Introduction to Performance Evaluation


This section lays the groundwork needed for understanding and utilizing the performance evaluation model. The purpose, scope, and assumptions underlying the model are explained.

It should first be stated that the cost-based comparison of the two process variants is only half of the picture, and that a process comparison model is incomplete without some sort of relative performance evaluation. Hence for the sake of completeness, a simple methodology for such is proposed. Based on collaboration with several SM and MM molders, including Space Limited, Multishot, and Black & Decker, several important performance issues were identified. Based on this information, corresponding performance aspect definitions and measurement techniques were developed as detailed below.

Model Purpose

The purpose of this model is similar to that of the cost estimation model. The only difference is that performance measurement metrics will be used in stead of a cost-based metric. Specifically, the model is designed to help decision makers choose between two process variants based on several measurable performance characteristics. The CAD model of the products would be input to the model, and with minimal other user inputs, the performance aspects would be evaluated and output side-by-side so a decision could be made. Based on the numerical results obtained through evaluation of the model, the exact decision process is left up to the product designers. Specific decision-making methods (such as decision-based-design or multi-objective optimization) are not considered.

Model Embodiment

The model is designed to evaluate several performance aspects of a product through geometric analysis of the CAD model. Although full automation is the eventual goal of this model, much of the CAD analysis will be manually performed for the time being (i.e. certain things must be input, like the mold parting direction, etc.).

Unlike the cost estimation model, this model would not have a graphical output like quantity-cost curves. This model's output is simple numerical values that can be displayed as a table for relative comparison purposes. Each row of the table would correspond to an independent measurable physical characteristic, and the columns would represent the measured values for both SMM&A and MMM variants.

Unlike the cost estimation model, this model would require very few equation evaluations - much of the needed information should be extracted directly from the CAD files. The model will not use a weighting system to rank the various performance aspects. This is because the relative performance of the performance aspects is both subjective and application-specific. For instance, for a certain product, weight may be a very important consideration, while assembly clearances are not. We will leave this weighting to the model user.

In stead of providing a concrete decision-making model, this model is simply intended to help evaluate certain relevant physical characteristics of the molded products. The output of this model can be input to a suitable decision-making model, such as a multi-objective optimization routine. It is recommended that a strategy, such as plotting Pareto-optimal frontiers be used to help the decision makers choose the better variant based on their specific needs.

Model Scope and Assumptions

The scope of the model is the same as the cost estimation model. All of the assumptions still hold, but are less important for this model, as they are not used to evaluate most of the performance characteristics.

Model Application Example

This model could be used in any of the application examples of the cost estimation model. Here, instead of evaluating the manufacturing costs of each variant, their relative performance in several separate categories will be evaluated. This model is intended to supplement the cost estimation model by providing several separate performance metrics that can be used in conjunction with cost to choose the better variant.