The experience of the Anticipatory Failure Determination (AFD) method applied to an Engine Concern

	Bookmark This Page Email This Page Format for Printing

First presented at the Altshuller Institute TRIZCON2000, May 1, 2000

Vladimir Proseanic (Ideation Int.)
Dmitry Tananko (Ford Motor Co.)
Svetlana Visnepolschi (Ideation Int.)
dtananko@ford.com

Abstract

The paper discusses the results of the AFD application to solving a Hitching/Ringing concern on an Engine. The AFD approach has been used by a joint Task Force team to identify the root cause of this Hitching/Ringing phenomena.

Using the AFD, the team was able to identify nearly all possible hypotheses of the Hitching/Ringing problem mechanisms and prioritize their verification process. Employing the AFD methodology allowed the team to greatly reduce the brainstorming time, and increase the efficiency of the team in subsequent testing and verification process.

In addition, the AFD method helped to improve the team’s understanding of the software and calibration and also their interaction with mechanical part of the Engine. It is also explained why the problems of this kind are so difficult to solve by means of other methods (FMEA, in particular) and demonstrated how specific steps of the AFD overcome these obstacles.

Problem description and why it was necessary to use the AFD

The problem is related to slow oscillation of the whole powertrain of the vehicle under steady pedal position (Ringing) or cruise control condition (Hitching). The subjective feeling is as if the vehicle towing a trailer which performs unexpected movements (change of the load), that is where the name of the phenomenon – Hitching, come from. The similar effect, which is usually observed during idle, has called Ringing. These problems sometime occur in the development of different types of cars.

According to the AFD theory, spending sufficient time and money is typical for problems related to failure with unclear mechanisms - Failure Analysis (FA). It is connected to the evident weakness of the traditional approaches in FA problem solving.

There are two points usually recommended for revealing the root causes of a negative event:

·        application of personal professional experience;

·        utilization of the worldwide knowledge, related to a similar problem.

Utilization of the personal experience is limited by similar problems one had experience with in the past. We should live in a really conservative environment to dispense this information.

When we try to utilize professional knowledge of other people, related to the solving FA problems, we meet a real obstacle, also known as Denial Phenomenon [1]. The human beings are subject to a psychological phenomenon called "denial" due to which we resist to the thinking about unpleasant things.  We say: "it can't happen here," "it will come out all right," "it's never happened before," etc.

In response to the question  "What can go wrong?" in respect to the functioning system or operational process, we would like to look worldwide for the records of the problems that have occurred during the past history of similar systems and processes. Unfortunately, the recorded databases are relatively meager – nobody is happy to share this information with anybody. People are not always anxious to document or publicize issues.

Therefore:

Using traditional methods to reveal the causes of an issue, we always deal with deficiency of the information, and that, sequentially, hinders the solution.

The AFD method is capable of overcoming the Denial Phenomenon and, therefore, it could provide required information of the possible root causes for the Hitching/Ringing issues.

Brief information on roots of the AFD methodology:

The AFD method may be considered as TRIZ application on the Failure Analysis (FA) and the Failure Prediction (FP) [2]. The authors of the AFD method are B. Zlotin and A. Zusman.

The early 1980’s, the AFD method has been developed and improved through meticulous analysis of the results of the various consulting projects. In the 90’s, AFD software was developed by Ideation International Inc. based on the AFD method.

Application of the AFD key steps to the Ringing/Hitching problems

In addition to the direct analysis of the conditions related to Hitching/Ringing phenomena, the following key steps of the AFD FA mode were applied:

Step 1 – Invert the Problem

Instead of guessing about the possible causes of the issue - slow oscillation of the whole powertrain related to the idle motion of the vehicle - we invert the problem and formulated it in a pro-active way:

It is necessary to produce the slow oscillation of the whole powertrain under the conditions that initiate and/or accompany the oscillation.

By rephrasing the question in this way, we change the attitude to the issue. The inverted question is very useful in counteracting the denial machinery [1].  Instead of asking ourselves the inverted question "How can I provide slow oscillation of the whole powertrain?" we put our attention on the offensive side of the game.  Thus, we engage our creative faculties actively.

