CONDITION BASED RISK MANAGEMENT (CBRM) ENABLING ASSET CONDITION INFORMATION TO BE CENTRAL TO CORPORATE DECISION MAKING David HUGHES

EA Technology - UK Dave.Hughes@eatechnology.com

INTRODUCTION
CBRM is a process developed by EA Technology in conjunction with several major electricity companies [1-4] to assist with the tasks of defining, justifying and subsequently targeting spending to achieve defined levels of performance. The process has already delivered major benefits. If fully embraced, it can become the basis for effective asset management programmes that address the critical issues of renewing networks and maintaining reliability within the ever more stringent regulatory environment. THE CBRM PROCESS The essence of CBRM is the provision of a structured framework that enables available engineering knowledge and experience of network assets to be readily and transparently linked to corporate decision making. Application of the process results in the fundamentally important decisions, relating to both operational and capital spending, being directly linked to asset condition and future performance. One of the great strengths of distribution network owners and operators is the wealth of available asset specific and generic knowledge and experience relating condition, performance, degradation and failure mechanisms, operating context, maintenance regimes etc. CBRM enables this to be effectively utilised to deliver corporate objectives. To understand the mechanics of CBRM it is useful to define the process by a number of sequential steps as listed below. A more complete understanding and appreciation of the potential of the process can then be obtained from discussion of its application and examples of the output in later sections. 1. Define asset condition. Derive heath indices for individual assets and build health index profiles for asset groups. Health Indices on a scale of 0-10, 0 indicating the best condition, 10 the worst. Link current condition to performance. Calibrate the health index against relative probability of failure (POF). Match the health index profile with current failure rate to determine health index/POF relationship. Estimate future condition and performance. Use knowledge of degradation processes to age health indices, ageing rates dependent on initial health index and operating conditions. Calculate future failure rates from aged health index profiles and previously defined health index/POF relationship.

4.

5.

6.

7.

8.

Evaluate potential interventions in terms of POF and failure rates. Factor in the effect of potential replacement, refurbishment or changes to maintenance regimes, modify future health index profiles and recalculate future failure rates. Define and weight consequences of failure (COF). Construct and populate a consistent framework to evaluate consequences in significant categories, safety, network performance, environmental, financial etc. Build a risk model. Combine POF and COF for asset groups to quantify risk. Total risk can be separated into previously defined categories. Total risk and risk within each category related to tangible quantities, , CMLs, frequency of fatalities or serious injuries etc. Evaluate potential interventions in terms of risk. Factor in the effect of potential replacement, refurbishment or changes to maintenance regimes, recalculate POF and COF and quantify risk reductions Review and refine information and process. Learn from applying the process, identify opportunities to improve asset information and refine models and algorithms. Define and progressively build an improved asset information framework.

APPLICATION OF CBRM Generating short term results It is important to emphasis that CBRM has been developed as a result of working with individual electricity companies to achieve specific, short term objectives. In particular to assist companies to prepare for regulatory submissions. To define asset condition and create the framework for evaluation of potential investment plans within a relatively short and clearly defined timescale necessitated an approach that did not involve major information gathering. Therefore the initial applications were based on the principle of maximum utilisation of existing information. At the start of each project there has been a widely held belief within the participating company that they did not have sufficient asset information to define asset condition and work through the process. Our experience led us to believe that this was not the case, and this has been largely confirmed. Within every electricity company we have worked with (on CBRM projects or in other areas) we have encountered extensive knowledge and experience of almost all asset classes. The difficulty is that in most cases the information is not held in a single central location and is not in a consistent

2.

3.

