A systematic way to solve technical problems

When confronted with a tough engineering problem, do you sit back, scratch your head,and stare into space hoping the answer will just come to you? Or do you apply the first fix that comes to mind? If so, experts in the art of problem solving would say you're going about it all wrong. Instead, they'd recommend a more systematic approach. One such method of solving knotty technical problems is called Triz, the Russian word for trees.

Triz is not new. It's the brainchild of Russian Genrich Altshuller. He's credited with discovering the method after studying thousands of worldwide patents. The system encourages thinking and prodding at a technical problem until one gets to the real trouble that needs fixing.

Triz has been around since the late 1940s but the computer revolution makes the method more practical. It can be self-taught, but as with most technical topics,formal training always helps focus the mind.

Here's an example
The process was applied this way to a production problem: When refining pure copper by an electrolytic process, a small amount of electrolyte remains in pores on the surface of copper sheets. The electrolyte evaporates in storage, creating oxide spots, which detract from the product's appearance and reduce its value.

"From the Triz perspective," says Boris Zlotin, a Triz Master with Ideation International Inc., Southfield, Mich., "the best way to solve the problem is to avoid producing pores at all. But engineers rejected that approach because it required reducing the current that drives the process which, in turn, cuts productivity." The next immediate solution has been to wash the sheets to remove the electrolyte prior to storage. This was labor intensive and pores remained a 15-year focus of improvement for the industry.

The Triz method suggests first stating the problem as a contradiction. In this case, the problem statement could read: The current must be low to avoid pores and high for required productivity. Now it's possible to apply tactics for handling physical contradictions. Zlotin suggests asking: Where (if we are talking about a resolution in space) or when (if we think about possible resolution in time) should the current be low to avoid pores? The answer comes quickly to those familiar with the production process — low current is needed only at the end to prevent pores while for most of the process the current could be high to ensure the required productivity. Because the complete cycle takes about 72 hr, it was found that reducing the current for the last 30 min is enough to produce pore-free copper.

"Two important points should be made here," says Zlotin. "First, it was not a simple problem. Manufacturers were wrestling with it for years using traditional methods, most of which focused on ways to wash off the electrolyte. People working on the problem in a workshop could not believe it took them less than an hour to find an answer. So it's important to emphasize what made the problem easy — breaking psychological inertia by understanding that it is better to eliminate the root causes of the problem, the pores, than fighting the harmful results of poor appearance. So stating the problem as a contradiction proved to be a door to the solution."

The second point is that experienced problem-solving specialists know trade-offs always produce lower expectations. But resolving a contradiction often pays off with unexpected benefits. "For example, while digging around the idea to lower the current at the end of the process, it was found that the current normally used was not optimal from the productivity point of view," says Zlotin. Instead, it was a result of previous trade-offs — a compromise to minimize the pores. With that discovery, the panel recommended increasing the current during the main production period while reducing it at the end, maximizing productivity, and eliminating pores."

The process
The innovation process takes five steps and starts with a brief description of the problem in a format the developers call the innovation situation questionnaire or ISQ. Later steps include problem formulation, identifying a priority direction and idea generation, developing concepts of solutions, and evaluating results.

A general overview in the software leads to the rules for effectively applying the process. It includes tips such as don't discard an idea just because a secondary problem crops up. Also, stop work when you become tired or an idea requires additional information.

For more detail, let's test the Triz system, called Innovation Workbench, on a simple problem. Its statement should include a positive requirement and a drawback. For example, our office workers labor in cubicles or workstations. The cube's small size allows packing more people in the same area of several traditional offices. But with so many people working in close proximity, the noise reaches near intolerable levels. How could this be corrected? The first step, completing the ISQ, organizes your knowledge about the problem. The more you fill it out, the more details you add to the problem and the more you get to the real crux of the problem. Answering the questions in the ISQ helps initiate thinking about the problem from a Triz perspective — in terms of an ideal solution, and useful and harmful functions. In the case of the work cubicles, noise is the harmful function.

The second step helps formulate the problem as a sort of cause-effect diagram that maps the situation in terms of harmful and useful functions (boxes) and their interrelationship. Once the diagram is finished, hitting a light-bulb icon tells the software to reformulate boxes in the chart into recommended directions for potential solutions.

When there were only about six boxes in the chart, the software generated eight basic possible ways to resolve the noise problem. For instance, one recommendation reads: Find a way to eliminate, reduce, or prevent [the] (People have to talk to get their job done) in order to avoid [the] (Work cubes are noisy).

This is where human-engineering evaluations come in. Of course it's not practical to keep people from talking, but the thought has possibilities. For example, let employees work from home or in a quiet area for part of the week. The more you think about the problem and add to the chart, the more avenues of investigation the software offers.

After adding a few more details to the chart, the software came back with 12 avenues. The eighth suggestion has merit and reads: Find a way to eliminate, reduce, or prevent [the] (Work cubes have no soundproofing) in order to avoid [the] (Work cubes are noisy).

To find ways to soundproof a workspace, users can examine the provided list of undesirable conditions. Noise appears in the list. Picking on it pulls up five more suggestions for quelling noise. The first is to use selectively permeable isolation. That sounds promising, so picking on it takes us to several discussions for how noise is quelled in industrial applications. These hint at active noise control as a solution. Such a device generates a negative noise pattern to the one received or heard in each workspace. By working through the 12 suggestions and the technical solutions provided by the software, engineers, or office managers, might find an acceptable solution to noisy offices.

Of course this is not an overly technical challenge. But it shows how the software would be used to formulate the problem and examine potential solutions. The accompanying boxes illustrate two additional technical problems and how they were solved.

If those here are not enough, link over to http://vubme.vuse.vanderbilt.edu/king/student_project_l isting_2000.htm. The list is the result of student projects at Vanderbilt University's medical engineering school. The green projects are password protected because the project findings might be patented.