What are the best practices for large-scale robot deployment, support, and maintenance?
Last month’s column briefly touched on the topic of incorporating mobile robot deployment into standard operating practices. This time let’s look at the best practices for large-scale robot deployment, support, and maintenance.
One way to breakthrough and expedite the process is to think of it in terms of a new technology introduction in the organization. Practically all companies that are in the process of digital transformations have tackled this issue. Managing the fleet of robots includes monitoring the robots and their execution, managing abnormal situations, logistics such as robot service, making sure robots are ready to do their work with the appropriate sensor payload and manipulating capability, and making sure all necessary parts are in stock locally or otherwise available in a timely manner to ensure operation and maintenance.
Robot manufacturers and integrated solutions or service providers also offer more specific guidance in this area including modular solutions and flexibility. For the end-user, a key pointer in selecting among the providers is how they match with the company's business needs from end-to-end, the cost of deploying such technologies and alignment with overall digital transformation plans.
Some good news is that end-users have recognized this issue and are actively pushing to improve fleet management, for example, robot-as-a-service, data-as-a-service, inventory management, remote operation and maintenance, and storage of the robot/asset. These have become routine matters, which are very well-supported by manufacturers and integrated solution providers.
Early on, management of the batteries that power the robots appeared to be a significant issue; however, very good pointers on battery life, operating practices to maximize battery power efficiency, and charging practices have advanced and are widely available, today. One potential plus is standardizing the charging station.
So, now, what are the lingering issues which are keeping end-users from pivoting off the plateau and making big progress with mobile robots?
Key issues continue to be mission management and data compatibility with other systems throughout the enterprise. Solutions are available, today. In this segment, we will dive into mission management. Data compatibility will be the focus of my next article.
Why is fleet management so important? It is because the ability to remotely control robots from hundreds or even thousands of miles away is one of the key objectives. It means that a human worker does not need to be present in the dirty, dangerous, demeaning environment.
If they don’t seek guidance from robot manufacturers or integrated solutions providers, end-users could struggle with mission management. A mission, during which a robot travels to one or more specified locations in the plant, could be for a broad range of assignments. Mission management ranges from ensuring robots received their assignments to confirming the tasks are completed. Under some abnormal situations, management must reassign unfinished tasks to a different robot.
Most often, the robots perform routine inspections that simulate operator rounds. A simple start would be to assign robots to the exact rounds that human operators or technicians would perform. For end-users who have optimized their rounds, this would be a great way to start. On the other hand, we have observed many opportunities for improvement. Missions can be streamlined to minimize the duration and battery power use. Rounds can also be designed to make the most of the robot’s capabilities. As I mentioned last month, the robot’s ranges in terms of hearing and sight far exceed those of humans.
Proof-of-concept (PoC) tests have shown that doors and stairs could present problems to robot rounds. Although models such as the Boston Dynamics SPOT can handle such obstacles, it is still best to minimize them during missions. It is a simple matter of risk management. The more stairs the robot must climb, the greater the risk that it will encounter a problem. For this reason, many of the green field designs or even brownfield modifications are considered to design robot-friendly facilities.
PoCs have also shown how to minimize the involvement of people who prepare robots for missions, transport them to mission starting points, perform simple local repairs, and return them to storage upon completion of the missions.
Some of the tests have encouraged users to occasionally install wireless sensors instead of using a robot in a particular location. If one location is an “outlier” physically, it could be very inefficient to send a robot there on an inspection round. A wireless sensor could be more cost-effective. The only drawback, potentially, is that wireless devices require maintenance. Sooner or later, a human technician will need to visit the site to change out the battery in a wireless sensor. Perhaps, in the future, a robot could perform this task! Finally, selecting the best technology is based on business value and future considerations.
In other cases, operators have found that they would prefer to see updates to some process variables more often than daily or weekly rounds allow. Process operations are better served by sensors that provide more frequent updates. Robots are typically used for asset management and maintenance purposes. They bring flexibility and, essentially, additional eyes, ears, noses, and hands to the field. While it is necessary for human operators to observe the plant on a 24/7 basis, robots are the best human companions and tools for remote operations.
Finally, managing payloads is another aspect of robot fleet management. Typically, there will be a variety of camera/sensor arrays that depend on the purpose of the missions. For example, visual inspections, emissions detection, and mapping surveys would use different payloads. Users have found that it is best to run all missions using one particular payload before changing it out for a different type of mission. In this case, multiple robots carrying different payloads are running in the facility for different purposes. Fleet management will function much like an aviation system in the field.
Penny Chen
Yokogawa
Senior Technology Strategist: United States Technology Center