ISO/TS 15066 and collaborative robot safety
Guiding collaborative robot design
By Mathieu Bélanger-Barrette
In recent years, many disruptive technologies have emerged. For example, driverless cars have gathered a lot of attention, because they are so different from what we are used to seeing. In fact, driverless cars have a bunch of sensors that make them safe. All these sensors allow the car to predict upcoming dangers and avoid crashing into other objects. However, even if they appear to be really safe, they are not yet completely allowed on the roads. They need to be tested and approved first.
The same thing is happening for collaborative robot technology. In fact, these robots have been on the market for a couple of years now and just recently have been the main subject of an International Organization for Standardization (ISO) technical specification. The big concern revolves around how these robots operate alongside humans, making sure each operation is safe.
A collaborative robot is a robot that is aware of its environment. In all four of their collaborative modes, robots have sensors that allow them to feel the presence of humans or any abnormal presence. The most recent type of collaborative robot is the power-and-force-limited models. They are equipped with force sensors that stop movement when they detect contact with some object.
To be even more precise in defining power-and-force-limited robots, as opposed to industrial robots, these robots stop immediately when certain forces are felt. They are designed with compliant materials and round shapes, which allow them to spread and thus disperse the pressure of an impact. You can also adjust the force and speed settings to make them work a little slower when they are near a human or a little faster when they are working solely around other machines. They are designed to be easy to program, since they use the same force sensors to read the force used in their motors and they allow you to program them by demonstration (i.e., teaching the robot the moves you want it to make). They are not as expensive as you might expect compared to traditional industrial robots. Collaborative robots are currently the best setup available for many tasks— safe and easy to use. Also, it is fast to recoup the expense of your investment in them.
In fact they are really convenient. On the other hand, to make sure they are safe, you need to perform a risk assessment. This process illuminates the problems you might encounter with your “cobot,” and it will determine the difference between a safe and a dangerous process. More importantly, the robotic application itself has to be assessed to make sure your employees are working in a safe environment.
Why do a risk assessment?
Ensuring the safety of your workers is your responsibility. The risk assessment is a tool to help you achieve worker safety. In some areas of the world, local laws and regulations require a risk assessment for any machinery installed on the factory floor. Not all governing regions require you to respect the ISO standards, but most safety institutions have laws incorporating or referring to safety standards. This also applies to large corporations with their own internal safety guidelines. So you will most likely have to comply with some standard, ISO or other, when integrating and designing your robotic cell in order to respect the laws in your region and ensure worker safety.
The risk assessment is then based on whatever standard you have chosen: either for your region or perhaps an even higher standard of workplace safety at a global level. If you want to ensure the safety of your robotic cell, you need to perform tests and adjust your cell in accordance with all the points required by the standards you have chosen. You should also document your cell’s performance in accordance with this standard. In any case, you would want to do this to reduce your liability in the event of an accident.
Risk assessment process
The risk assessment process basically means that you will analyze all the motions, interactions, and operations the robot performs during normal operations and ideally even those from installation to decommissioning. One complete operation might be divided into separated tasks. Then each task is evaluated and rated for its risk level. With certain criteria, these risks are classified, and those considered dangerous must be reevaluated to reduce the risk as much as possible within the limits of the operation.
That is the briefest way to see a risk assessment, however, it is a complex process, and all motions are considered risks. In fact, the simple motion of closing a gripper can be considered a risk. To know what type of risk is associated with the closing action of a gripper, for example, you need to consider several aspects: the severity of the risk, the possibility of avoiding the risk, and the redundancy of the risk. Each of these criteria can be evaluated in the following manner:
- Severity of risk: The principal way to evaluate this criterion is by evaluating the estimated force or pressure that is to be applied on the human body part by the robot. The new ISO technical specification has a complete chart equating human body parts with a pain value. There is a difference between a quasi-static impact and a transient impact. The first is an impact against a fixed object or a pinned object, and the second is an impact that allows for the free movement of the body part. In general, a factor of two is added to the maximum force/pressure, which basically allows the robot to hit you with two times as much force if you are free to move and recoil. The least severe or more acceptable risks are rated the lowest.
- Possibility of avoidance: This factor is normally evaluated in line with the speed of the robot or the environment where the robot is operated. For example, compare a robot running at full speed versus a situation where a robot is running at half its capable speed. You clearly have more chance to avoid the slow robot than you do the faster one. The more chance you have to avoid a risk, the lower the number.
- Risk redundancy: This factor is evaluated according to the number of times the risk is occurring within a fixed amount of time. For example, does this risk happen every 2 minutes during the application or once per month? Again, the less frequent the risk is, the lower the number.