CIRED2005 Session No 1

form. The most time consuming task in a CBRM project is identifying, accessing and manipulating many diverse and widely distributed information sources. In addition to the specific internal information, there is also extensive knowledge and experience of degradation and failure processes for most asset groups. There is a good understanding of the critical issues that relate environment, duty, maintenance etc to the condition and performance of assets and huge body of generic (type specific) experience within individual asset groups. CBRM is about using all this engineering knowledge and experience to maximum effect, creating a consistent framework to utilise this to define condition and link to current and future performance. It is about putting engineering knowledge and experience right at the centre of asset management decision making. The process starts with reviewing and collating all available information and experience for specific asset groups. At this point a decision is reached as to whether it is viable to generate a health index from the available information. Our experience is that for major asset groups, including substation assets, overhead lines and cables from 11kV upward, health indices that provide a reasonable definition of condition and can be justifiably related to POF, are viable. Clearly, the nature of information available is quite variable and it is accepted that these initial health indices are not perfect, but based on the extent of engineering knowledge available they are deemed to be a credible definition of condition and performance. A very important aspect of CBRM is that it requires a critical assessment of available information and identifies omissions that, if addressed, would greatly increase the confidence in the health index. It also clearly illustrates how information could be defined and collected in a manner that enhances future asset management processes. This creates opportunities to systematically improve asset information as discussed in later sections. Having completed projects with many companies we have realised the value of working within a common framework relating health indices to POF. This had led to use of standard curves that are fitted to specific health index profiles and failure rates for each population. This enables direct comparison between different assets or similar assets in different locations and enables health index values to relate to consistent definitions of condition and performance. In essence, we now calibrate each set of results against a common definition of condition and performance. By cross referencing results for similar assets in different locations with different performance levels we can greatly increase the confidence in the results. A feature of the health index, POF and ageing methodologies used in CBRM is that at all stages they are related back to physical condition and degradation and failure processes and assessed in light of the practical experience of the assets. In CIRED2005 Session No 1

this way there are a number of reality checks that can be applied to the results and the predictions of future performance. In the same way, the assessment of consequences and building of the risk model is based on practical experience and calibrated against tangible values. In most cases the actual total risk in a specific category (safety, financial, network performance) can be defined in real terms. Thus the final risk figure for an asset group and the benefit in terms of risk reduction resulting from a particular investment programme can also defined using the same real values. The CBRM process has been successfully applied to the full range of assets that make up distribution and transmission networks. The value obtainable in the short term is illustrated by example in the following section. Examples of CBRM output Health indices, POF and failure rates. The first output and the foundation for the rest of the process is the definition of asset condition in the form of a health index derived for individual assets. These are built into a health index profile, a distribution of health indices, for a population. As described earlier, considerable thought has gone into the relationship between the health index and POF. An exponential relationship was initially identified as a reasonable description and subsequently this has been used to derive and calibrate health indices. The form of the basic relationship is shown schematically figure 1.

M asurable deterioration but no e significant increase in P(f)

Probability of failure (Pf)

Significant deterioration sm ll a increase in P(f)

5 Health Index

Serious deterioration significant increase in P(f)

10

Figure 1

By adopting a consistent basis for the relationship between health index and POF, a health index profile gives an immediate appreciation of the condition of assets in a group and an understanding of the implication for future performance. A health index derived for a population of HV OHLs based on a combination of condition and design information is shown in the profile in figure 2. Ageing this profile and factoring in the improvements in performance that would be

obtained by rebuilding lines enables an estimate of the overall improvement in performance for different levels of investment, as summarised in tables 1 and 2.
Combined Health and Design Index for HV lines, 0,5,10 years

Grid and Primary Transformers

250

200

150
25000

20000

100

No of poles

15000

50

10000

Range 0-1 8 63 8 100

Range 1-2 13 106 26 100

Range 2-3 11 62 24 45

Range 3-4 8 26 8 16

Range 4-5 5 31 7 15

Range 5-6 2 20 6 8

Range 6-7 3 12 3 2

Range 7-8 1 7 0 0

Range 8-9 2 1 2 1

Range 9 - 10 3 4 2 2

5000

Load CR 4 or 5 & <10km from Coast Load CR 4 or 5 & >10km from Coast Load CR <4 & <10km from Coast Load CR <4 & >10km from Coast
0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10

0 Health and Design Index

Health Index Value

Figure 2, Combined condition and design index for HV OHLS

Figure 4, Profile in year10, assuming no significant intervention

Current rate Year 5 Year 10 % of lines rebuilt over 10 years 2.5 5.0 8.0 20 33

Condition and design 0.064 0.072 0.085 Condition and design 0.063 0.059 0.048 0.035 0.028

