22 Oct 8 killer ways to apply analytics in oil and gas
Many industries today benefit from applying analytics to their more vexing problems. Examples include life insurance companies who try to predict customer acceptance of new insurance products, retailers who predict the sales of new outlets, cable companies who try out new cable bundles, and food companies who plan menu choices (such as an all you can eat shrimp dinner special) based on expected availability of key commodities (like shrimp).
Alberta’s oil and gas sector has all the characteristics to make it an ideal place to apply analytics to its key problems:
- There’s oodles of data to analyse, and the volume is perpetually growing.
- There’s lots of variables at play.
- It takes scarce expertise to do the analysis.
- The operations will run for decades.
Alberta’s oil and gas fields will generate enormous amounts of information for processing. Simply strapping a dozen sensors to a producing well to record volumes, pressures, temperatures, mass, and throughput (just a small number of the possible variables of interest) can generate millions of discrete data measures. Trying to make sense of all this data, and relate it to each other, to time of day, to weather events, to work shifts, and provide real-time guidance to decision makers is only practical through modern analytics.
The prize is attractive too. The costs of downtime or missing production are huge, and the number of very similar assets in the industry means small benefits can scale quickly.
In my view, there are eight big categories of performance optimisationavailable by applying analytics to unconventional basins.
1) Reduce downtime or business interruption
Oil and gas operations include equipment counts that will number in the tens of thousands and figuring out which are about to fail so that preventative services can be deployed is a serious challenge. And it won’t be clear cut what drives failure. Consider down hole pumps. There are thousands installed throughout the oil and gas fields, and many factors could contribute to pump failure (manufacturer, age, run time, fluid composition, operating temperature). Servicing pumps just before they fail on a planned basis would be significantly better than dispatching emergency-type services after a failure has occurred. Experience in shale basins in the US suggest predictive analytics can lengthen pump life by more than 25%. Other failure prone assets include power supplies, turbines and valves.
Weather events (intense storms, fires) can play a serious role in interrupting product flow. Predicting the weather is challenge enough, but analytics could help in predicting the impact of these events on the business, as well as predicting the actions that could best remediate the impact of events.
Finally, an example from mining, where safety analytics is used to predict which crew shifts are most likely to experience a safety incident (turns out it’s 24-34 year old males with new born kids who are a bit sleep deprived. In Africa, it’s workers returning from Christmas holidays who have partied a little too hard). Safety analytics can help improve safety performance by helping to target appropriate interventions.
2) Manage social license
Alberta’s oil and gas industry has its share of spirited and well organized opposition. These social movements are almost exclusively reliant on digital social media to self organise, coordinate and mobilise, which mean they meet the criteria for analytics (lots of data, growing over time, etc). I suspect the oil and gas industry monitor social media today, but the mounting opposition to pipelines certainly suggests that there’s room to improve that analysis. Consumer-facing businesses are far more adept at understanding their customer base and use analytics for that purpose. Ways analytics help manage social license include:
- predicting where social movements are mobilising
- identifying the role played by various forces (schools, social groups, activists)
- improving the quality and impact of communications with stakeholders
- improving the ability to attract and retain a workforce that is more responsive to social license issues
3) Improve overall performance
There’s no shortage of good ideas that the industry could chase to improve performance. But which ones? And which will have the “best” overall impact on the business? The oil and gas industry tends to look for the small number of big ideas that will make the most difference as quickly as possible. Analytics can help there. But pure digital companies such as Google use analytics to test and refine hundreds of tiny changes rather than creating an internal competition to chase after scarce capital for a few big ideas. I suspect analytics could be more profitably deployed in oil and gas to predict how small changes can add up to big impacts given the long term nature of the industry (drilling thousands of similar wells, for example).
Other ways to use analytics include optimizing the order of doing things (like well delivery) to minimise unnecessary moves, reducing mobilisation and demobilisation costs and consumables handling (like pipe, water and gravel). Consumer goods companies use analytics to help optimise driving patterns and routes, whereas airlines (like Fedex) use analytics to optimize flight fuel performance and flight planning, a concept that should readily apply to the oil sands industry with its trucks and vast geography. If the broader industry, perhaps with some government encouragement, were to share their GPS data, truck loads, delivered weights and storage depots, the industry could optimise the locations for laydown yards, quarries and ponds to minimise drive times.
4) Optimise the resource
Finding the sweet spot for a well, and just as importantly, avoiding poor quality spots, is a job for analytics. The experience in the Powder River Basin in the US (which drilled thousands of gas wells) really illustrates the power of in-the-moment analytics. Independent analysis shows that the basin drilled too many dud wells that either produced no gas, or worse, just water, and that a superior selection of well sites, using just a few data variables, could have saved the industry billions in value. Kaggleis one solution that US shale players are using to analyse well site selection. Drilling just the right wells can save hundreds of millions in avoided well costs.
5) Accelerate value delivery
The sheer number of wells the industry delivers (shale wells, steam injection wells) means that small incremental time improvements translate to big savings. If a well can be drilled 30 minutes faster, there’s a savings in costs from rig rentals, labour hire, and so on, times the number of wells to be drilled. If wells can be brought on line faster, the resource is converted into value quicker, which can dramatically improve company economics. The number of wells, their specific performance attributes, the features of the geology, all lead to the application of analytics to support decisions.
6) Reduce the amount of capital deployed
The oil and gas industry is very capital intense with a huge range of costly items, and it’s not always clear how to optimise the assets. For example, how much inventory of specific items, such as pumps, should be held in inventory, and where in the geography, to keep repair times minimised? The size and scale of installed assets, with their varying levels of operating performance, will make predicting the performance of those assets very difficult without robust tools. Should the capacity of an asset be expanded? Should an asset be abandoned? In my experience, when asset performance is hard to understand, the easiest solution is often to duplicate the asset. This isn’t really an attractive answer in Alberta with its already high resource cost. A better answer is to model out the industry and its operating performance to predict where bottlenecks are going to appear, where optimisation potential can be captured and where assets can be minimised, downtime minimised, asset availability maximised.
7) Achieve better pricing
From time to time, the oil and gas industry will go through periods of high levels of product availability (ramp production as new facilities are put on line, or during shut downs) and well as shortages (when facilities may be interrupted or not as productive as planned). The better players will be modeling out the industry and its infrastructure to gain a lead on pricing by predicting when shortages or surpluses will appear, the likely shape of the forward pricing curve, and the likely competitor responses. And pricing is not limited to product – down the road, the industry should be anticipating that carbon, water, power and input fuels, as well as the global commodity trading environment will all benefit from more thoughtful analytics to understand their behaviour in a complex market. The parallel is the financial industry who use analytics to help predict similar situations involving interest rates and currencies to help develop appropriate trading strategies.
8) Manage risks
Last but not least is the ability of analytics to help manage business risks. Analytics could play a role in helping manage financial and compliance risks by identifying and quantifying risks that play out over multiple assets and time frames, across contracts and sites. Anticipating risks based on fast moving data and getting remediations in place would help the business avoid down time, interruptions, unexpected claims and non-compliance failures.
The case for analytics is very powerful in oil and gas, and gets stronger over time. I would anticipate the global industry to be big consumers of analytics services and solutions in the future.