METHODS ENGINEERING:ENGINEERING DESIGN

4. ENGINEERING DESIGN

After reviewing the job design guidelines from Section 2 and the information from Section 3, you can design alternatives. Remember the five steps of engineering design with the acronym DAMES: Define the problem, Analyze, Make search, Evaluate alternatives, and Specify and sell solution. See Table 22.

1. Define the problem broadly.

Usually the designer is not given the problem but instead is confronted with the current solution. The current solution is not the problem but just one solution of among many possible solutions. The broad, detail-free problem statement should include the number of replications, the criteria, and the schedule. At this stage, putting in too much detail makes you start by defending your concept rather than opening minds (yours and your clients’) to new possibilities. At this stage, the number of replications should be quite approximate (within ±500%).

Criteria (see Section 2.1) usually are multiple (capital cost, operating cost, quality) rather than just one criterion. The schedule defines priorities and allocation of resources for the project.

2. Analyze in detail.

Now amplify step 1 (defining the problem) with more detail on replications, criteria, and schedule.

• What are the needs of the users of the design (productivity, quality, accuracy, safety, etc.)? See Sections 2.1, 2.2, 2.3, 2.4, 2.5, and 2.6.

• What should the design achieve?

• What are limits (also called constraints and restrictions) to the design?

• What are the characteristics of the population using the design? For example, for designing an office workstation, the users would be adults within certain ranges (age from 18 to 65, weight from 50 to 100 kg, etc.).

The design should consider not only the main activities (the ‘‘do,’’ e.g., the assembly) but also the ‘‘get-ready’’ and ‘‘put-away’’ and support activities such as setup, repairs, maintenance, material handling, utilities, product disposal, and user training.

Since people vary, designers can follow two alternatives: (a) Make the design with fixed characteristics; the users adjust to the device. One example would be an unadjustable chair. Another example would be a machine-paced assembly line. (b) Fit the task to the worker. One example would be an adjustable chair. Another example would be a human-paced assembly line (i.e., with buffers) so all workers could work at individual paces.

3. Make search of the solution space.

Now design a number of alternatives—not just one.

Now use the information you gathered using the techniques of Section 3. Also get design ideas from a number of sources: workers, supervisors, staff people, other engineers, vendors, suppliers, and so on. (Benchmarking is a technique of obtaining ideas from other organizations.) The solution space will be reduced by various economic, political, aesthetic, and legal constraints. Among the feasible solutions (solutions that work), try to select the best one—the optimum solution.

A problem is the tendency of designers to be satisfiers rather than optimizers. That is, designers tend to stop as soon as they have one feasible solution. For example, when designing an assembly line, the designer may stop as soon as there is a solution. For an optimum solution, there must be a number of alternatives to select from. Alternatives tend to suggest further alternatives, so stopping too soon can limit solution quality and acceptance.

4. Evaluate alternatives.

To get better data for your evaluation, consider trying out your alternatives with mockups, dry runs, pilot experiments, and simulations.

You will need to trade off multiple criteria—usually without any satisfactory tradeoff values. For example, one design of an assembly line may require 0.11 min / unit while another design may require 0.10 min / unit; however, the first design gives more job satisfaction to the workers. How do you quantify job satisfaction? Even if you can put a numerical value on it, how many ‘‘satisfaction units’’ equal a 10% increase in assembly labor cost?

Consider a numerical ranking, using a equal interval scale for each criterion. (Method A requires 1.1 min / unit, while method B requires 1.0 min / unit; method A requires 50 m2 of floor space, while method B requires 40 m2.) However, managers generally want to combine the criteria. Table 23 shows one approach. After completing the evaluation, have the affected people sign off on the evaluation form. Then go back and select features from the alternatives to get an improved set of designs.

5. Specify and sell solution.

Your abstract concept must be translated into nuts and bolts—detailed specifications. Then you must convince the decision makers to accept the proposal. A key to acceptance is input

from others (especially users and decision makers) early in the design stages (steps 1, 2, and 3). At the meeting where the final proposal is presented, there should be no surprises; the participants should feel you are presenting material they have already seen, commented on, and approved. That is, the selling is done early, not late; if they did not buy your approach then, you have modified it until it is acceptable. One common modification is partial change instead of full change; a test market approach instead of immediate national rollout; change of some of the machines in a department instead of all the machines; a partial loaf rather than a full loaf.

The installation needs to be planned in detail (who does what when). This is a relatively straightforward, though time-consuming, process. The installation plan typically is approved at a second meeting. It also should have preapproval by everyone before the decision meeting. During the installation itself, be flexible—improvements may become apparent from the sug- gestions of supervisors, operators, skilled-trades workers, and so on.

Finally, document the results. A documentary video and photos give opportunities to praise all the contributors (and may even reduce resistance to your next project).