3rd Party 0.028 0.028 0.028 3rd Party 0.028 0.028 0.028 0.028 0.028

Total 0.092 0.100 0.113 Total 0.091 0.087 0.076 0.063 0.056

Table 1: Fault rates (per km) assuming no rebuild

Applying the calculated HI/POF relationship to the Year 10 profile indicates a 3 fold increase in rate of failure. Factoring in interventions reveals that to maintain the current (acceptable) failure rate would require 92 (out of 850) transformers to be replaced over the 10years. By considering the relevance and effect of refurbishment/remedial measures on an individual transformer basis and factoring these in, the same overall result can be obtained by a significantly cheaper option involving replacement of 45 transformers and refurbishment/remedial work on 70. Risk calculations. The initial applications used health indices and POF calculations to define investment requirements to achieve a specified level of failure rate. More recently risk models (combining COF with POF) have been populated to provide a more complete means of optimising investment. Risk is evaluated in 4 basic categories (financial, safety, network performance, and environmental). In each category the consequence of failure (the range and the distribution) are determined by consultation with experienced engineers for each asset group. The output is calibrated against real values wherever possible, i.e. the number of CMLs resulting from failures of the asset group. The relative weighting between each category is determined by relating each consequence to a common framework, value in monetary terms, e.g. the monetary value of a CML. This enables the overall risk for any scenario (with or without an investment package) to be calculated as a single number (of risk units, RUs). The total number can then be split down into the 4 categories and the effect (reduction in risk) achieved by different investments packages can be compared, as illustrated by the example in table 3 below.

Table 2: Fault rates in year 10 for different levels of rebuild

Another example (figures 3&4) shows the health index profiles for a population of large transformers. In this case, different ageing algorithms are applied to transformers based on loading and environment.
Grid and Primary Transformers
400 350 300 250 200 150 100 50 0

Range 0-1 27 149 31 168

Range 1-2 14 84 29 72

Range 2-3 6 50 11 22

Range 3-4 5 27 11 15

Range 4-5 2 12 0 4

Range 5-6 2 2 3 0

Range 6-7 0 2 0 1

Range 7-8 0 0 0 2

Range 8-9 0 0 1 0

Range 9 - 10 0 0 0 0

Load CR 4 or 5 & <10km from Coast Load CR 4 or 5 & >10km from Coast Load CR <4 & <10km from Coast Load CR <4 & >10km from Coast

Health Index Value

Figure 3, Current profile for population of grid and primary transformers

CIRED2005 Session No 1

A risk summary for a population of fluid filled cables. Base Case, at the end of next regulatory period with no significant investment. Option 1, including a programme to replace short sections, install additional stop joints and remake terminations at a cost of 12m. Option 2, Option 1 with increased cable replacement at a cost of 16.5m Risk Category Network performance Safety Financial Environmental Total Benefit Cost Cost Benefit 1RU 1500CML, 50,000RU 1 fatality, Base Case 1,893 584 18,510 21,653 42,640 Option 1 759 352 6,308 6,980 14,399 28,241 12m 424/RU Option 2 670 292 5,552 6,114 12,628 30,012 16.5m 549/RU

asset information and therefore the key to building CBRM into business as usual is the creation of an information framework that gives asset managers access to appropriate information routinely. Improved asset information to enhance asset management decision making is seen as the key to success. Discussion of improved asset information usually leads to thoughts of huge IT projects, which involve mammoth cost and have very distant deliverables (probably the last thing electricity companies need at present). Contrary to this, the experience of implementing CBRM has revealed a real opportunity to achieve this objective (routinely available, quality information) relatively painlessly, incurring only modest initial costs with overall cost savings in the medium term! During the development and application of CBRM, a real understanding of the nature and quality of information currently held by electricity companies has been obtained. This has highlighted the present deficiencies and provided an insight into the means of progressively improving the quality, relevance and consistency of asset information in an economic and practical manner. A central feature of the CBRM process is the critical review of the information current available. This inevitably leads to identification of additional or alternative information that could be particularly useful in improving the ability to define condition and future performance. Furthermore, activities involving extracting data from various sources and combining it to give a numerical value of health index, has demonstrated the difficulty of working with data in the variety of different forms currently available. The potential value of adopting a system that defines the results of all inspection or tests by a relatively simple numeric condition rating system then becomes clear. The CBRM process, and some of the work to optimise maintenance that preceded it, also demonstrated the significance and importance of maximising opportunities to obtain critical condition information, in particular to integrate the information collection across the inspection and maintenance activities. This leads to significant opportunities to optimise the whole management process and generates significant OPEX savings, better targeted inspection leading to less invasive maintenance and the collection and storage of less (but more useful) data. Another feature of the CBRM process is the principle of progressive improvement. The aim has always been to produce the best result possible with the currently available information and then define the means to progressively improve the output by improving the information. Improvements in information are obtained not with large expensive condition assessment activities, but by identifying potential useful information that can be obtained during existing, routine inspection and maintenance activities. This is perhaps the most critical lesson. Distribution assets are visited frequently. By careful consideration of the

