It is about Evolution, not Revolution!
One of the biggest misconceptions about simulation and modelling is that one needs to add all information before getting useful information as a payback. I strongly disagree! It is about Evolution, not Revolution!
Models can yield useful information even in its lowest level of detail. Here are some tips:
1) Use your existing knowledge!
You know your plant better than anyone else.
Sometimes, it is hard to find evidence to prove this intrinsic knowledge and that’s when some methods become quite important. For instance, being able to simulate the asset behaviour (or how it is likely to behave in the future) provides scientific evidence that one could use to go after that budget to find improvements.
Your knowledge also provides a powerful initial screening to what needs to be taken care initially “this is where I will invest my time because I know there will be some Return On Investment”. This connects quite well with the next tip…
2) Find the main problem then go into detail!
You don’t need to model an entire refinery or all the production wells of platform to support your decision-making process.
For example, if you are running a performance forecasting analysis, you can assign a single failure and repair pattern that describes all the subsystems and equipment items described for that unit. If you do this for all units, the most critical one can be easily identified. This unit requires extra attention and it needs to be detailed to support a solid decision-making process. Once you finish your optimisation process for this unit, move to the second on the list. This process does not need to be complete in a specific time frame – each step is actually optimising the system performance and will result on improvements to the performance.
3) Show the results – it is about continuous improvement
Show the results for your analysis. We have been to several meetings where a nice “WOW” indicates a happy client. This does not mean we have done something sophisticated but modelling and simulating provides insights to data and information that were not previously available. Just by starting somewhere, some counter-intuitive problem might appear. Sometimes the solution is simple!
One great example that I remember was one of our clients working as a FPSO operator – their starting point for the modelling process was the topsides focusing on the oil and gas systems – with this consideration, they couldn’t reconcile the performance data from the field to the model. As they moved on and increased the complexity of the model, they decided to take into account the water injection system. Only at this point, they have realised that the inspection regime for one of the filtration packages was hurting their bottom quite a lot.
4) Keep the model live!
In DNV GL, for example, we have been working with many companies for years to maintain their models live. Some models evolved from a simple gas station to a country-wide gas network incorporating 15 gas stations, all pipelines, all fertiliser plants and power stations. So it is really about evolving the model, getting feedback from operational data and suggesting improvements along the way.
5) Get yourself into the Virtuous Cycle of Data collection
One new trend that we are identifying in the market relates to all this process. We need to start somewhere and companies that engage with small models are in a much stronger position to benefit from a complete view of the asset. Every step of the process – from data acquisition to the transformation of information to knowledge – can be successfully replicated and a series of “lessons-learned” can be derived. This means that the next units you are modelling will only take 50% of the time when compared to the first project.
So, we see this “Virtuous Cycle of Data collection” as a company that start to gather data to transform into knowledge. Initially, the support of the decision-making process is not good and the results are not perfect. But next time they try something, they will start from another level. This process continuous until the company hits a critical mass which will allow them to move to more advanced methods.
For example, we have seen companies moving from qualitative RBI to quantitative RBI. This was only possible because they’ve managed to collect the necessary data and are mature enough to move to more advance methods. So, qualitative methods are a great starting point! From Preliminary hazard analysis (PHA), Hazard Identification (HazIds) to Consequence Analysis and Quantitative Risk Assessment (QRA)! From a Failure Mode Effect (and sometimes Criticality) Analysis (FMEA/FMECA), Hazard and operability study to a RAM analysis or Performance Forecasting study (yes, they are different J).