You are Home   »   News   »   View Article

When automation isn't mechanical

Tuesday, September 4, 2018

Automation isn't necessarily about replacing people's jobs with machines. It can also be about automating processes with software to put the right information in front of the right people to help them do their jobs better. We interviewed Peter Zornio, CTO of Emerson Automation Solutions, to find out how it works in industry today.

When most people think of 'automation' they think of automating machines, to do mechanical jobs, like robots.

But to Peter Zornio, CTO of Emerson Automation Solutions, perhaps the more interesting applications of automation today are where software automation helps people to do a better job.

One of the best examples of this, one we are all familiar with, is Google Maps' capability to suggest routes to us, and make use of all kind of dynamic information to do this, such as changing traffic congestion or public transport delays.

Mr Zornio says - this is not a computer telling us what to do next, but a computer guiding what we are going to do, and alerting us when something has changed so we should do what we normally do differently (such as a traffic jam 30 minutes away on our usual route). The software has all kinds of analytics and expertise happening in the background to provide us with the most useful information. The process could also be called 'decision support'.

Today's automation tools can also help workers acquire knowledge faster, by providing them with what they need to know, for example guiding them about upcoming problems with a piece of equipment. It can provide access to information onto mobile computers - making data freely and securely available wherever it's needed

You can differentiate the various types of automation by talking about 'physical automation,' where a robot does a mechanical process and 'software automation', where a software tool is making your work easier or better. For example, autonomous cars are physical automation.

The oil and gas industry does have some mechanical processes which could benefit from physical 'programmed' automation - drill rig operations, dynamic positioning of vessels, automated streamer positioning on seismic vessels, programming a drone inspection along a pre-planned path, or a system to inject chemicals when certain decisions are made. These are important, but only a small part of the overall industry. Working in the oil and gas industry, whether in exploration or production, is far more about making decisions based on relevant, real-time information. So many tasks come down to someone making the right decision, for example with anything associated with reliability and maintenance.

Mr Zornio manages Emerson's automation technology strategy and standards.


Computers have many strengths compared to people, but the one strength people have compared to computers is 'synthesis', the ability to put together many different pieces of information and working out a pathway forward, Mr Zornio says. Computers cannot do this as well, but are improving.

Google Maps understands its limitations putting information together, which is why it often suggests a number of different routes you might like to take. There could be other information which even Google doesn't have, for example that you prefer buses to trains or like quiet roads. Or you might know that a rail replacement bus is often unreliable and don't want to take Google's first suggestion.

Similarly, a good system for an equipment repair engineer might give a number of different explanations for a fault, and the engineer can use the explanations together with other knowledge she may have to decide what to do.

The more sophisticated the computer system, the more confident it can be in its suggestions.

Consider the way that doctors work, Mr Zornio says. They may have computer tools which can help interpret specific results or scans, but only a doctor can synthesise the available information and make a diagnosis, computers are nowhere close - yet.

Or consider the challenges with autonomous cars. A human driver will see two 5 year olds playing with a football by the road and know to drive more carefully. A computer will see a road and obstacles in front of the car, but not have the experience to be more aware of children. But the computer can still provide value.

Industrial experts can use computer information to know that something may be close to failure. But only a person can work out the best path forward when taken in consideration with other information. How critical is that equipment in the current situation? If there is other planned work going on next week, perhaps it would be OK to risk things for a few days and change the component while the other work is going on. The decision requires 'context', also something which computers are very bad at compared to people. Computers are far stronger when they only need to work with the information right in front of them.

Keeping equipment autonomous

An interesting trend is that companies are proving much keener on keeping decision making about autonomous machinery made within the equipment or plant itself, not over a distance.

For example, Emerson is the main automation contractor for Shell's recently launched Prelude Floating LNG vessel.

All the control systems for the LNG equipment are onboard the ship itself, so there is no dependence on satellite data connection for core operations.

Similarly for autonomous vehicles, all of the decision making for controlling the vehicle is onboard the vehicle itself.

Onboard decision making is faster, and seen to be more reliable.

There can be remote monitoring, however, for autonomous cars and for the Prelude FLNG.

'It is the exact same thing for oil and gas as you would find in autonomous cars'. Today, people don't want critical functions to be dependent on an internet connection, he says. 'People want the first level of control on the ground'.

However, there are appropriate remote monitoring services that can offer important data for decision making. There can be remote companies studying how the algorithms work and evaluating how well they work - for example autonomous cars have companies like Google continually fine-tuning their algorithms, perhaps finding ways to improve how the machinery operates.

