The basic idea of model migration is illustrated in Figure 1.…”

Transfer of a nanoparticle product between different mixers using latent variable model inversion

“…Data‐based approaches to guide technology transfer activities have been proposed to transfer analytical methods and calibration models between different instruments . Contributions have been provided to assist the transfer of soft sensors and models between different plants to predict the final product properties based on process similarity . Other studies have focused on the transfer of models for process monitoring purposes, and on the use of multi‐variate charts to monitor the quality of products manufactured in different plants with different production targets or to establish multi‐variate specifications for incoming raw materials employed in plants of different scales …”