In batch process, operation conditions change to meet the requirements of market and customers. For example, in injection molding, the same machine may be used with different molds and materials to make different parts. For each different mold, material, and machine combination, data based modeling process has to be repeated for the development of a quality prediction model. This may involve repetition of a large number of experiments, if common process characteristics are ignored. Obviously, this is inefficient and uneconomical....[ Read more ]

In batch process, operation conditions change to meet the requirements of market and customers. For example, in injection molding, the same machine may be used with different molds and materials to make different parts. For each different mold, material, and machine combination, data based modeling process has to be repeated for the development of a quality prediction model. This may involve repetition of a large number of experiments, if common process characteristics are ignored. Obviously, this is inefficient and uneconomical.

Despite that operating conditions may be different for different batch processes, certain process behaviors and characteristics are common under these conditions. Effective using and extraction of these common process behaviors and characteristics can allow fewer number of experiments for the development of new process model, resulting in savings of time, cost and efforts.

With this as the key objective, process similarity is defined and classified based on process representation. Model migration strategies are proposed for four types of process similarity, scale similarity, family similarity, inclusive similarity and unknown attribute similarity, which involve developing a new process model by taking advantage of an existing base model and process attribute information. For each type of process similarity, an illustrative example is given to demonstrate that the proposed method and results are satisfactory.