Chapter 4. Focused Improvement. Part 5
14. P-M Analysis
14.1 The Weaknesses of Traditional Cause Analysis
Cause analysis (fishbone analysis, or the cause-and-effect diagram) is a traditional quality control method that can be used easily and effectively to solve many types of problem. Its weaknesses become apparent, however, when attempts are made to use it to analyse the multiple interacting causes that lie behind chronic problems. These weaknesses, which relate to the way in which potential causes are rounded up, are described below.
(1) The phenomenon is not checked and stratified precisely enough
Although several events or phenomena may seem at first sight to be identical, further observation and analysis often shows that they are not, so it is important to classify the observed phenomena carefully according to the conditions under which they occur. If this is not done thoroughly enough, any corrective action is likely to miss the mark. It is essential to pinpoint each separate phenomenon precisely in order to avoid this.
(2) The phenomenon is not analysed adequately
There is a tendency in conventional cause analysis to list the possible causes that first come to mind, without doing a proper physical analysis of the phenomenon. Failing to identify potential causes systematically, without a sound physical analysis of the phenomenon, leads to irrelevant factors being included and relevant ones being overlooked.
Traditional cause analysis is also often based on an incomplete understanding of the conditions under which the phenomenon can arise. Since the full range of possible situations is not considered, there is an inevitable bias towards a particular set of conditions. The analysis is thus applied to a limited group of possible causes, while others are left out of the picture. The upshot is that chronic losses remain at the same level. Failing to identify all of the conditions giving rise to the phenomenon and all possible relevant factors is a serious error. Conventional cause analysis often short- circuits the process by jumping ahead to corrective action before the phenomenon has been properly analysed.
14.2 What is P-M Analysis?
The technique known as P-M Analysis was developed to compensate for these deficiencies in conventional cause analysis. In P-M Analysis, the P and M do not stand for preventive or productive maintenance. As Figure 4.35 shows, the P stands for phenomenon and physical, while the M stands for mechanism, machinery, men/women, materials and methods. It is defined as an approach that elucidates the mechanism of a chronically-occurring undesirable phenomenon by analysing it in terms of the physical principles and parameters involved. Having identified the mechanism by which the chronic phenomenon (a chronic quality defect, breakdown or other loss) occurs, it goes on to list all the possible factors that could logically be thought to affect this mechanism, in terms of the machinery involved, the men and women who operate it, and the materials and methods used. It could also be described as a way of examining a situation analytically and systematically in order to eliminate a chronic undesirable phenomenon by reviewing the causal system, analysing the causes and identifying and eliminating all relevant defects. The point is that P-M Analysis is not just another improvement tool; it is an entirely new way of examining and interpreting a situation.
14.3 Target-Setting: The Basic Approach to Improvement
The first thing one usually notices when visiting companies that have not adopted TPM, and hearing presentations on improvements they have made, is how low their targets have been set. This is true whatever the industry, size of the company, or type of operation. Typical examples would be a 30% reduction in quality defects, or a 20% reduction in changeover time, with the most ambitious targets being 50% reductions. Another common feature is jumping straight from the problem to the solution. It is often unclear why the particular solution was adopted, what the causes of the problem were, and what thought processes led to the solution.
To maximise the cost-benefits, improvement topics should be addressed in the order of the largest financial losses – that is, starting with the item on the far left of the Pareto chart. The projects should be prioritised in this way (see Figure 4.36). However, it is important not to be satisfied with a conventional target of, say, a 30% reduction, but to aim to get rid of the problem entirely (i.e. to adopt the zero-focused approach). People may well start off by thinking that zero targets are unattainable, but many companies have demonstrated that they can in fact be achieved surprisingly often through the correct use of P-M Analysis.
14.4 Good P-M Analysis
A good P-M Analysis satisfies the following three conditions: (1) It follows the correct thought process
There must be a logical thread running from the phenomenon via physical analysis through contributing conditions, then 4-M correlations, to the phenomenon’s possible causes. This point requires careful attention, because many P-M Analyses include lists of factors that are irrelevant to the phenomenon or bear no causal relationship to what has gone before (for example, when secondary 4-M correlations are mixed up with primary 4-M correlations). Of course, care must also be taken not to omit any of the potential causes.
