Scheduling Multiproduct Chemical Batch Processes using Matrix Representation

• Main Author: Shafeeq, Amir
• Format: Thesis
• Language: English
• Published: 2008
• Subjects: TP Chemical technology
• Online Access: http://utpedia.utp.edu.my/8558/1/2008%20PhD%20-%20Scheduling%20Multiproduct%20Chemical%20Batch%20Process%20Using%20Matrix%20Representation.pdf; http://utpedia.utp.edu.my/8558/

Batch process plants are usually designed for the production of specialty and fine chemicals such as paint, food and pharmaceutical to meet specific product requirements as set by current market demand. Batch process plants can be operated as single product in which only one product is produced and multiple products which allow production of more than one product using same batch facility. The economics of the batch process heavily depends on efficient scheduling of the different tasks involved in manufacturing the range of products. The main objective of scheduling is generally to minimize completion time known as the makespan of the batch process. Product sequencing, which is used to set order of products to be produced, has a direct impact on the makespan particularly in the multiple products case. Another effect on makespan is observed for different transfer policies used to transfer the product intermediates between process stages. The generally adopted intermediate transfer policies are (i) zero wait (ZW), (ii) no intermediate storage (NIS), (iii) unlimited intermediate storage (UIS) and (iv) finite intermediate storage (FIS). In the past, the determination of makespan for each transfer policy has been done using a number of mathematical and heuristics approaches. Although these approaches are very efficient and are currently being applied in many chemical process industries but most of them end up with the solution in terms of complex mathematical models that usually lack user interactions for having insights of the scheduling procedure. This motivated the current work to develop relatively simple and interactive alternate approaches to determine makespan. The proposed approach uses matrix to represent the batch process recipe. The matrix is then solved to determine the makespan of a selected production sequence. Rearrangement of the matrix rows according to the varied production sequences possible for the specified batch process recipes enables the makespan to be determined for each sequence. Designer is then provided with the production sequence options with its corresponding makespan from which a selection could be made according to the process requirements.