Step 2 – Find the method(s) of producing the phenomenon

After the problem is inverted, our attention is automatically diverted from "things that can happen" to "things that can be produced".   Therefore, the next logical step is:

Identify the areas of science, engineering, or even everyday life, where the same phenomenon of oscillation is intentionally created. 

We are directed to a different field, namely method of production.  This is important, because this field is:

·        widely explored;

·        always different from the area where the problem occurs.

Having formulated the inverted problem, we can search the Internet and the patent library, or interview subject matter expert in the corresponding technology.  By utilizing these new possibilities, we can usually get an exhaustive set of "standard" ways of producing the desired phenomenon (oscillation, in particular). 

The inverted formulation of the problem becomes an inventive problem (“How can I do [something]?” or “How can I cause [something] to occur?”).  This conversion allows us to employ the inventive apparatus of TRIZ, as well as knowledge base and analytical methods.

Step 3 - Utilize Resources

The utilization of TRIZ resources in the AFD process is based on the following postulate:

For any issue or drawback, all the necessary components of the issue mechanism must be present within the system or its immediate environment as available resources.

Therefore, in order to solve the problem of a slow oscillation of the whole powertrain it is sufficient to:

1.      invert the initial problem;

2.      identify all standard ways of creating the slow oscillation;

3.      verify that all resources required to produce the concern are present in the corresponding vehicle.  

Utilization of resources directs us to determine if all necessary components are present in our system as available Resources. 

Upon listing all available resources in the system, the most like hypothesis can be carried out. In majority, if all resources within the system or its environment support the considered hypothesis, the problem is solved. 

In case when the considered method has almost all resources within the system except one or two, we can try to employ them. Keeping in mind the requirement that mechanism must be spontaneous, we cannot use any objects from outside of the system. So, the simple rule has to be followed:

The new resource should be produced from those that are already available in the system or its environment.

In spite of the complexity of generating the hypothesis, this method completely reflects the main idea of resources utilization, and it gives us a way to successfully identify the most tricky and subtle root cause.

The case of slow oscillation in idle mode of the vehicle (Ringing phenomenon) is a unique problem. We came up with the following list of methods for providing oscillation, Hypotheses:

After identifying the components required for each Hypothesis realization and comparing them with system resources, we came up with the following conclusion:

For almost all methods of creation slow oscillation related to idle motion of the vehicle, all resources existed within the system.

It means that all available hypothetical mechanisms were likely to contribute to the issue.

It is not reasonable, however, to try to eliminate all the revealed mechanisms, because:

·        some of them might be quite weak and their elimination could not solve the problem;

·        all of them are closely connected with the key elements of the vehicle, changing which might create even worse problems.

In order to select the Hypothesis for further elimination, we have to identify which particular Hypothesis could provide critical contribution into the issue. 

How the intermediate steps of the FA process increase the advantages of the method.

We have shown above how the key steps of the AFD FA had been applied to the Hitching/Ringing problem. It could be used, if:

·        a quick answer is needed (in emergency situation or for preliminary estimation of possible solutions);

·        the problem is not complicated.

Problems that are as difficult as ours require the complete FA procedure including several intermediate steps. We would further focus on these steps that had played the main part in selecting the right Hypothesis.

Completing the FA Questionnaire and building cause-effect diagrams

 Before the automatic formulation of the inverted problem, we have to answer several questions in order to get the complete picture of the problem, in particular:

1.      What set of closely connected elements, within which the issue takes place, can be referred to as a system. In our case it was an Engine that included:

1.Cylinders system;

2.Injectors;

3.PCM (Powertrain Control Module);

4.High Pressure Oil Pump;

5.IPR (Injector Pressure Regulator);

6.ICP (Injector Control Pressure);

7. Camp-sensor (RPM sensor)

e needed to know these elements as they usually provide resources for the problem mechanism.

2.      How to describe the system functioning. For this purpose, we have built diagrams that represent existing useful functions of the Engine, cause/effect relations between them and their side effects.

