1. Summary of project, product, framework
Sydney Trains (ST) is introducing new business processes and decision support tools to evaluate and justify, in quantified cost, risk and performance terms, the optimal time to replace aging and obsolete equipment. These are a significant departure from traditional data-analytical methods or conventional investment evaluation techniques. They incorporate technical, commercial and sustainability factors and enable expert knowledge to be quantified for filling gaps in hard data evidence and forecasts. The techniques have now been applied to a range of ST asset classes, yielding multi-million dollar added value through changes in Capex and Opex requirements, risk reduction and performance improvement.
2. Description of LCO project: best practice asset management – in practice
Two big initiatives have occurred recently in the development of asset management in ST:
- Building an Asset Management System (ISO 55001 certification – 2017)
- Consolidation of asset information and work management systems (EAM system roll-out – 2017)
These have delivered improved line-of-sight alignment, strategic planning, value and efficiency in how ST’s asset portfolio is managed. The other priority is to optimise effectiveness. This is where good decision-making is vital: determining what is worth doing, when, to deliver best value-for-money over the whole asset lifecycle. We conducted a world-wide study of best practices and technical solutions to address these questions, concluding that the SALVO (Strategic Assets Lifecycle Value Optimisation)[1] methodology and DST[2] toolkit from The Woodhouse Partnership Ltd represented the most robust, advanced and practical solutions for our business needs. They align with the ISO 55001 requirement for a clear decision-making framework and decision criteria.
SALVO is a six-step process developed by an international group of asset-intensive organisations to:
- Force the right questions to be asked, within a systematic framework
- Focus on business value opportunities (combining costs, risks, performance, sustainability and other decision-drivers)
- Identify alternative solutions or options, including non-obvious ones that may not require an asset intervention
- Evaluate options in consistent value-for-money terms, including the degree of confidence (quantified ‘cost of uncertainty’), the value of ‘intangibles’ and the premium paid for compliance with obligations or constraints.
- Train technical staff to explain and justify a business case for investment (and optimal timing) in terms that financial staff can understand.
We adapted SALVO to establish an innovative Life Cycle Optimisation (LCO) process, involving a ST-customised mix of business process changes (i.e. who makes what sort of decision, how), training courses (rigorous consideration of cost, risk and performance elements, along with all other decision drivers) and decision support tools (DST) that enable instant ‘what if?’ calculation of optimal intervention timing and value impact. The combined process has opened people’s eyes and established a new way of breaking down functional silos to rapidly achieve consensus. We have created a fundamentally new way of looking at our assets, our interventions and our demonstration of value to our stakeholders.
A distinctive feature of this approach has been how it draws together and visibly applies core asset management principles into every-day decision-making. SALVO forces us to always consider whole lifecycle value (business impact). It connects good decision-making to context-specific stakeholder expectations and AM objectives. And produces clear-cut justification (and assurance) that we are doing the right things, for the right reasons, at the right times.
Figure 1. Example of DST ‘storyboard’ and cost/risk/performance-optimised decision results
Figure 2. Summary of Sydney Trains Lifecycle Cost Optimisation process
The LCO process has also proven extremely flexible; evaluating options and quantifying the business cases for diverse cases as:
- Track sleeper/transom replacement options and timings
- Condition monitoring strategy for points
- Maintenance frequency for traction motors
- Optimal spares for MV transformers
- Replacement timing for old 66KV circuit breakers
- Wooden pole inspection strategies
- Ballast cleaning machines (maintenance and spares)
- Life extension options for aging ‘heritage’ bridges
- Signalling system Upgrade timing
- Point machine replacement/upgrade justification
- Rail head wear monitoring
- Rail renewal/lifecycle
- Track circuit system obsolescence/upgrades
User feedback:
Originality and ingenuity
We believe this is the first fully-rigorous development and implementation of such a decision-making framework for Asset Management within Australia. Many organisations have addressed parts of the requirement (e.g. certain decision types) and have component tools for specific purposes, but our LCO is a systematic decision-making methodology, covering all types of asset management decision and at criticality-proportionate levels of application. It combines human psychology (e.g. how to elicit ‘tacit knowledge’ from subject experts while minimising bias and identifying/quantifying the inevitable uncertainties) with advanced technology (e.g. artificial intelligence algorithms incorporated within the DST modules to seek the optimal work program for total cost, risk, performance impact).
Other elements of innovation lie in:
- The ability to model how risks affect each other (one risk or degradation mechanism affects exposure to others, and vice versa), to ensure their combined effect is correctly quantified and treated.
- Human factors: the structured, team-based determination of the true nature of the problem/opportunity, what could be done, and what is the best value option (and optimal timing).
- Enabling complex compromise strategies to be expressed in clear business value terms, so non-technical stakeholders can understand why a proposed strategy is optimal.
- Cumulative learning and decision refinement made simple; the transparency and real-time ‘what if?’ capabilities mean that any new information, or changing context, can be instantly incorporated to refine and adjust asset strategies.
- Quantifying the dollar significance of ‘intangibles’, data uncertainty and the premium paid for compliance with non-negotiable constraints; enabling proportionate effort to manage expectations, refine data or even negotiate more appropriate boundary obligations.
Program and project management
Implementation of an LCO process requires new skills and a significant change in organisational culture, besides the process changes and tools/systems configuration. The successful change from established (strongly departmental) methods of working, to cross-functional, value-based decision-making required a people-centric transformation program. From the outset, therefore, significant investment was made in training those directly involved in asset management decision-making, as well as an awareness and education program for ST stakeholders and sources of relevant asset and intervention expertise.