(2) The fault-finding process is correct
- The search for faults must be rigorous and uncompromising
- Measurement techniques must be appropriate (measuring instruments precise enough for the tolerances, properly chosen datum points, and the right measurement methods)
- Tolerances must be set clearly and correctly
- The correct restoration and improvement techniques must be used
(3) The results are satisfactory
If the chronic undesirable phenomenon still remains even if the thought process and the process of identifying faults were correct, it shows that there is some problem with the application of the technique.
14.5 The Definition of P-M Analysis
P-M Analysis is a step-by-step procedure designed to eliminate chronic losses. Its definition, and the approach it adopts to eliminating chronic losses, are:
The definition of P-M Analysis: An approach in which an undesirable phenomenon occurring in an operation is physically analysed in accordance with the operation’s principles and parameters in order to elucidate the mechanism by which the phenomenon arises.
The idea of eliminating chronic losses: List all the factors that could theoretically be considered to affect the phenomenon; check them all without regard to how large or small their contribution might be, and take action against all the discrepancies discovered.
Figure 4.37 shows the eight steps of P-M Analysis, with a brief description of each.
14.6 The Structure of P-M Analysis
As Figure 4.38 shows, the structure of P-M Analysis is like a branching tree. The sections of the diagram from left to right correspond to the first four P-M Analysis Steps (Step 1 – Clarify the phenomenon, Step 2 – Perform a physical analysis of the phenomenon, Step 3 – Identify the phenomenon’s contributing conditions, and Step 4 – Examine the relationships with the 4Ms). Although the diagram only goes as far as the secondary 4-M correlations (which is usually all that is necessary), it can be extended as far as the tertiary and quaternary correlations if required.
14.7 Where to start and finish a P-M Analysis
People tend to want to begin a P-M Analysis on whatever they have identified as a problem from examining the routine production data. However, as Figure 4.39 shows, it is important to identify the specific reasons for the problem first (using Why-Why Analysis) and then take one of these reasons as a starting point for the P-M Analysis. The P-M Analysis should be stopped when it has reached the level of the deterioration of individual components (such as the wearing-out of a shaft, or the loss of strength of a spring). As mentioned above, this will usually be at the secondary 4-M correlation level, although it may be necessary to go further in some cases. Once the deterioration modes have been identified, they should be then surveyed to find out if they are actually happening. The ones that are actually happening should then be investigated further to determine why the people looking after the machines had not prevented them, and measures should be put in place to make sure they are prevented in future.
14.8 The Steps of P-M Analysis
14.8.1 Step 1: Clarify the phenomenon
A ‘phenomenon’ is an observable event or occurrence, and ‘clarifying the phenomenon’ means identifying what is causing the problem by careful observation of the facts. If the phenomenon (what is actually happening) can be pinpointed, the problem is half-way to being solved. Conversely, a P-M Analysis conducted on a phenomenon that has not been accurately identified will miss the mark and be merely a waste of time.
The important thing when observing the phenomenon is to actually see it for yourself, without relying on what other people have told you. The problem should be carefully observed, and what actually happens, the way in which it happens, its location on the machine, any differences between one machine and similar ones, and other factors should be examined to see if exactly the same phenomenon is happening every time, or whether there are any differences. Analysing the phenomenon in this way is called ‘stratifying’. The key point is to analyse the problem carefully and objectively, without jumping to conclusions. Serious errors can result from making judgements without having sufficiently stratified the phenomenon.
It may be possible to observe a phenomenon as it is actually happening (catching the criminal in the act, so to speak). Often, however, only the end result can be seen (like turning up at the scene after the crime has been committed). The former is often possible with minor stops, while the latter is more common in the case of breakdowns and quality defects.
When a phenomenon happens very rapidly, it is extremely effective to film it with a high-speed video camera and then watch the video in slow motion to see exactly what is taking place. This is a good technique to use with minor stops, which are mainly caused by products catching or jamming somewhere in a machine, and with some quality defects.
Even if the phenomenon cannot be observed as it is happening, there is usually sufficient evidence left behind to formulate a hypothesis about its causes – if a machine has broken down, for example, the broken part will always show signs of having received energy. It is important to infer from such signs what kind of energy was applied to the part, and from what direction, to cause it to break. The broken surfaces should be examined under a magnifying glass or microscope to determine their condition and structure, and the part should be kept until the cause of the problem has been tracked down.
