Anticipating Failures with Substance-Field Inversion - A TRIZ Methods Case Study
Editor | On 06, Mar 2003
By: Thomas W. Ruhe
Anticipating failures is a powerful tool for shortening the time required to develop robust products. Inverting Standard Techniques of Substance-Field analysis allows anticipation of failures caused by a useful function not being performed. Value Definition and Product Trade-offs
â€œValue can only be defined by the ultimate customer. And itâ€™s only meaningful when expressed in terms of a specific product . . . which meets the customerâ€™s needs at the right time at the right price. Value is created by producers â€“ from the customer standpoint, this is why producers exist.â€
In this age of global competition and Internet marketplaces, the only certainty for producers is that customers define â€“ and rapidly redefine â€“ value. The producer with the product that best meets the customer needs first has market advantage. Producers commonly model customer and producer value as having three aspects:
It is common to treat the producer model as having three conflicting goals, producing a trio of â€œTyranny of the ORâ€ problems:
â€œYou can have lower cost or shorter schedules.â€
â€œYou can have shorter schedules or higher quality.â€
â€œYou can have higher quality or lower cost.â€
Using this model, producers work to drive product costs down while, at the same time, they shorten schedules to reduce development costs. The result is often unsatisfactory because of uncertainty how trade-offs of goals affect customer perception of value. This uncertainty isnâ€™t necessary. In the 1950â€™s, Lawrence Miles laid the groundwork for customer perception of value as the relationship of benefits to costs.
Applying this view to the Cost-Quality-Schedule triangle would define producer value in a way that clarifies the relationships, offering options beyond â€œTyranny of the ORâ€:
This model supports the producer habit of reducing costs and shortening schedules.Further, it clarifies that increasing function improves producer â€“ and customer â€“ value.
Defect Discovery and Time to Market
Producers increase function, either by new or revised design, through a product development process. A critical part of the product development process is proving the reliability of the design.
Traditional development processes rely on completing passes through Design-Prototype-Test- Analyze (DPTA) process cycles to discover and correct defects (Figure 1). The effectiveness of this empirical method is in direct proportion to how many times the cycle is completed. However, the time available to complete trips through the cycle is limited by ordinary delays and pressure to shorten the schedule. The limitation is worse if the process cycle relies on testing of the complete system. Because the time for each subsystem to reach stability depends on its complexity, the first few turns around the test cycle at the system level are often disabled by failures in just one subsystem. This inherently random defect discovery process makes time-to-market unpredictable.
Reducing Time to Market by Anticipating Defects
An effective way to improve functional quality and shorten schedules is to anticipate failures during the design process (Figure 2). By inserting failure anticipation cycles, the number of full DPTA process cycles required to achieve functional goals is reduced.
Many Risk Analysis tools exist, for example Failure Modes and Effects (FMEA), Fault Tree (FT) and Event Tree (ET) and Hazard and Operations (HAZOP) Analyses. Each of these, by predicting and relating initiating events, midstates and end states, attempts to produce a severity-ranked listing of all the undesirable phenomena possible for a system. For example, FMEA assumes failure modes of the individual system components as initiating events, and then attempts to create the mid- and end-state consequences. It is, like most Risk Analysis tools, answering the question â€œGiven a known state or event, what can go wrong?â€ Failure anticipation based on the Theory of Inventive Problem Solving (TRIZ) takes a fundamentally different approach. TRIZ failure anticipation seeks to answer the question, â€œGiven the system, how can I most effectively invent failures?â€ From the original application for failure determination by V.V. Mitrofanov, it has been extended to Subversion Analysis by Boris Zlotin, and to Anticipatory Failure Determination by Zlotin, et. al.4
Basic Function, Substance-Field Analysis and Failures
Dr. Taguchi elegantly illustrates that the basic function of an engineered system is to transform input energy to an output function in the â€œP-modelâ€ diagram (Figure 3).
For a system to perform its basic function, the transfer of input energy to output function must be completed within useful limits of magnitude and duration. Conversely, if a useful function is not completed, then a failure may occur.
G.S. Altshuller originated the concept of Substance-Field Analysis for systems.5 The basic Substance-Field model is a function pair (Figure 4).
In this model, S2 is the tool, component or substance transferring an action to an object S1. The field represents the energy transfer between the substances, components or systems. Example field types are mechanical, electromagnetic, gravitational, thermal and strong nuclear. Again, if the energy transfer is completed within useful limits of magnitude and duration, the useful function is performed.
By creating situations in which the useful function is not performed, Substance-Field Analysis may be used as a failure prediction tool.
TRIZ Tools â€“ Standard Solutions and Standard Techniques
The TRIZ toolkit for problem solution includes Altshullerâ€™s 76 Standard Solutions for Inventive Problems.6 The Standard Techniques are a further development of the Standard Solutions.7 One group of these contains six techniques for directly eliminating the effect of a harmful action (Table 1)
Each of these techniques presents a Substance-Field solution model for one case of harmful energy transfer producing a harmful function. The standard knowledge bases, databases and solution tools of TRIZ may be used to creatively generate solution concepts
Substance-Field Inverse Analysis
As stated previously, it is common in TRIZ to use the â€œother way aroundâ€ principle to invert problem statements and solution techniques for failure anticipation. One way is simply to invert the objective. That is, a harmful function is treated as useful. The objective then is to magnify that harmful function, its effects and to invent new harmful functions.
