[Editor's note: This story originally appeared in the March 2020 edition of E&P. Subscribe to the magazine here.] 

When the digital oil field arrives, which it really already has, there will still be a significant human component, at least for the foreseeable future. While artificial intelligence (AI) and machine learning can process huge amounts of data to make operational decisions (e.g., detecting a full tank and shutting off a pump or rerouting its flow), many of those decisions involve weighing too many and often subjective priorities to be left to AI and machine learning alone.

What these ever-growing digital tools can do, however, is allow a small number of humans at monitoring centers to track dozens of wells across hundreds of miles of remote landscape, deploying additional resources only when there is a perceived problem detected by this human-machine operations team.

The field of automation has advanced greatly in a short time. Just a few years ago, water treatment plants required people on site to attend to valves, pumps and sludge tanks or to manage the flow of vehicles in and out of the facility. They were standing still doing two things: watching for a problem to report or listening for a call telling them to adjust a pump, spin a valve or run a water quality test. There would seem to be few things less humanly productive than that antiquated operation.

Early SCADA systems reduced the number of field personnel as sensors began doing the monitoring, sending data to computers programmed with alarms that alerted a much smaller staff to make adjustments or repairs. By taking over the most monotonous and unproductive aspect of human work, automation made personnel more efficient and valuable.

Automation’s next evolution connected actuators to control valves in the field, enabling one or two technicians in an onsite control room to respond to many, often simultaneous, alarms by telling the computer to make specific adjustments.

The latest systems take that one step further, making it possible for all monitoring and control to be done off site. It does this by adding camera surveillance to the mix.

For example, should a tank level gauge trigger an alarm because a tank is filling up, the latest software allows remote technicians to access a camera view associated with the alarm to verify the situation.

With sophisticated hardware, software and communication systems, a small number of people at a single monitoring point, such as SitePro’s 24/7 Remote Operating Center (ROC), can monitor and control facilities in multiple fields across an entire basin.

The irony in removing oftentimes all people from the site is there are actually more eyes on the operation than there were when people were on site because those alarms can be sent out to multiple personnel instantly. Those individuals also can pull up all the historical data including the visual to then decide how to proceed and if a site visit is needed. Anywhere from 60% to 80% of onsite issues can be resolved remotely.

While this kind of technology benefits all aspects of the energy industry, it is particularly suited to the water midstream sector, which involves assets that spread widely across basins. Midstream’s entire existence revolves around transporting water across great distances, which would make in-person monitoring extremely inefficient.

In addition, remote monitoring greatly reduces the miles driven, which increases safety. It means technicians sent to sites are more likely to be fully prepared for what they will find with parts, tools and expertise, and therefore able to get the system back up and running without extra trips.

Even when everything is in order, this kind of system is vital in managing the flow on a daily or even hourly basis. Should the input level at one disposal well reach capacity, remote monitoring personnel can divert an appropriate amount of water to a different well that has available capacity.

Next steps
The near future looks even brighter for remotely aggregated monitoring and control. In addition to processing and analyzing floods of data, AI aggregates information as it goes. The industry is working to train AI to assess trends for their predictive value.

One option will be to monitor water flow throughout a system to determine the source of the highest volume of skim oil. Armed with those data, operators or disposal companies can divert the most oil-rich water back through a separator to collect and profit from a product that would otherwise be lost to a saltwater disposal well.

This system also can remotely monitor water flow and quality to adjust chemical input as those elements change throughout the day. This can save on costly chemical injection when flow rates are down, yet make sure they are adequate to protect the formation when flow rates or contamination levels rise.

Because SitePro’s ROC system also monitors an extensive list of pump information beyond flow rates and voltages (e.g., temperature, stroke, fault codes and more), the plan is to use those data to make pump adjustments to maximize efficiency and predict maintenance needs. By reducing downtime in this way, producers face less risk of produced water backups that in a worst-case scenario can require wells to be shut in while the problem is resolved.

Autonomous monitoring and the bottom line
With private-equity firms changing their requirements from growth to current profitability and cashflow, everyone in the oil patch, including water midstream firms, are investigating ways technology can reduce costs and improve efficiency. The most efficient way to employ technology on this front is to outsource it to third-party experts with existing systems, rather than spending vast amounts of time and money on creating an internal automation system. Investors may not be willing to wait the months or years for internal development. Most are looking to invest in companies that can demonstrate a solid bottom line now.

Will automation ever completely replace humans? While science fiction likes to explore how that would look, the truth is that in today’s world computers cannot think creatively. They can excel at one thing: rapidly doing what humans tell them to do. And that is exactly how they make humans much more efficient in a variety of industries, including in water treatment and transfer.