The causes of quality problems can also sometimes be inferred by examining the defective product under a microscope or with measuring instruments, particularly if the problem relates to the product’s shape. In other cases, it is important to observe the phenomenon from the 5W1H standpoint, trying to identify whether its occurrence depends on the machine or lot, the operator, the season, the day of the week, the time of day (first thing in the morning, just after a changeover, etc.) and so on, and whether there are any identifiable cyclic or time-based patterns in the way it happens.
Figure 4.40 summarises what needs to be done in order to clarify the phenomenon.
‘Clarifying the phenomenon’ means understanding the phenomenon precisely, observing exactly how it occurs, and knowing what patterns it exhibits. To ‘stratify’ it effectively (i.e. to determine what factors its occurrence depends on, and how its occurrence depends on those factors), the following key actions should be taken:
(1) Observe the facts carefully, using the ‘3 Actuals’ approach (actual location, actual objects and actual facts).
(2) Use the 5W1H approach (What? Where? When? Which? Who? How?) to stratify the phenomenon until it can be stratified no more.
(3) Compare the normal situation (or good product) with the abnormal situation (or defective product) and identify all significant differences.
14.8.2 Step 2: Perform a physical analysis of the phenomenon
All phenomena can be explained in terms of physical principles and parameters. Physically analysing a phenomenon means examining it in order to establish the physical principles behind it. For example, if an object has become scratched, it must have brushed against or collided with a harder object. Thus, if a product keeps getting scratched during a manufacturing process, looking for places where it could come into contact with a harder object will indicate which parts of the machine need to be investigated and make it easier to identify the causes. This step is crucial for the following reasons:
- A sound physical analysis of the phenomenon will reveal all the relevant potential causes.
- Since it is logical and systematic, it ensures that no potential causes are overlooked.
- It prevents conclusions being drawn on the basis of intuition and guesswork.Before conducting a physical analysis, however, there are two important sub-steps that must be completed – understand the mechanism, and identify the principles and parameters. These will now be described.
(1) Understand the Mechanism
The key to success when analysing the potential causes of an undesirable phenomenon is to start by learning about the equipment where the phenomenon occurs – how it works, and how it is constructed. If this is well understood, the mechanism by which the phenomenon occurs will naturally become apparent. Despite this, however, many people jump straight into investigating the causes of problems without first going through this step.
To find out how the equipment does its job, and how it is put together, start by doing a three-dimensional diagram like that shown in Figure 4.41. Mark out the equipment’s functional units with dotted lines, name the units, and write a brief description of what each does. Next, make a list of all the individual components each unit is assembled from, with their names and functions. Doing this should make it clear that every component has a role to play and none is superfluous. It also becomes apparent what kinds of problem would happen if any of the components were not maintained to the required degree of precision – useful knowledge for establishing zero-defect conditions in Quality Maintenance. It has been demonstrated time and again that a sound knowledge of the structure and functions of the equipment and its component parts makes the subsequent investigation of the causes of problems far quicker, more accurate and more reliable.
A word of warning – do not confuse a sketch of the outside of a machine with the three-dimensional diagram required here. The drawing required here – called a ‘mechanism diagram’ – should clearly show how the parts fit together, how the motive force is transmitted, and how the machine actually does what it is supposed to. It is also essential to hand-draw the diagram yourself. Photographs or copies of manufacturers’ drawings should never be used instead. Remember that the point is not to produce a drawing but to use the act of drawing as a highly effective means of learning all about the way the equipment is constructed and how it works. Figure 4.42 lists the key points to bear in mind at this stage.
(2) Understand the principles and parameters
Once the mechanism and construction of the equipment have been understood, the next important step is to sort out the principles and parameters, as defined in Figure 4.43. All too often, people operate and maintain machines without understanding the principles by which those machines operate or the parameters under which they must be operated, but this makes it almost impossible to understand the mechanisms by which quality defects and other problems are produced. It is essential to understand the principles and parameters in order to conduct a physical analysis and establish the contributing conditions.
To illustrate what is meant by principles and parameters, let us consider an ordinary battery-powered torch. The principle by which the torch lights up could be described as ‘An electric current is made to pass through a resistance wire, heating it up to 1,000°C, the temperature at which it emits light’ (see Figure 4.44). The parameters (conditions) that must be in place in order for this principle to operate are (1) The battery produces a certain minimum EMF, (2) There is adequate contact between the battery and the bulb’s filament (i.e. no breaks in the circuit) (3) The bulb is in good condition and of the specified rating.
