Configuration of Flexible Manufacturing Systems

Project Contact: Karl Kempf, Intel Corp
Proposal Contact: Raj Kawlra, GM Tech Center

Go to LFM Research Group 5 Mission Statement page

---

Background: With stiff global competition, the "Voice Of the Customer (VOC)" has been demanding increased diversity in product offerings, along with frequent updates to each product. This in turn is influencing manufacturing to invest in more flexible equipment for new programs. A piece of flexible equipment will be defined as one that has the ability to perform multiple processing tasks on a wide range of products (example, CNC Machining Centers for machining prismatic parts, or robots for material handling).

An opportunity that flexibility provides manufacturing is the possibility to explore various processing sequences available through alternate system configurations. Ideally, the configuration that yields maximum quantity of parts, at best quality, and at least cost, with high responsiveness to product change is desired.

Proposal Description: The focus of this research will be to provide the basic building blocks which can be used to assess the impact of different system configurations on the Quantity-Quality-Cost space.

Quantity, Quality, and Cost could be measured by the following parameters:

• Quantity: throughput (pieces/hours), inventory levels, uptime, ...
• Quality: ppm (parts per million) non-conforming, scratches/dents due to material handling
• Cost: investment cost (machines, material handling), operational cost (operators, set-ups, maintenance, rework), ...

Given that it is not always possible to simultaneously achieve all the three goals (of maximum quantity, best quality, lowest cost by any specific system configuration, the model would aid the manufacturer in determining the optimal manufacturing system configuration that will deliver a specific program objective, which may be:

• Maximize Quantity for a given set of Cost and Quality constraints.
• Maximize Quality for a given set of Cost and Quantity constraints.
• Minimize Cost for a given set of Quality and Quantity constraints.

A simple example that could be studied in detail would be a manufacturing system with (20) flexible machines. The issue in this case would be to evaluate the alternate system configurations, which may include:

• Serial System: a transfer line (all 20 machines doing only a part of the processing task).
• Parallel System: (20) parallel transfer lines with one machine each (each machine processes the entire part).
• Hybrid System:
  • (2) parallel transfer lines with (10) machines each
  • A serial line of (10) machines followed by (2) parallel lines with (5) machines each

The research topic requires:

• identifying all the system variables that affect each of the three performance measures -- Quantity, Quality, and Cost,
• understanding and developing the relationships between these sets of variables,
• building the optimization model so that meaningful trade-offs can be made. These trade-offs may be important at the time of conceptualization and design of the manufacturing system (if material handling will constrain the system from being reconfigured as required, or for daily/weekly running of the system,
• and possibly, validating at least part of the results of the optimization model using a real case study.

The model could later be generalized to evaluate alternate manufacturing system concepts having a mixture of flexible and special purpose machines.

Schedule: Research should last anywhere from 12 to 18 months, depending on the level of detail that is scoped for the project. The project would entail:

• Initial visits too different sites for a total of (8) to (10) weeks (roughly 2 to 3 weeks at each factory), including interviews with senior manufacturing engineers and possibly equipment builders. These interviews should result in a better understanding of the system issues and identification of the system variables.

• Literature search and brainstorming sessions with senior research personnel -- approximately (10) to (15) weeks.

• Modeling and software development -- (15) to (25) weeks.

• Validation of the model developed using a real case study -- (15) to (25) weeks.