Session: CIE-07-01: CAPPD General

Paper Number: 116347

116347 - Optimal Design and Process Planning for Product Adaptation With Initial Evaluation of Solutions Based on Information Entropy 

Adaptable products are used to satisfy different and changing customer requirements through changes of products such as reconfiguration and upgrading during their utilization stages. Design of adaptable products can reduce product costs and extend product lifespans, thus improving competitiveness of products in the marketplaces. Product adaptation is the process to modify the existing product to satisfy the new requirements. The process for adapting an existing product, compared with the creation of a new product from scratch, can result in savings in product costs for the customer, thus making the product more competitive in the marketplace. The environmental impact can also be reduced through modifying existing products/designs instead of creating new ones.

In this research, an optimal design and process planning approach is developed for product adaptation with initial evaluation of solutions based on information entropy. In this approach, various candidates of design configurations and operation processes for adaptation of a product are modelled by an AND-OR tree with design and process nodes. A design node such as a component or an assembly describes a partial design solution, while a process node such as an assembly operation or a disassembly operation describes partial process solution. Design nodes are classified into two categories, un-adaptable nodes and adaptable nodes. Un-adaptable nodes cannot be changed for product adaptation, while adaptable nodes can be changed (e.g., removed or replaced) for product adaptation. Design and process nodes are further associated with design and process parameters. Optimization to identify the adaptable product design and process is conducted at two levels: configuration/process optimization level and parameter optimization level. For each configuration/process candidate, parameter optimization is carried out to identify the optimal design and process parameter values based on optimization objective functions such as performance and cost. Among all configuration/process candidates, configuration/process optimization is carried out to identify the best product adaptation configuration and process. Since the design solutions of adaptable products are evaluated by multiple evaluation measures with different units, these evaluation measures are converted into comparable evaluation indices to build the overall evaluation index considering importance factors of these evaluation indices. In this work, satisfaction indices are used as the evaluation indices. Data points considering evaluation measures and evaluation indices are obtained from customers and/or engineers. Then, the least-square curve fitting method is then used to obtain the numerical relations between evaluation indices and the corresponding evaluation measures. The multi-level optimization method is not effective when large numbers of adapted design candidates, adaptation process candidates, and design and process parameters are considered to identify the optimal design solution. Information entropy is employed to evaluate the partial configuration/process candidates modelled as branches in the AND-OR tree to eliminate the branches that are unlikely to lead to the optimal solution to improve the optimization efficiency when extensive computation efforts are required in parameter optimization. Information entropy is an effective tool to evaluate the amount of information considering multiple states when the probabilities of these states are not the same. Information entropy, as one evaluation measure, together with other evaluation measures such as design adaptability, product adaptability, performance, costs, etc. are used primarily due to multiple solutions are usually considered for design adaptation and product adaptation. In this work, the design and process nodes in the AND-OR tree are associated with probabilities representing the different opportunities that these nodes are selected for the final solution. Since design configuration and adaptation process candidates are created from the AND-OR tree through tree-based search, useful information for these design configuration and adaptation process candidates can be evaluated using information entropy. Numerical example and engineering application are developed to demonstrate the effectiveness of the newly introduced approach. The results of the studies have shown that the branches of AND-OR tree with more useful information represented by higher entropies are more likely to lead to the optimal design solution.

Presenting Author: Mohammed Saad University of Calgary

Presenting Author Biography: Mohammed H. Saad received the BSc. in mechanical engineering in 2016, and the MSc. in mechanical engineering in 2018, both from the American University of Sharjah (AUS), Sharjah, United Arab Emirates. He is currently a Ph.D. student at the Department of Mechanical and Manufacturing Engineering, University of Calgary, Calgary, Alberta, Canada. His research focuses on design of adaptable products, engineering optimization, and sustainable manufacturing. Mohammed is a recipient of the Alberta Graduate Excellence Scholarship (AGES), and the Ph.D. excellence scholarship from the Department of Mechanical and Manufacturing Engineering at the University of Calgary.

Authors:

Mohammed Saad University of Calgary
Deyi Xue University of Calgary