If the torch failed to light when switched on, the physical analysis of the phenomenon would be that ‘Not enough electric current to cause light emission flowed between the battery and the bulb’, and the ‘contributing conditions’ (any condition that would give rise to the situation described by the physical analysis) would be the failure of the parameters described earlier.
Parameters can be thought of as the minimum conditions necessary in order for the principle to operate correctly. They can most easily be identified as the minimum conditions that must exist for each of the functional units of the system to do its job. In the case of a torch, for example, the functional units are the battery, the circuit wire and the bulb.
(3) Conduct a physical analysis
Conducting a physical analysis means logically establishing how an event would take place. It consists of elucidating the mechanism by which the phenomenon occurs, via a change in a particular physical quantity, based on the principles and parameters already identified. It may in fact be easier to think of it as analysing the physical quantity involved. To help explain what it means, let us take a certain type of machining operation as an example.
To process an object or material is to change its shape or nature, which requires the application of some form of energy to the object or material. In fact, processing is no more than the application of a defined amount of energy at a defined position.
Incidentally, the amount of energy applied is given by the following equation:
Amount of energy applied = rate of application of energy (force, heat, electric current, etc.) x time
In a machining operation, energy is applied to a workpiece through the point of contact between the workpiece and a cutting tool. This point of contact is known as the ‘processing point’. This means that there would be no possibility of a defective product being produced if the position of the processing point and the amount of energy being supplied to the workpiece via the cutting tool remained correct. A defective product is always the result of a deviation in the position of the processing point or the amount of energy from what it should be.
Therefore, to consistently make good products, these two things (the position of the processing point and the amount of energy supplied) must be kept correct at all times. The physical analysis of a problem phenomenon is a description of how this is not happening.
An appropriate physical analysis of a phenomenon can usually be arrived at by following the procedure shown in Figure 4.45.
Figure 4.45 Step 2: Physical Analysis
1) Locate the processing point (the point at which energy is transferred).
2) Sketch the processing point.
3) Clarify which object is giving the energy (A) and which object is receiving the energy and thereby changing its shape and/or characteristics (B).
4) Clarify the physical, quantitative parameter (C) that governs the change in shape and/or characteristics. Physical parameters include those related to position of the processing point, such as distance and angle, and those related to the quantity of energy, such as force, speed, acceleration, temperature, electrical current, voltage and time.
5) Clarify the change (D) in the physical parameter.
6) Express this relationship in writing, typically in the form ‘A and B cause C to change D’.
7) If the occurrence of the phenomenon can be expressed in terms of an equation or formula (e.g. a reaction formula), this itself can be used as the physical analysis.
Let us explain this in more detail by looking at a real-life example, that of a metal cylinder being machined to size with a precision grinding wheel (see Figure 4.46), where the undesirable phenomenon is that the finished outside diameter of the cylinder varies. The object imparting the energy (A) is the grindstone, and the object that receives the energy and changes shape is the cylinder (B). The principle of operation of a cylindrical grinding wheel could be expressed as ‘Abrasive grains (N.B. these are equivalent to cutting tools) in a grindstone rotating at high speed bite into the surface of a revolving workpiece, shaving off its surface in the form of filings. By cutting continuously into the workpiece, the grindstone finishes it to the specified shape and dimensions’.
In an actual grinding operation, the grindstone works in cycles, first moving forward (while grinding the workpiece) to a pre-set foremost position, then halting briefly, then retreating to its starting position. This means that the physical parameter C that determines the outside diameter of the cylinder, as the grindstone cuts, is the distance between the surface of the grindstone and the centre of rotation of the cylinder.
The physical analysis of the problem phenomenon (that the outside diameter of the cylinder varies) can therefore be expressed as, ‘The distance (C) between the surface of the grindstone (A) in its foremost position and the rotational centre of the cylinder (B) exhibits variance (D)’.
It should be noted here that each distinct phenomenon can only have one physical analysis. If you see more than one physical analysis for a phenomenon, it means that the phenomenon has not been stratified in sufficient detail, and that Step 1 (Clarify the phenomenon) needs to be repeated. Also, if it is difficult to express the physical analysis in words, a diagram, mathematical formula, reaction equation, etc. can be used instead.
