How to cut your shutdown frequencies by 50 per cent
Extending intervals between shutdowns is the holy grail for people with the headache of managing a large, complex, critical plant that’s continuously running. This has been further heightened by the lamentable slide in oil prices in the past 24 months.
That said, there is an existing methodology that, when properly implemented and understood, could cut shutdown frequencies by 50 per cent.
To get your house in order and optimise your shutdown cycles, you really need to understand three key areas. Firstly, the human factors: psychology and self-interests have to be addressed. In that, I mean proactively handling the difference between what is right for the business in total cost/risk terms compared to the localised self-interest and short-term gains that tend to affect personal motivations.
Secondly, you need to appreciate the assets, their diverse needs and differential importance; the risks, failure modes and mandatory obligations of the daily running of a facility. This involves the cross-disciplinary collaboration and discipline to sort out the ‘must do’s from the should-dos from the could-dos from the won’t-need-to-dos.
Third, you need to look at the pounds, shillings and pence. But it’s not just cash, it’s resources and ‘lost opportunity costs’ as well. This addresses the capability constraints and value-for-money dimensions – funding and skills and so on; what can be done, to achieve what business benefit.
Looking at this through the prism of the six-step SALVO Process (see Figure 1), the optimisation of a shutdown strategy requires steps one to five all to be joined up. The first ensures diagnosis of which problems are actually the most pressing, not just which are the highest risk (most organisations already have concepts of criticality analysis, risk ranking and so on).
Figure 1. The SALVO Process
We also need to look at the rate of change of risk. This needs to be built into the activity selection and prioritization criteria. It is not just a matter of how big is the cliff but also how close to the edge are you standing.
And then SALVO steps two and three sort out what the underlying issues are, and what is possible to do about them (including avoiding the need for shutdown). These steps start with the principle that you don’t need a shutdown at all, unless there’s a bottom-up justification that something is unavoidably, necessarily going to require a shutdown.
So, at this stage we are doing an enhanced version of scope challenge, with a timing dimension built in and considered. Best practice shows that there are actually five filters of such scope challenge needed. And it’s astonishing how, with some structured guidance, some multi-disciplinary lateral thinking can eliminate the need for shutdown-dependent tasks.
These may involve minor plant modifications or you may need to adjust operating parameters or even simply ‘manage the expectations’ of clients or contracting partners.
So then, for the tasks that survive these challenges, there comes the ‘bang for your buck’ test of tasks and their timings. It’s where we bring technical versus commercial interests into play because the technical need is rarely absolute. It’s tradeable – in terms of the level of risk that is worth taking. SALVO step four is a guided evaluation, with ‘what if?’ software tools, to evaluate different task types, their cost/benefit/risk impacts and (individual) optimal timings.
Navigating the competing interests
We use the ‘Shamrock’ diagram (Figure 2) to make sure that we’ve covered all the stakeholder interests. The outer ring represents typical ‘business drivers’; taken clockwise, we must do things safely, reliably, sustainably, cost-and-performance-efficiently, at high quality, satisfying customers, being environmentally responsible and compliant with absolute obligations.
The five inner groupings enable us to quantify these pressures – they are different mechanisms for quantifying the significance of these requirements i.e. scaling the elastic that is pulling us in all the different directions. For example, we use the term ‘Shine’ to cover the intangibles or ‘perception’ influences, where the best technique for quantification is indirect, based on the difference between what we would like to do (for Shine reasons) and what we would otherwise be justified in doing: this gives us a ‘premium paid for Shine’ that can be managed more transparently and consistently.
Figure 2. The Shamrock diagram
Optimal bundling and timing alignments
Typically, for pretty complex process units, the ‘bundling’ stage (SALVO Step 5) is dominated by somewhere five to ten interval-constraining tasks, representing the highest frequency, cost/risk critical and most unavoidable activities. These are individually represented in the search for the optimal combined programme.
Other, medium criticality, activities exert a more collective influence on the shutdown strategy (e.g. scope, cost and duration growth with longer shutdown intervals). And there is a third family of activities that is best treated as ‘opportunistic’ or ‘parasitic’ – they are never able to justify or modify the shutdown timing by themselves, but are worth doing if an appropriate shutdown opportunity exists.
All three groups need to be explored for the optimal combined work scopes and timings of shutdowns. This is an immensely complex search, and the SALVO consortium commissioned and developed a unique tool to do it – DST Schedule Optimizer™.
This uses an extremely advanced and rapid search engine, combined with life cycle costing, reliability engineering and financial cash flow analysis maths, to explore the best work combinations and cost/risk compromise for jobs that have different urgencies and timing sensitivities.
The results of this process (see Figure 3) have invariably shown big surprises and opportunities for eliminating shutdowns, extending intervals or, on some occasions, achieving longer run lengths between major turnarounds by creating targeted ‘pit-stops’ for smaller work scopes.
It also provides, with complete transparency, the business case for the shutdowns that are worthwhile, and the cost/risk impact of deferment. So we can see the ‘culprits’ or bottleneck tasks that are constraining longer run lengths.
Furthermore, as everything is converted into dollar impact, and integrity risks are also quantified, the potential benefit for ‘debottlenecking’ is also quantified. This represents the potential ‘payback’ for any minor engineering projects that might design out the need for that activity.
(a) Before (each task performed at its ‘personal’ optimal)
(b) After (re-optimised for shared costs, downtime and overheads)
Most people spend their time planning and executing each shutdown in turn, having skipped the first, much more important consideration of what is worth doing in the first place and why - they jump straight into the how.
When they finish one shutdown, they start planning the the next, relying on some post-mortem findings, presumed task lists, backlog and some common sense scope challenge. Shutdowns can even become self-fulfilling prophecies, emerging from two cyclic assumptions: “we have XYZ work to do, so when shall we shutdown to do it?” and “We have a shutdown scheduled, so what work should we do in it?”
The SALVO process injects the necessary business case discipline onto a proper determination of what shutdowns (if any) are really needed, and how often, with clear demonstration of the impact of alternatives. That's the basic architecture of it. And it has been proven repeatedly in different plant types, operating environments and cultures.
In an early example, in Sasol's Secunda plant in South Africa, we helped to shift the two-yearly cycle for core process reactors straight out to four-year cycle. This involved identifying, justifying and eliminating 11 critical "weak links in the chain" to enable the longer run lengths.
The same process was applied to all of the SABIC Innovative Plastics sites in Holland, Spain and USA – again, doubling the interval between shutdowns in a single three-month study, including developing the business case for all the (very minor) plant changes that enabled this.
And similar findings have been achieved in HV electrical transmission networks and car tyre production lines – the process seems generically applicable and, in each case, has realized a major change in shutdown strategy. The smallest benefit so far has been a 28 per cent reduction in planned downtime.