Table 3, Summary of risk for fluid filled cables

1RU 70 OPEX, 1000litres of oil 140RU

This risk model is still in the development stage and needs some refinement, nevertheless the results we are achieving provide an effective means of defining a complex and rather abstract quantity (risk) in a manner that can be related to tangible values. Encouragingly where the results have been presented and discussed with asset managers there has been a high level of satisfaction, both in terms of the relative values between asset groups and the overall presentation of risk, in a manner that can be directly related to engineering knowledge.
SHAPING THE FUTURE

Building CBRM into business as usual The initial impetus to develop CBRM came from the need to define and justify spending plans in regulatory submissions. Great success has been achieved but it has been a very labour intensive process to provide a one off snapshot of asset condition and risk. The potential value of building CBRM into everyday activities such that asset managers routinely have access to well defined asset condition information that can be used to actively manage asset is clear. Indeed the ability to target spending (both OPEX and CAPEX) to achieve defined benefits will be essential to maintain acceptable reliability while working within strict financial restraints. Built in CBRM would give asset managers the ability to direct resources to achieve specific performance objectives, something that is absolutely essential in the increasingly demanding regulatory and financial environment. Improved asset information the key to future success The key to implementing CBRM is access to appropriate CIRED2005 Session No 1

information that will enhance decision making processes it is possible to define inspection protocols, using simple numeric condition ratings that will rapidly lead to a data resource that can routinely enable much more effective decision making. By targeting the information to meet specific requirements, the volume of data and the whole inspection and maintenance regime can be rationalised. All this should be possible using existing IT infra structure. Creating an improved information framework is an engineering issue not a major IT project! Applying the principles discussed above therefore provides electricity companies with a significant opportunity to inexpensively create an information framework that will enable targeted spending to deliver specific network performance at minimum cost. This is critical to future success. For UK companies this will enable delivery of DPCR4 targets and be excellent preparation for DPCR5! CONCLUSIONS CBRM provides an effective means of linking the extensive engineering knowledge and experience of network assets, available in most companies, to corporate decision making. Implementation has demonstrated it can deliver significant short term benefits. The CBRM experience offers the opportunity to build an improved asset information base that would greatly enhance decision making capability. It would enable future spending to be defined, justified and targeted to deliver specific levels of network performance at minimum cost. CBRM is a vital component of successful asset management in the ever more demanding regulatory and financial climate.

Acknowledgements The author is indebted to the many colleagues at EA Technology who have contributed significantly to the development of CBRM and to the numerous partners in the participating electricity companies. CBRM only works if there is effective collaboration between all those who share knowledge and experience of the assets. References [1] E Hamilton and D Hughes. CBRM, the practical application of intelligent asset management to United Utilities distribution network, Distribution Europe 2004, April 2004. [2] D Hughes. The use of health indices to determine end of life and remnant life of distribution assets, CIRED, May 2003. [3] T H Jartouson, B Jesus, D Hughes and M Godfrey. Development of health indices for asset condition assessment. IEEE, Transmission and Distribution, Dallas, USA, September 2003. [4] T Richards and D Hughes, The use of condition information for strategic risk management, Euro TechCon Birmingham, November 2002.

CIRED2005 Session No 1