14.8.3 Step 3: Identify the phenomenon’s contributing conditions
In this step, we use the underlying physical principles and parameters to identify all the conditions (the ‘contributing conditions’) that will inevitably produce the phenomenon if they are present. The key point here is to identify all of these conditions. The phenomenon is physically analysed, the mechanism by which it occurs is elucidated, the conditions that logically must give rise to it are identified, and all conceivable situations in which it could arise are logically worked out, without relying on experience, subjective judgement or gut feelings.
In conventional cause analysis, only some of the many contributing conditions are taken into account. Causes are sought for them, and corrective action is taken, but the losses often do not decline, because other relevant conditions have been ignored. It is important to do a comprehensive analysis that does not exclude anything just because it is thought unlikely to have much bearing on the problem.
It was explained earlier that parameters are the minimum conditions necessary in order for the principle to operate correctly, and that they are the minimum conditions that must exist for each of the functional units of the system to do its job. In many cases, the contributing conditions that give rise to a particular phenomenon consist of insufficient accuracy in the maintenance of these parameters.
As a guideline, when using P-M Analysis to unravel the chain of causes behind a phenomenon, contributing conditions can be thought of as operating at the level of the system’s functional units (assemblies that perform a single defined function), primary 4- M correlations as operating at the level of sub-assemblies (smaller assemblies that compose the functional units) and secondary 4-M correlations as operating at the level of individual components (see Figure 4.47).
Also, when listing the contributing conditions, the situation should be considered in the light of each of the 4-Ms (machines, methods, men/women, materials). Start by considering the mechanical precision of the equipment, then go on to consider methods (inadequate processing conditions, standards, etc.), people (not observing the standards) and materials (quality from the previous process). Figure 4.48 shows the specific procedure in more detail.
Let us continue with the grinding-wheel example to illustrate in more concrete terms how to identify the contributing conditions. As we have seen, the physical analysis of the phenomenon is that ‘The distance (C) between the surface of the grindstone (A) in its foremost position and the rotational centre of the cylinder (B) exhibits variance (D)’ (see Figure 4.49). This sentence contains two conditions: ‘the surface of the grindstone (A) at its foremost position’, and ‘the centre of rotation of the cylinder (B)’. These two conditions can be taken as the contributing conditions just as they are. All we need to do is to express in words what must happen to them in order to produce the variation in the physical quantity identified in the physical analysis (‘the distance C exhibits variance’). Quite simply, the two contributing conditions are:
(1) The surface of the grindstone (A) at its foremost position varies (2) The centre of rotation of the cylinder (B) varies
Next, we look at the other functional units that compose the grinding wheel to see whether any change in their working would produce a change in the physical quantity. There are in fact two more functional units, the grindstone dressing unit and the cutting compensation feed unit, and examination shows that each is associated with one of the following two contributing conditions:
(3) The dressing volume varies (the dressing unit)
(4) The cutting compensation volume varies (the cutting compensation feed unit)
All four of these contributing conditions relate to equipment precision. However, there are also two others relating to methods:
(5) The cutting movement cycle of the grinding head varies (6) Items in standards are deficient or nor observed
And one other relating to materials:
(7) The quality in the previous process is unstable
Figure 4.50 summarises the key points for identifying contributing conditions, and Figure 4.51 illustrates two common patterns by which contributing conditions arise.
Figure 4.50 Step 3 Key Points in Identifying Contributing Conditions
1) First, take the two conditions identified in the physical analysis (the object that gives energy and the object that receives energy) as contributing conditions.
2) Then look at the remaining functional units to see whether changes in their functioning could produce a change in the physical parameter.
3) Review the processing conditions, work standards etc. to see if any are deficient or not being followed.
4) Examine the materials and the quality of previous processes.
14.8.4 Step 4: Examine the relationships with the 4 Ms
In this step, we examine each of the contributing conditions identified in Step 3 in relation to the ‘4 Ms’ (men/women, machinery, materials and methods), logically identifying all the causal factors that could possibly give rise to them. It is important to identify every possible relevant factor, regardless of whether it has actually caused problems in the past, and to make sure that none is left out, by working back methodically from the processing point to the mechanism that ultimately supports it.
An important point when identifying the 4-M correlations (which does not necessarily apply to the contributing conditions) is to ensure that they are described in such a way as to make them measurable, or at least verifiable. Avoid vague statements that cannot later be verified, such as ‘part X was not properly attached’.
Figure 4.52 summarises the key points to bear in mind when identifying the 4-M correlations.
