Robot Removes Operators From Extreme Environments
Robots have the potential to move human operators away from uncomfortable, potentially risky environments and into comfortable, safe control rooms. Remotely operated vehicles have already achieved this for subsea fields; however, before this approach can be extended to surface facilities, the robots must be reliable and safe in potentially explosive environments. The Sensabot robot has addressed these challenges and could be the foundation on which future generations of robots are built.
In 2010, a technology plan was prepared that focused on the specific challenges facing projects in the Kashagan Field in Kazakhstan:
- Climatic temperatures typically ranging from –25°C to +35°C
- High hydrogen sulfide concentrations in produced gas
- Raw-gas-injection pressures as high as 690 bara
- Ice-bound unmanned artificial islands in the winter
These challenges require operators to wear breathing apparatuses and cumbersome insulated clothing in winter that hampers their movement. In summer, the breathing apparatus creates the risk of heat exhaustion.
One contribution to the technology plan was the concept of remotely operated robots. These could remain permanently on location and could be driven by operators located in a safe environment.
The Robot Concept
The first stage of the Sensabot project was to identify the robot’s high-level functional requirements. These fall into three categories: user acceptance, safety, and independence.
User Acceptance. The project team recognized that, although the upstream oil and gas industry has been using remotely operated vehicles for many years in subsea environments, there has been little significant use of robots for surface facilities. The team felt that onshore operators would view the robot with suspicion, so it needed to work reliably to build confidence. By simplifying its functional requirements, the project team could focus on fewer features and decrease the risk of failures.
In this context, it was decided to focus on simple sensing tasks rather than the manipulation of items such as valves and switches. Therefore, Sensabot was designed to perform daily operator inspections of the plant. It was equipped with a range of sensors that replicated those of a human operator (Fig. 1): cameras (sight), gas detectors (smell), a vibration detector (touch), and a microphone (hearing). Later in the development, a thermal imaging camera was added.
Another important design decision was to minimize the amount of automation. Sensabot was not designed to eliminate the need for human operators, just to relocate them to a safe and comfortable environment.
Finally, Sensabot was designed to work on plants that have been designed for human operators. Sensabot was scaled to mimic humans (Fig. 2) and to navigate 1-m-wide corridors with 90° corners. This also gave Kashagan management the comfort of knowing that, should Sensabot fail to perform, human operators could readily take over its tasks.
Safety. The overriding safety requirement is that Sensabot should be able to operate in oil and gas environments. In this respect, the project team set its most demanding requirement. Although Sensabot will usually be operating in Zone 2 and explosion-safe environments, it is being certified for International Electrotechnical Commission—Explosive (IECEx) Zone 1 IIB environments (hydrocarbon gas will be present or can be expected to be present for long periods of time under normal operating conditions).
Independence. When establishing the user requirements, it was recognized that there was no point in removing oilfield operators from the field if the robot required frequent maintenance interventions. Therefore, the design goal from the outset has been that Sensabot should operate for at least 6 months without human intervention.
In addition to its reliability, Sensabot needed to be charged while on location. Therefore, an integral part of the robot system is its kennel. The kennel is sufficiently robust to protect Sensabot during transport. Once on location, the kennel is plugged into a 110/240-V power supply. When Sensabot docks in the kennel, a charging connector engages and the operator can switch on the power to charge the batteries for the next mission. The batteries are specified for 3 km of driving, navigating 40 m of rise and fall, and performing 4 hours of sensing operations between recharges.
Challenges and Solutions
IECEx Certification. Conventionally, IECEx certification for explosive environments is applied to simple stationary instruments or containers. However, Sensabot contains 19 different assemblies that need to be certified, 43 in total if duplication is included.
Sensabot Mark 1 attempted to tackle this complexity by installing the majority of these assemblies in an overpressured body. However, this presented a number of technical challenges: The kennel needed to replenish the internal air pressure automatically, the pressure-control valve proved to be unreliable, and leakages around the body’s seals meant that the operational life between replenishing pressure was too short.