Another method is to invert one or more of the TRIZ tools. The method used here inverts a set of Standard Techniques by replacing harmful effects with useful effects (Table 2).
Each of these techniques presents a Substance-Field solution model for a way to impair a useful function. In each, if the function is not completed, then a failure may occur. Using these inverted models, the problem now is to create situations for each case that will impair or eliminate the useful action. That is, to invent failures, anticipating them so they may be eliminated. Again, the existing knowledge bases, databases and solution tools of TRIZ are all useful, being inherently designed to solve creative problems.
A trial application of this failure anticipation method was conducted on the paper output system of the DeskJet 990C printer (Figure 5). Customers value real printing speed. The 990C achieved draft black text speed of 17 pages per minute (ppm), a significant jump from the 12ppm of its 970C predecessor. A key enabler of the speed gain was removing the need to perform a full ejection cycle for every page. The required redesign of the media output system could introduce new failure modes. This presented a valuable opportunity to exercise Substance-Field Inverse Analysis.
The 990C media output system appears to be quite simple. A trio of fingers on the pusher moves the media into the output tray. So the customer function of the media output system is to move one sheet into position in the output tray for each ejection cycle. This can be represented as changing pusher position qi into media position yo (Figure 6). However, the energy-level engineering definition of the basic function is to convert input drive torque ti to output media motion, produced by pusher torque to acting on the media (Figure 7).
Beneath this simple basic function, however, is a complex support and cam drive system to actuate the pushers at only the desired time. To be sure that the failure anticipation would focus on the substance-field pair with the highest likelihood to introduce new failure modes, a functional diagram was constructed (Figure 8), including super-system resources and control system actuation algorithms.
With the Substance-Field system of interest selected, the goal is to generate solutions to the inverse model cases. While these may be found by simple system inspection, TRIZ solution resources may identify richer possibilities. Examples of anticipated failure modes for the pusher system are shown in the following tables:
The new failure mode predicted by the concepts in Case 5 was demonstrated in testing. The output media encountered resistance from the other media already in the output tray. The pusher force then created a force couple resulting in media buckling, rather than moving forward into the tray. Several solution concepts were generated, with final modifications reducing the occurrence of these failures to a negligible level. Thus, the case study demonstrates how inverting one group of Standard Techniques allows the use of standard TRIZ creative tools to anticipate failures.
Future Method Extensions
The techniques presented only address those cases where the failure is caused by noncompletion of the useful action. Application of inversion to other groups of Standard Techniques would expand the scope of failures anticipated.
Also, no energy transfer is perfectly efficient. Some energy is wasted, perhaps in a harmful form (Figure 10). 10 For example, torque is lost in a bearing interface, dissipated as heat that may damage the bearing or its lubricant. Further, if the noise includes a source of energy, or if a feedback loop is present, the magnitude of the output function can exceed useful levels, itself becoming harmful. The Tacoma Narrows bridge failure is an infamous example of this type of failure.
To mature the capability of Inverse Analysis methods, models that address these energy situations should be developed and tested.
1. James P. Womack, Daniel T. Jones â€œLean Thinkingâ€, Simon & Schuster, New York, 1996; www.lean.org
2. James C. Collins, Jerry I. Porras, Built to Last: Successful Habits of Visionary Companies, HarperCollins, New York, 1997, p. 43.
3. Lawrence D. Miles, Techniques of Value Analysis and Engineering, McGraw-Hill, New York, 1961.
4. B. Zlotin, A. Zusman and S. Visnepol, â€œNew Tools for Failure and Risk Analysis,â€ Southfield, MI, Ideation International Inc., 1999; www.ideationtriz.com
5. G. S. Altshuller, Creativity as an Exact Science, Gordon and Breach Science Publishers, 1984
6. B. Zlotin, A. Zusman G. Altshuller, V. Philatov, â€œTools of Classical TRIZ,â€ Southfield, MI, Ideation International Inc., 1999; www.ideationtriz.com
7. Zinovy Royzen. â€œTool, Object, Product (TOP) Function Analysis,â€ TRIZ Journal, September, 1999, www.the-trizjournal.com
8. Zinovy Royzen, Innovation in Manufacturing using TRIZ, workshop, Seattle, 1997; www.trizconsulting.com
9. Functional diagram and some effects generated using TechOptimizer 3.0, Invention Machine Corp., Boston, 1998; www.invention-machine.com
10. Figure adapted from William E. Eureka, Nancy E. Ryan, editors, Quality Up, Costs Down, ASI Press / Irwin Professional Publishing, New York, 1995; www.amsup.com