Note that specific data for body parts, force/pressure, and pain levels are listed in ISO/TS 15066 – Robots and robotic devices – Collaborative robots.
As mentioned earlier, the score represents the danger of the risk. The highest scores should be reduced as much as possible to obtain acceptable levels. But it is very important to keep in mind that every time you reduce a risk, you need to make sure that it does not increase another risk at some other point in your application process. At the end of the day, there is no rigid guideline on what is dangerous and what is not. It is your responsibility to make sure your robotic cell is safe enough for human-robot collaboration.
As Lars Kieffer of Universal Robots said, “Risk assessment is not about avoiding hazards altogether, but about choosing one risk over another. For example, if you’re looking to cross the street and a bus is coming and there’s a 5 percent chance that you would be run over, then it’s unlikely that you would take the risk. But if there is a child running after a ball crossing the street in front of the bus, you would not hesitate to run in front of the bus to save the child.”
In most robotic installations, ISO 10218 – Robots and robotic devices – Safety requirements for industrial robots is used as the standard. This standard was release in 2011, before the general introduction to the market of collaborative robots, so very little information was available on collaborative robots at this time. This is why the development of ISO/TS 15066 is so important. It is a technical specification that gives guidelines specifically for the use of collaborative robots. Because the field of “cobots” is still developing, the general standard for cobots is also still developing. This is why the technical specification will not really be official until all the data and authorizations have been confirmed.
ISO/TS 15066 and risk assessment
These new specifications have a direct effect on the risk assessment process, because they provide a method of calculating the direct data in the risk assessment process for the severity of risk and the possibility of avoidance evaluations. In fact, to determine whether or not a risk is dangerous, you need to know where to draw the line. If you do not know what type of impact will cause what type of injury, you cannot just throw numbers on a sheet and expect your robot to be safe.
Since the technical specification estimates many different human pain levels according to different types of impact, it is easier to draw that line. For example, a human hand can accept 135 N or between 200 and 295 N/cm2. With that information, you can determine the maximum force a robot can apply for a given robot configuration. Suppose a robot is moving along the z-axis with its gripper closed and crushes a hand between itself and a table. What is the best way to evaluate the severity of an impact like this?
- Determine if the impact is quasi-static or transient: In this case, the impact is quasi-static. In the case of a transient impact, the force and pressure can usually be doubled. This means that an impact with a hand that can move freely can have a greater force than one where the hand is pinned between the robot and an object.
Determine the robot’s maximum force: This information can be found in the safety settings of the robot. If this information is not available, you can test the force of the robot with a force gauge.
Determine the body part area, the pressure, and the pain level: Depending on the region of the body where the force is applied, the amount of pressure required to cause injury will change. This can be the difference between a safe and an unsafe robotic action.
Do the math: Calculate the pressure (N/cm2), and analyze if the pressure is lower than technical specification recommendations.
Conclude: If the pressure applied is lower than the ISO/TS 15066 recommendation, you are good to go! If not, you need to perform a risk reduction process.
Obviously, this is just one piece of an example, and the standard does not boil down to this simple a process, but perhaps this example can direct you and help you to figure out how the new technical specifications can be useful for your collaborative application. The fact that the new specs provide data and a framework for analyzing the risk of injury is particularly helpful when completing a risk assessment.
Completely safe robots?
We are often asked if it is possible to buy a totally safe robot. And the question is quite hard to answer. Some robots are marketed as being safe. They are indeed aware of their environment and will stop if they hit you, but the application itself has to be evaluated in its entirety. For example, the risk assessment process will be different if the robot is carrying a sharp metal part or if it is carrying a soft rubber duck. The same argument applies if the robot is operated on a mobile base or on a fixed table. The context of different operations has to be evaluated and considered in the risk assessment. So buying a safe robot, even if it is possible, still needs to be analyzed afterward to make sure the operating environment is safe. This is especially true in cases where there will be human-robot cooperation.
People also often ask if they need to update their collaborative robot cell since this new technical specification was introduced. The ISO/TS 15066 was released in January 2016. The data and information it provides is available as a recommendation or for guidance. This means that you do not need to adapt your actual robotic cells to this new specification in order to be approved by a third party. However, I do recommend compliance with this new technical spec for two reasons.
First, the safety of your employees is critical. You want them to be working in a safe environment, and this technical specification will help you move in this direction. Second, this technical specification is very likely to be adopted in the next few years. If you are not complying with the standard now, you will need to do so one day. Why not start today?