The combined verbal and graphic analysis of the engine system gave us the possibility to indicate “main participants” of the problem and their real (not formal) relations.

The diagrams also served as a base for further automatic problem formulation.

Localizing the Issue

At this point, we tried to identify and focus on particular time and place where the issue (low oscillation related to idle) happened.

For this purpose, we had to go through the following milestone:

1.      Identify the Last Event (i.e., the system function or operation during or immediately after which the issue appears);

2.      Describe the conditions that initiate or accompany the issue (i.e., the specific parameters of the system or its nearby environment which correlate with the appearance of the issue.)

We understood that Ringing/Hitching possibly happens after turning to idle/cruise control. We also realized that in both Ringing and Hitching problems the issue takes place upon one indispensable condition - absence of manual control.

When the localizing step was carefully fulfilled the inverted formulation for the Ringing problem looked as follows:

1. Find a way to provide [the] (Oscillation of RPM) with the help of [the] (Turning to idle), (Absence of manual control), and any other function, which happens before the Last Event.

Verifying the Hypothesis

The information obtained from the previous step became helpful for verifying the Hypothesis.

There was only one Hypothesis that could satisfy both verification criteria:

1.      Define all the components necessary for this mechanism to be utilized in the system as present available resources;

2.      Provide the complete solution to the Inverted Problem.

It was the H2: “Providing instability in controlling system.”

The calibration system (control system) that is designed to compensate any undesired signal deviation, fails to do that. Instead, it amplifies the initial oscillation with inadequate time delay of the controlling signal thus causing Ringing or Hitching. (Both phenomena show themselves when the system of engine control works without human participation in idle mode or cruise control.)

The tests of feedback and control signals in idle (Ringing) or cruise control (Hitching) modes had verified the mechanism of the H2 as possible root cause of the slow oscillation of the whole powertrain.

This plot shows three main signals of the control system: actual RPM, filtered RPM, and a command signal. As one can see, the command signal is completely out of phase with filtered RPM due to efforts of keeping constant speed in cruise control mode. Actual RPM is delayed from the command signal due to the mass inertia. Moreover, filtered RPM is delayed from actual RPM due to filtering calculation. The specific combination of these delays, or, in other words, combination of unified control system with individual characteristics of the particular engine, produces Hitching phenomenon. From the control system point of view, this is a classical instability problem due to the delay in the system feedback loop.

Discussion

All the possible Hypotheses for the Hitching/Ringing mechanisms were identified through existing “standard” ways of producing these phenomena. That provided the complete accuracy of the Failure Analysis (no possible mechanism could be missed).

The provided AFD FA gave us the possibility to verify the following mechanism of the Ringing/Hitching phenomena (Verified Hypothesis):

1.      The mechanical system of the engine includes many elements that can provide non-stability, however, none of revealed factors can be estimated as sufficient to be considered as a “trigger” of oscillation. Therefore, improvement of the system’s specific elements cannot solve the problem in general.

2.      The calibration system (control system) that is designed to compensate any undesired signal deviation fails to do that, instead it amplifies the initial oscillation with inadequate time delay of the control signal thus providing Ringing or Hitching. (Both phenomena show themselves when the system of engine control works without human participation in idle or cruise control mode.)

The Task Force team greatly reduced the brainstorming time and increased the efficiency of the team in the following testing and verification process. The whole procedure took one month.

AFD method helped to improve team’s understanding of the truck software and calibration, and also their interaction with mechanical part of the Engine.

REFERENCES

1.      Kaplan, S., Visnepolschi, S., Zlotin, B., Zusman, A. 1999. New Tools for Failure and Risk Analysis. Southfield MI: Ideation International Inc.

2.      Altshuller, G.S. 1984. Creativity as an Exact Science. Translated by Williams, S. NY: Gordon and Breach Science Publishers.

3.      Ternenco, J., Zusman, A., Zlotin, B. 1998. Systematic Innovation: An Introduction to TRIZ (Theory of Inventive Problem Solving). Boca Raton: SRC Press.

4.      Ideation Int. Inc., Final Report on the project “Revealing root causes of Ringing/Hitching phenomena on the truck engine”, 1999.