4.52 Key Points in Identifying 4-M Correlations (P-M Step 4)
1) Scrutinise the mechanism diagram and parts table in order to logically identify all the potential causes.
2) Find the potential causes by dividing the mechanism into what seems to you to be appropriate sub-assemblies, starting at the processing point and working back to whatever it is based on (e.g. a power source).
3) Paying attention to even minor defects (fuguai), list all potential causes regardless of how much you think they contribute to the problem or whether they have actually contributed to a problem before (especially when identifying the secondary 4-M correlations).
4) Check your cause-and-effect logic by working backwards through the steps from right to left (e.g., check whether, if the primary 4-M correlations are happening, they will really produce the associated contributing condition).
5) Always describe the 4-M correlations in words in such a way that they can be measured or inspected.
14.8.5 Step 5: Work out what the ideal state (the standard values and criteria) should be
The next step is to examine each of the possible causes of the phenomenon identified in the previous steps, and establish a criterion for determining whether it is as it should be or not. As users of production machinery, we are very often not provided with any standards by which we can judge whether the equipment is in the condition it should be, in order to produce the required quality level in the product. In fact, at the component level, such standards are virtually non-existent. We therefore need to start by looking at the equipment’s operating manuals and installation records to find out what standards we do have (e.g. for levelness, vibration and other aspects of installation precision, or for parallelism, straightness, shaft runout and other aspects of assembly precision). To establish standards for other factors, we need to obtain detailed parts lists, assembly diagrams and other information, or seek assistance from the machine’s manufacturer.
One practical way of establishing standards where none exist, particularly if there are many similar machines in the plant, is to measure the precision of the best- performing machine (the one with the highest process capability index, for example, or the one that turns out products with the greatest precision) and take that as a tentative standard. When specifying the precision of individual components, it is important to avoid qualitative expressions such as ‘not worn’ and try to use quantitative expressions such as ‘not worn by more than 10 μm’. Figure 4.53 summarises the key points for establishing standards, standard values and inspection criteria for equipment, methods, people and materials.
14.8.6 Step 6: Consider what investigation methods to use
Here, we work out exactly how we are going to investigate and measure each causal factor. We decide what to measure, how to measure it, and what we are going to use as a reference plane if we are measuring dimensions. When doing this, we may find that we have no idea how to measure a particular factor, but there is always someone, somewhere who does, so it is important to cast the net wide. The activity can also give a tremendous boost to the company’s level of metrological expertise. As Figure 4.54 indicates, we need to decide the When?, Who? and How? (or What with?) of carrying out the necessary measurements on all the potential causes of the phenomenon.
14.8.7 Step 7: Identify abnormalities (fuguai)
‘Identifying abnormalities’ means looking at each of the causal factors, comparing their actual values against the standards we established in the previous step, and seeing if there are any deviations. When performing the measurements, we must do everything possible to ensure the results are reliable, e.g. by using instruments whose precision matches the tolerances we are trying to measure (if the tolerance is 0.01 mm, for example, we need an instrument that can measure in 0.002 mm units), and by establishing clear reference surfaces and measurement points. Figure 4.55 shows the key points to bear in mind when looking for deviations from the standards (such deviations are also referred to as ‘deficiencies’, ‘minor defects’, or by the Japanese word ‘fuguai’, which means anything that is not as it should be and might cause a problem).
Figure 4.55 Step 7: Key Points for Identifying Minor Defects (Fuguai)
1) Painstakingly identify every single fuguai by examining the situation from the standpoints of what it should ideally be, and what minor imperfections you can find.
2) Don’t be satisfied with finding just a few major defects. Keep going until you have found all the defects, big and small, regardless of how much they might contribute to the problem.
3) Analyse the fuguai to determine their causes and find out why the less than perfect situation arose. Then use this information to devise improvements that will prevent the fuguai from ever reappearing. Note whether each fuguai is due to natural deterioration or to forced deterioration.
14.8.8 Step 8: Restore or improve; sustain and control
(1) Restoration and improvement
Restoring something means getting it back into its original, correct condition. Production equipment that has deteriorated should always be restored first, before any improvement is attempted. However, improvements will be required if restoration cannot solve the problem (for example, if the machine has a structural weakness or is insufficiently rigid), or if it is technically outdated. Figure 4.56 shows the key points for restoring and improving.
(2) Sustainment and control
No matter how good an improvement may be, it will only produce useful results if it is sustained, and making sure it is locked into place is the only way of securing a firm foundation on which to build the next level of improvement.