Following signoff of a project charter and appointment of the steering committee, the program started with SALVO/LCO introductory training for potential users/contributors/beneficiaries. The next phase developed an integrated process map for the LCO process, showing its fit with existing ST business processes and systems. Detailed methods training was then grouped by decision-type audiences, such as those responsible for projects/change evaluations, strategic spares, operational maintenance, inspection and lifecycle (e.g. Refurbishment, Renewal) interventions. In all cases, the formal training was followed by immediate on-the-job facilitation of ‘live’ current cases, and coaching to ensure conceptual understanding was reinforced and consolidated by practical application. Throughout the program, skills and competency assurance was a priority, using a formal certification scheme (mini exams and independent case validations) to ensure that standards and consistency are maintained.
An agile project management process was used throughout, to adapt as skills, knowledge and confidence grew. This has resulted in a naturally-evolved group of ‘super-users’ providing peer support and quality validation to those with more intermittent involvement. This has delivered a strong sense of ‘ownership’ in the resulting solution, embedded in strategic planning and the AM management system (e.g. a clear-cut decision-making framework/criteria) and transparently evident justification for AM decisions that show how costs, risks and performance are optimised in day-to-day decision-making. The next phase of the transition is therefore in sight – namely to consolidate such decisions into corporate resourcing and strategic budgeting processes, including the business interface with Transport for NSW, the ultimate asset owner.
Business impact
The LCO process is consistently revealing scope for improving the mix of costs, risks and performance (compared to expert judgement or current/historical practices). From a representative sample of eight asset management decisions, one case typically confirms current practice to be correct and the others show improvement opportunities of A$20k/year to A$300k/year each (through a mixture of costs avoided, risks reduced and/or performance improved). And, in a few cases, the impact is substantially greater: already we have several cases where the process has shown benefits of A$20-plus million NPV.
Other organisational benefits, we have found:
- LCO provides a much-improved level of consistency in asset lifecycle decision-making.
- It provides a structured method for SMEs to actively contribute in decision-making.
- The methodology enables complex combinations of risk, cost and performance factors to be assessed and simplifies the demonstration of their impact upon a decision.
- It overcomes one of the biggest challenges in AM decision-making, achievement of consensus between multiple stakeholders with different interests in, or opinions about, the case (such as failure rates or event consequences). LCO cuts through emotions, ‘vested interests’ and ‘paralysis by analysis’.
- DST enables extremely quick turnaround to tangible outcomes (within a two-hour workshop) for stakeholders to evaluate options, test for sensitivities and achieve consensus.
- The tools ensure lifecycle decisions are recorded in a centralised library, improving audit trail and capturing history about how a decision was made and why.
- The methodology and software enable complex decisions to be simply and clearly presented to external stakeholders for endorsement and approval.
Case example
The ST Asset Condition Assessment method (ACAA) gives an estimation of remaining technical life of assets. The LCO process is then used to evaluate options and make the cost/risk/performance justification of the optimal timing. This stage is particularly important when:
- The cost of asset renewal is high compared to ongoing maintenance, repair and component renewal.
- There is a potential impact on system reliability and safety if renewal is deferred.
- There are many units or systems requiring renewal in similar timeframes (competing for resources/budget).
One example is the A$300 Million replacement decision for signalling interlocks within the Sydney Signalling Box. With a design life of 25 years, they are expected ‘due’ for renewal/upgrade in 1-2 years time. The LCO process considered alternatives such as refurbishment (to extend life) and various replacement design options and timings.
In the case modelled in Figure 3 below, the black line shows the total business impact of different asset replacement timings. Contributing to this, the purple line shows asset replacement and future asset lifecycle costs, discounted according to when the replacement occurs, the yellow and blue lines show how the current asset maintenance and risks increase with such delay. The maintenance includes both ongoing routine maintenance and minor component renewals which increase in both frequency and cost-per-occasion as the asset ages (and obsolete components are harder to obtain).
The risks incorporate various failure mode probabilities, and their impacts on safety, remedial costs and system downtime (quantified in terms of delays to customers). These are based on ST standards for the cost of train delays and the significance of safety incidents. DST software tools were able to combine information sources from both historical records and range-estimated future projections.
Figure 3 SALVO/LCO Analysis for Signalling Interlocking; Base Case & Sensitivity Testing
The DST modelling showed, in contrast to ‘design life’ (replace in next 1-2 years), that the ‘most likely’ optimal asset replacement is in 20-24 years. This result varies with data assumptions over the period 12-40 years (see sensitivity graph in Figure 3), but in no case could renewal be justified earlier than 12 years. Furthermore, it can be seen the ‘total impact’ line is relatively flat within the min/max range of replacement intervals, so it was decided to put the investment into strategic plans for c.15 years time. This minimises the current ‘cost of uncertainty’ and would enable a revision of the decision, for potential further deferment, in c.10 years (by then supported with the additional evidence of ongoing reliability).
The option for a refurbishment to extend asset life was also assessed, but the modelling quickly showed that this would not reduce total lifecycle cost/risk enough to justify the necessary project investment, even though the asset life would be c.10 years longer.
The results were reviewed against the ACAA model and the judgement of maintenance engineers and other subject matter experts. The optimal strategy correlated reasonably well with the lower end of expert judgement: as in other studies, this reveals that people can be somewhat risk averse in their judgement of future system performance. The impact of optimised renewal timing, compared to the original technical/design life, represents a net benefit of c. A$90 Million NPV.
These relay interlockings are typical of installations of that type across the network, although larger in scale, and it was determined that similar lifecycle patterns would apply to future degradation and risks at other locations, so adaption of the same study to other locations is simple and rapid. A separate study is also now being undertaken for computer-based interlockings, since the risks and impacts are very different, and the obsolescence timescales for components considerably shorter.
