Capability

 Reality 

Reality, because of its complexity, can be limiting in practical and analytical terms. Figure 2 below is used in some arenas as an illustration of the contracting solutions for equipment procurement and support, progressively leading to acquisition of capability.

Fig. 2 - Transformation Staircase
UK LoD Acquisition Management System - 2005 ⁽²²⁾

Some take the view that the illustration is a display of absolute intent because the progression appears to be onwards and upwards, as if a staircase to better things, which may or may not be true. It is true however that it is recognized that there cannot be a 'one size fits all solution'. Developing strong and healthy Customer/Supplier relationships is key in this as a means of achieving value, being output rather than input focused, forging the value chain, and taking a whole-life perspective. Consequently there is a declared intent to have more and improved strategic, professional partnering arrangements, in the public, private and public/private arenas. Although the idea of partnering and collaboration with key suppliers is not new in itself, in this sense it requires a cultural shift to "an approach with an attitude and a management ethos, (towards selected suppliers) of openness, effective communication, close collaboration and co-operation, trust, honesty, transparency, sharing and mutual benefit." CIPS ⁽²³⁾

Clearly, it is accepted that there must be greater involvement between Customer and Supplier to make transformational progress. What is not always so readily recognised is that Customer and Supplier sophistication must also be improved if there is to be a real chance of the Customer properly articulating expectations and the Supplier meeting them. They must understand the outputs and the means to deliver them. Ultimately this means entering the maelstrom together to make sense of the data whirlpool.

"Capability" is easily defined, but it is not simple. For example, it can be defined as follows ⁽²²⁾:

An operational outcome or effect that users of an equipment or service need to achieve. (Sense One)

It can also be defined as:

The operational need which is satisfied by the deployment of an operational system integrated with other cooperating systems. (Sense Two)

"Effect" can be defined ⁽¹¹⁾ as: 1. "Something that is produced by a cause or agent; result. 2. Power or ability to influence or produce a result. 3. The condition of being operative."

It is obvious that we cannot guarantee that users could or would achieve or influence the required operational outcome or effect; we either do or we don't. If we do, we don't necessarily know if we are being efficient. If we don't, we don't necessarily know that we are being inefficient. Therefore it is unreasonable to attempt to measure Capability in this sense.

However, if we consider Capability in terms of Sense two, then it may be possible to consider the value contribution of the deployment of an operational system to output effectiveness.

There are broader views of effectiveness. Jane Ellis cites Seashore and Yuchtman ⁽²⁴⁾ who define 'effectiveness' as 'the ability of an organisation to exploit its environment in the acquisition of scarce and valued resources to sustain its functioning'. Martin Christopher ⁽²⁵⁾ supports Seashore and Yuchtman in their assumption about scarce resources and develops it further; "Since we can assume that money spent on service is a scarce resource then we should look upon the service decision as a resource allocation issue." In both cases efficiency and effectiveness are inextricably linked to outputs.

If we can understand the contribution of our resources to our output, potential and/or actual capability, and the end to end efficiency of our processes, we can begin to recognize their value contribution and make allocations appropriately. We should separate the tangible from the intangible and have real performance measures that can be seen to relate to product or service value and effectiveness ⁽²⁶⁾. These metrics should be the drivers for decision making both from the operating and operating cost perspectives. Easier said than done!

Recently I was party in a discussion with representatives of a global consultancy firm who had been engaged to improve availability in a specific area. When I asked them what their definition of availability was, I got a slightly pained and unclear response; but I am certain they went away somewhat better informed. Also recently I was party to discussion with a major prime contractor who was similarly looking to improve availability in a different, but related, area of business. Therefore it seems that there is recognition that availability of some sort is key to delivering outputs but there is uncertainty about the understanding of exactly what and how. It also confirms Stephens et al's view that "overall behaviour emerges only when the complete system can be seen as an entity." I interpret this to go beyond the systems engineering environment because if we accept that complexity is a key issue, then we should accept that there is also uncertainty, even beyond the point of acceptance that cannot be quantified without empirical data. Furthermore, Stevens et al ⁽¹³⁾ define a system as "A human-made entity with a distinguishing and defined purpose that draws on integrated, constituent parts, each of which does not individually possess the required overall characteristics or purpose." Whatever, the elements forming an entity are interrelated, interdependent or interacting and it is only when a system is brought into use that comparison between the design abstraction and the physical reality can be properly made; how does the actual failure rate compare with the output design failure rate.