Figure 4.57 illustrates the importance of this idea.
To sustain improvements reliably and keep the situation under control, we need to do four basic things: (1) sort out what needs to be controlled (and hence what needs to be periodically checked), (2) consolidate the checks and reduce the time needed to carry them out, (3) establish zero-defect conditions and (4) monitor and sustain zero-defect conditions. These four aspects of sustainment will now be discussed.
1 Sort out what needs to be controlled
To maintain the improved situation after a successful P-M Analysis, we need to ensure that all the potential causes of the phenomenon are kept at all times within their allowable ranges.
Although we tend to think that we only need to monitor the ‘Not-OK’ items (the potential causes that on investigation were found to have deviated from the standard, and were therefore actual causes of the phenomenon) by incorporating them into the Autonomous Maintenance standards, this is wrong. It is wrong because it might just have been by chance that those particular items were outside their allowable ranges when they were measured, as shown in Figure 4.58. There is no guarantee that the other items, the ones that just happened to be OK when measured, will stay that way for any length of time.
All of the potential causes identified in the P-M Analysis are therefore candidates for being monitored, but this does not mean that they all need to be. They should be classified as ‘fixed’ (unlikely to change), ‘semi-fixed’ (liable to change, but only slowly) and ‘variable’ (liable to change at any time). The fixed ones do not need to be monitored, but the others do (see Figure 4.59).
A fixed factor is a potential cause that, once improved, stays put (like a nut fitted with a locking device) and therefore need not be monitored. A variable factor is a potential cause that changes daily, regularly going outside its allowable range if not closely controlled. A semi-fixed factor is intermediate between a fixed factor and a variable one; it is a potential cause that changes over a certain interval owing to wear and tear (such as a cutting-tool tip or a workpiece holder) and eventually goes outside its allowable range. A potential cause that is fixed when the equipment is new may turn into a semi-fixed one after time if the equipment deteriorates, necessitating regular replacement.
2 Consolidate and shorten checks
Since many of the potential causes of problems identified in the P-M Analysis will need to be monitored and controlled (the semi-fixed and variable ones), the workload of operators and maintenance engineers increases. In fact, the more thorough the P-M Analysis, the more new checks will be identified. Improvement team members tend to think they have done their job if they build the checks into the inspection standards. However, while the result may indeed be a wonderful set of standards, they will be of little use if they are too burdensome for the operators and maintenance engineers to apply.
Resources are limited, so to enable us to reliably sustain and control everything that we need to, we must minimise the time and effort required. This means that we must fix as many of the potential causes as we possibly can. It is absolutely crucial to do as many improvements as possible to turn variable factors into semi-fixed factors, and semi-fixed factors into fixed ones. On top of this, we must also reduce the time and effort required to perform the checks by reducing their number, extending the intervals between them, and making them easier (and hence quicker) to complete (see Figure 4.60).
To minimise the number of checks that have to be carried out, we should try to establish checks as near to the beginning of the P-M Analysis as possible, looking first to see if we can do so at the contributing conditions level, and then moving down to the primary 4-M level and so on (see Figure 4.61). We can also consolidate checks by monitoring groups of potential causes through parameters such as vibration values, and modify equipment to make static precision checks easier to perform.
3 Establish zero-defect conditions (prepare QM Matrices)
To ensure that the checks can be carried out easily and without any omissions, the relationships between the quality characteristics of the product and the 4 Ms for each piece of equipment should be plotted on a QM Matrix as illustrated in Figure 4.62. As the figure shows, a QM Matrix relates the quality characteristics (Q – down the side) to the 4-M conditions that must be controlled (M – across the top), using circles in the relevant cells to show where relationships exist.
4 Monitor and sustain zero-defect conditions (monitor deterioration trends and perform daily checks)
Some of the items to be checked that are identified on the Q-M Matrix will deteriorate as the machine’s hours of operation accumulate (for example, a part may be subject to progressive wear). The measured data for these items should be plotted on a graph at predetermined intervals in order to monitor the deterioration trend. Other variable factors should be included in Autonomous Maintenance standards or in the maintenance department’s periodic maintenance standards and checked on a routine basis.
14.9 Points to Note when Carrying out a P-M Analysis
Each of the steps of P-M Analysis has now been explained, together with the important points to bear in mind at each step. Figure 4.63 summarises these points.