You could also have software in the cloud which helps people manage their work, for example storing photographs taken in an inspection to be shared. You can have remote systems for managing equipment reliability or sensor data about equipment health so you can predict potential problems and prevent unscheduled downtime.

Top quartile

Emerson has done some analysis into how oil companies compare in their operational performance. It found that the top quartile of companies tend to have both higher reliability and lower maintenance costs. Emerson found that top quartile performers spend half has much on maintenance compared with average performers and operate with an incremental 15 days of available production each year. They also spend 20 percent less on production-related expenses.

This is partly because typically about a quarter of companies are now active in using condition based maintenance, Mr Zornio says. This is tricky to do, but means that companies can get a sense of when something is going wrong, based on actual condition, before it does. So where other companies will change the oil every 8,000 hours, or whatever the manual says, a company with condition based monitoring might see that it only needs to be changed after 12,000 hours.

Confidence in the guidance

One of the biggest factors differentiating stronger and weaker companies is that for stronger companies, the staff have more confidence in the guidance that the computer offers, and so are more comfortable acting on it, and spend less time gathering their own data.

For example, many companies rely on Emerson's new suite of software solutions (from its recent acquisition of Paradigm) that provide improved modelling and better targeting for drilling, which will help them maximize recovery from reservoirs.
Mr Zornio is careful to say that the computers are providing guidance rather than rote instruction.

In the oil and gas industry there are still people who say they want to inspect equipment in person rather than trust what a computer says. Sometimes they are right. But still one of the biggest hurdles in getting a computer based monitoring system used is persuading people to accept what the computer says, he says.

Improving the performance of your equipment can be seen as trying to lose bodyweight, in that there is no big secret to it, but lots of tools that can help; provide you are willing to do the hard work to make it happen, Mr Zornio says.

FPSOs and unmanned platforms

In February 2018, Emerson announced it had completed a $90m automation project for BP's 'Glen Lyon' FPSO (floating production, storage and offloading vessel) West of Shetland. Emerson serves as main automation contractor. The project is expected to ultimately produce 130,000 barrels a day.

The contract covers 5 years and includes remote monitoring and predictive maintenance technologies, as well as support.

The systems should help BP improve reliability and availability of the equipment on the FPSO, optimise production and reduce operating costs. BP will also use Emerson's simulator tool to train operators and engineers on 'real world' scenarios before doing it on the real thing.

Emerson provides similar services for Clair Ridge, another 'mega project' from BP in the North Sea. It also manages the fiscal metering systems on all of BP's North Sea assets, generating data for accounting, custody transfer and taxation purposes.

Emerson also has executed a $17m automation contract for Premier Oil's 'Solan B' platform, one of the first oil platforms in the North Sea designed to be unmanned (most of the other unmanned systems are gas only). The systems for managing gas are a simpler than oil, without the complex separation processes, and so unmanned equipment is much easier.

Methane leaks

One automation topic of growing interest to oil and gas companies is how to manage methane leaks.

There is a backstory here that oil companies want to make a persuasive argument that gas is a much greener fossil fuel for electricity generation than coal - and a number of NGOs have claimed that the methane leaks are so high in the gas production process that is hardly true. Methane is a much more potent greenhouse gas than CO2 (although typically will only stay in the atmosphere for 8 years).

Gas can leak through gas-driven controllers and pumps, equipment leaks and seals, well venting, among other sources.

Straightforward gas sensors, which are fitted as standard on offshore platforms as safety equipment, can also be used to check for fugitive emissions. They will show that the gas concentration in the air is higher than it should be, or rising. Their use can be followed up with manual gas sensors which people can carry to find the leak more precisely.

Ultrasonic leak detectors can be used in relatively quiet facilities, covering around an acre (4000m2) of land. They can hear the sound of a leak.

For longer pipelines, computer modelling can, to some extent, provide an alternative to regularly spaced leak detectors. Emerson recently acquired a software company called ESI which has technology to make a computer model of the pipeline and calculate how flows should change along the pipeline, due to friction and turbulence. The flow can be measured every few km or longer if needed. If the measured flow is different to how the model predicts the flow should be, that can be taken as an indication of a leak.

Understanding flows using modelling is similar to a technique doctors use, when they want to work out if an artery is clogged without opening up the artery. They inject a dye in some part of your body, and monitor how the dye is flowing around your body, and compare that with how the dye would flow if no arteries were clogged.

Associated Companies
» Emerson
comments powered by Disqus


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


Skills and development pathways for the energy sector
Dr Lau Cher Han
from UCAN Technologies


Latest Edition Jan 19
Jan 2019

Download latest and back issues


Learn more about supporting Digital Energy Journal