But how often are we the architects of our own downfall because we have accepted what we are told by the designers and manufacturers, its reliability? Do we really understand what our expected value of availability should be? Do we ever quantify it in any truly meaningful way? Is it appropriate to plan based upon a negative exponential distribution and expect that reality will match that expectation?

Albert Einstein ⁽²⁷⁾ said that "As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality."

The shape of the curve which is a function of the actual failures cannot be deemed to be a product of an exponential distribution unless analysis reveals it as such. Just because three objects have the same surface area does not mean that they are the same shape. It only takes one parameter, the radius, to calculate the surface area of a circle, whilst for a rectangle it takes two, length and height, and for a trapezium it takes three, lower length, upper length and height. Therefore it is unreasonable to summarize the shape of a curve unless there is sufficient information to properly describe its attributes.

It is the ability to characterise the distribution of failures mathematically as a function of the observed failures in finite populations against a time related independent variable and approximate these to infinite population conditional probability models that enables probability analysis to begin. The measures of location and of spread or variability can be determined and these in conjunction with the characterization can describe the curve under consideration and enable the calculation of probability of failure and conversely survival. Barlow offers a definition of probability as "a degree of belief held by an analyst or observer". He says "The idea here is that probability is an analytical tool based on judgement useful for making decisions. Of course we are free to adopt observed frequencies as our probabilities if we are consistent, but observed frequencies are not otherwise probabilities according to this definition."

It is interesting to note that in consideration of the maelstrom, organisational transformation, partnering and the various views of reliability and availability that Reliability Centred Maintenance is in vogue and seen by some as medicine for making things better.

Reliability Centred Maintenance (RCM) is a "method for establishing a scheduled (preventive) maintenance programme which will efficiently and effectively achieve the inherent reliability and safety levels of equipment. It is methodology which can be applied to the development of a preventive maintenance programme and results in improved component reliability and minimised overall programme costs. The intended end result is improved overall equipment safety, availability and economic operation." ARMP ⁽²⁸⁾

RCM as a methodology takes a Physics-of-Failure perspective including root-cause analysis, and evaluation of the impact of defects and stresses on product reliability. Based on the analysis, the failure is eliminated, or abated, by re-engineering the design and/or the production, assembly, and/or support processes.

But is RCM as an oxymoron? If we accept that RCM can improve the operating reliability of equipment and consequently increase availability, then by that acceptance are we not denying that failure is generally exponentially distributed?

Fig. 3 - the P-F Cirve
Adapted from Cranfield University, RMCS, 2004

If failure is exponentially distributed, through RCM, broadly speaking, we must be aiming to increase the Mean Time Between Failures and/or decrease costs. However the P-F Curve above belies the exponential failure of an item under consideration because, although only an illustration, it does characterise the LRU/component under consideration as having a minimum period of time before it will degrade and ultimately fail; the distribution of failures is not exponentially characterised. This begs a number of questions.

If the support system is functioning as designed, does the empirical data indicate a correlation relationship between the failures of the LRU/component under consideration and its B10, B50, B90 lives?

Will the re-design or re-engineering of the LRU/component positively affect the Mean Time Between Failures of the system such that availability improves?

How would we know that availability has improved?

I am not going to speculate on the first question because that is a matter for analysis. On the second, I only want to say that it is not necessarily the case that redesign or re-engineering will actually improve availability. It may be the case that we are using less of the component under consideration, but how has that impacted upon the overall system availability? In his book "System Engineering Science" ⁽²⁹⁾, Arie Dubi shows that it is scientifically possible that an LRU/component with a higher failure rate can produce a higher overall availability of a system. My third question is largely unanswered, in my opinion, because of the lack of understanding of availability.

 My Concern