You are Home   »   News   »   View Article

Tessella; how to do analytics

Thursday, May 9, 2013

To do useful data analytics in the oil and gas industry, you are better off defining the question you want answered and then gathering data to answer the question, rather than gathering data and then trying to work out what you can do with it, say Nicholas Clarke and Bas Hendriksen of Tessella

There is a lot of talk at the moment about making advanced use of 'analytics', analysing your company's data to work out clever ways to improve your business.

The best way to do data analytics is to first figure out what questions you want answered, and then gather the data you need to answer this question, says Bas Hendriksen, business development manager at Tessella.

This is different to just setting off data analytics tools on the data already in your data warehouse, he says.

Mr Hendriksen was previously IT manager at Shell's corporate CIO office in The Hague, and strategy manager for Shell's Smart Fields research and development.

Analytics can be effective 'when the business says, here is a problem that we don't understand that happens often, can we pull data from different sources and work out if there is any logic by which we can forecast or at least recognise what is happening,' says Mr Hendriksen.

Analytics 'can only be effective if people have identified what their problem is.'

For example, in Nigeria, some oil and gas companies have been trying to work out if oil is being stolen from long pipelines, by taking pressure readings at both ends of the pipeline and comparing those . 'Here you see an example of using available data in a smart and not common way to improve your performance,' he says.

The oil and gas industry has been gathering more and more data over the past few years, Mr Hendriksen says, with more sensors and control systems. But it is still seeing more delays or deferments than it would like, which shows that it could maybe get a lot more value from the data than it does at the moment.

'The data that is there to monitor what is happening, is only partially being used,' Mr Hendriksen says. 'Most commonly the data is being used to ensure safe operations, and with current staffing levels, that is quite often all staff can do.'

'There's a lot of technology out there, all the big vendors have vast arrays of big data tools, analytics tools, there's no shortage of tools. But we're not seeing that [used] to great effect. The focus is still far too much on the data. It is so much more straightforward to collect the data.'

One example of an industry making good use of analytics data is supermarkets' loyalty? schemes, which enable the supermarket to gather large amounts of data about individual's spending patterns.

The smart supermarkets do not just dump their data in a data warehouse and then let some geeks loose on it with analytical tools; they build up an entire data infrastructure to make sure they are generating the information they need, about how different marketing promotions are performing.

The marketing planners never see, or work with the raw transaction data, it is too detailed. Rather they work from the derived information they need to get insight on how to make better decisions about specific marketing programs.

'You have to be absolutely clear about what you want to have, don't lose the key data amongst the mass of all the other data. Formulate your questions and get them answered,' Mr Hendriksen says.

'Our concept is that the data should live in an analytics laboratory, where it is there to be used, living,' says Nick Clarke, head of analytics with Tessella. 'The data only exists in the system to answer questions that you'd want to have answered about your business.

It is an active space, focussing on answering known questions.'

By contrast, a data warehouse 'is somewhere with bins to store things in. It is storage based around future retrieval and future use. Very useful and necessary in its own way but limited as a concept.'

Your organisation

'To make analytics work it requires the business strategy people at board level being involved, all the way down through to the IT infrastructure people, who supply the data at the right time,' says Mr Clarke.

Companies usually employ engineers to study the operations of specific pieces of plant and keep them running, but they often do not have time to understand how different aspects of the plant interact together.

'Those highly segregated siloed parts of the organisations previously haven't had to work together,' Mr Clarke says. 'That's the stumbling block that has to be overcome.'

Finding people

Achieving data analytics takes a wide range of different skills, including people who are specialists with working with large data sets, people with a specialist knowledge of the industry, people who are good at working with probabilities (risk managers, statisticians), and people who can train others how to work with the data.

Tessella aims to help here, by providing staff with all the different skills required on a consultancy basis, who can come to an oil and gas industry and help put together a program to solve the problem.

Tessella can provide consultants who have worked in all areas of the oil and gas industry, both at operators and service suppliers, and also analytics, simulation and modelling specialists, who can come into the company and try to work out the right answers.

'There is a big variety of skills you need in an analytics laboratory to solve a specific problem,' Mr Hendriksen says. 'You don't need all of them, but you need every time a different mix of them. Tessella has all of these skills in our organisation.'

Tessella has UK offices in Stevenage and Abingdon in the UK; Stevenage is close to Cambridge and Abingdon is close to Oxford, and it employs many people who are graduates of those universities, who have the technical analytical skills.

Project examples

Tessella was engaged in a project along these lines for the nuclear re-processing industry, where there client was struggling to understand a mechanical problem. Tessella staff analysed the acoustics of (sounds made by) the machinery.

Another project was for a UK water company, which had to make predictions of what it would spend over the next 5 years, to be presented to the regulator, which it would then use to calculate water bills.

The analytics included expected degradation of the assets, planning for extreme events, and understanding the real time data coming in, and looking at financial risks.

Tessella's staff could understand how the company works, build small software tools to improve the numbers being fed into the existing enterprise systems, and build business processes, and train company staff so they can do it over the longer term

Associated Companies
» Tessella
comments powered by Disqus


To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.


Latest Edition Mar-Apr 2024
Apr 2024

Download latest and back issues


Learn more about supporting Digital Energy Journal