Another solution was to use already available certified assemblies. However, because of Sensabot’s small space envelope, most of these were too large.
The approach that was eventually adopted for Sensabot Mark 2 was to design customized assemblies that could be certified individually. This means that the certification process is far more complex than the norm, especially because each assembly must be certified across Sensabot’s full operating range.
Wireless Communications. The wireless link between the robot and its driver is critical if Sensabot is to operate effectively.
Sensabot Mark 1 was designed to operate using a conventional WiFi system similar to that in homes. However, modeling and trials revealed that a large number of WiFi access points would be required to cover even a small production island. More seriously, because of the way WiFi works, the signal to the robot would often be lost when a handoff occurred between access points.
A range of alternative wireless networks was evaluated, and the conclusion was that 4G Long-Term Evolution (LTE) networks offered the best solution. While LTE has less bandwidth available per radio than WiFi, it has more than enough to operate Sensabot and an extensive network of other wireless instruments. Also, signal loss is progressive and gradual, so, if the driver sees that Sensabot is entering an area with a weak signal, the operator can reverse away before the signal is lost.
Sensabot Mark 2 is equipped with both WiFi for small-scale demonstrations and 4G LTE for full operational deployments.
Functionality. Sensabot Mark 1 underwent a series of trials in Houston in early 2011. Overall, the trials were a success, with Sensabot proving easy to drive even for inexperienced operators. It performed a wide variety of inspections and traveled many kilometers around the plant, accessing elevated platforms and navigating dark rooms.
However, a number of design weaknesses were identified that have influenced the design of Sensabot Mark 2. In addition to the move away from an overpressured body and the shift toward 4G LTE wireless,
- Solid tires were replaced with pneumatic tires. These are less prone to wear on metal grids, create less vibration when driving over rough surfaces, and provide better traction across a wide range of surfaces.
- A move was made from the use of light detection and ranging to stereo cameras to map the surrounding terrain. This is an advantage for Sensabot because stereo cameras can create complex 3D models of the environment that may eventually allow Sensabot to operate with more autonomy. They also can be programmed to detect drops and to intervene before the operator drives over one.
- A thermal imaging camera was added.
- Maximum speed was increased from 5 to 7 km/h.
Deployment and Benefits. Once Sensabot Mark 2 is certified, the plan is to deploy it at an upstream production facility. Ultimately, the intention is to install the wireless network, kennel, and Sensabot at an unmanned location where Sensabot can fulfill its full potential. However, in the short term, it needs to be tested and proved with a minimum of disruption. Therefore, the entire system, including the control panel, is being built into a shipping container.
Once the container is on location and plugged into a power supply, the system is ready to operate. This minimizes the effect on busy oilfield operators in the early stages of deployment.
Enhanced Functionality. As the project has developed, it has become clear that the components and principles that make up Sensabot could be combined in a variety of packages to suit a range of tasks and operating environments.
For example, as an offshoot of the development program, a manipulation arm was developed and demonstrated. It performed a range of tasks including operating valves and electrical switches. In this respect, it has one major advantage over remotely operated vehicles in that it has massive inertia, a result of its 450-kg mass and low center of gravity. This means it can operate heavy-duty oilfield equipment without special reactive tooling.
Finally, it is entirely plausible that a higher degree of autonomy could be incorporated into Sensabot Mark 2 and future robot generations. An early target will be the incorporation of collision avoidance. Other simple ideas include self-navigation from the kennel to the operational location, auto-reversing in the event of wireless-signal loss, and auto-diagnosis of the sensing data with alarms to alert the operator in the event of anomalies.
This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 181409, “A Robot That Removes Operators From Extreme Environments,” by Ian Peerless, SPE, IPKA Consultancy, and Adam Serblowski and Berry Mulder, Shell Global Solutions International, prepared for the 2016 SPE Annual Technical Conference and Exhibition, Dubai, 26–28 September. The paper has not been peer reviewed.