Home >En route to robot safety
En route to robot safety
31 July 2019
As the trend moves away from full enclosures for robot cells in favour of human-robot collaborations (HRC), which manage without guards where possible, but still guarantee operator safety, it is nevertheless still important to remember that each application will require a full risk assessment and verification of the safety measures required
Robots are classed as partly completed machinery in terms of the Machinery Directive. The two harmonised C-standards ISO 10218 “Safety of Industrial Robots” Part 1: “Robots” and Part 2: “Robot systems and integration” – which also contains information on collaborative operation - are available for detailed safety requirements and there is an accepted de facto requirement for all safety functions to meet Category 3 / PL d of EN ISO 13849-1.
Steps toward a safe HRC application
When the normative specifications are implemented, the fact that robot cells are classified as a machine under the Machinery Directive means that each step of the conformity assessment procedure must be completed. As the robot itself is only regarded as partly completed machinery, it is not until the end effector or the necessary tool for the respective application is in place that the robot achieves a specific purpose and must then be regarded as final machinery. The integrator – who could also be the user - becomes the manufacturer of the machinery and is responsible for the safety-related inspection, including CE marking.
One of the most important points en route to achieving a safe robot application is to produce a risk analysis in accordance with EN ISO 12100. On robot applications, the challenge for the “risk assessment” is the dissolution of the boundaries between what used to be clearly separate work areas for human and machine. The operator’s movement must be considered in addition to the hazards emanating from the robot. However, these cannot always be calculated in terms of speed, reflexes or the sudden approach of additional people.
Based on the risk analysis, the next steps are the “safety concept” and “safety design”, including component selection. The results from the “risk analysis” and “safety concept” are used to document the selected safety measures in the “risk assessment” and to implement these in the “system implementation”. This is followed by the “validation”, in which the previous steps are re-examined. Validation is essential for proving that machines are safe.
The checklists in EN ISO 10218-2 provide additional guidance for robot applications. By attaching the CE mark, the integrator ultimately confirms that the robot cell meets all the legal requirements of the Machinery Directive 2006/42/EC when used in accordance with its intended purpose and is safe for use.
Selection of robot and safety components
A wide range of robot systems is available on the market, which are suitable for various application areas. In order to implement a safe HRC application however, we will increasingly need safety systems that are considerably more intelligent so that work areas in which humans and machine collaborate can be designed safely.
These systems may be part of the actual robot control system, for example for calculating the robot’s movements safely. In many cases however, safe motion functions such as these will not be enough to achieve the safety objective. Combinations will often be needed, which will include near-field protection (e.g. tactile safety sensors) and safe sensor technology for monitoring the detection zone, such as the PSENscan from Pilz, for example.
Integration of sensor, control and actuator technology opens up new freedoms when it comes to planning dynamic process cycles and work areas in which human and robot interact safely.
Ultimately a safe HRC application is the result of several factors: reference to the normative framework conditions and a complex risk analysis on this basis, selection of a robot with the relevant safety functions, selection of appropriate, additional safety components and finally, validation.
The main areas for HRC in the future will be in industrial applications but there will be some challenges with regard to safety; firstly, there isn’t one safe robot or one safe sensor technology that covers every eventuality, in terms of safety. There will be a range of technologies and these must be safety certified. It’s the job of the integrator to describe the hazards and implement solutions, based on a risk analysis. For example, contact with humans should be ruled out where there are hot work pieces or sharp tools. In this case it makes sense to detect the coming together of human and robot using for example tactile mats or laser scanners. It depends on the respective scenario and environment.
Practical guidance for integrators
When planning an HRC application, robot selection is an important element for the integrator. Various methods should be applied to verify and validate safety requirements, including visual inspections, practical tests and measurements. The integrator has to verify or validate over 200 points in all.
To simplify implementation of the normative specifications for HRC and demonstrate the ways forward, the Technical Specification ISO/TS 15066 “Robots and Robotic Devices – Collaborative industrial robots” was developed. This substantiates solutions for safe human-robot collaboration in an industrial environment.
Four types of collaboration are described in ISO/TS 15066 as protection principles.
- Safety-rated monitored stop
- Hand guiding
- Speed and separation monitoring
- Power and force limiting
When implementing a safe human-robot collaboration (HRC), integrators can choose one of these “types of collaboration” or a combination of them for their application.
A body area model has been defined in the Annex to the Technical Specification (TS). The model defines points in the corresponding body areas, with details of the respective pain threshold. These pain threshold values can be applied as validation for safe HRC. The body area model provides details of the respective pain threshold for each part of the body (e.g. on the head, hand, arm or leg), which marks the start of the pain threshold. If the application remains within these thresholds during any encounter between human and robot, then it complies with the standard.
There are no safe robots, only safe robot applications!
Implementing human-robot collaborations in an industrial environment is definitely going to increase; however, its growth will be heavily dependent on innovations in the fields of sensor technology and robotics. Together, automation engineers, robot manufacturers, integrators and notified bodies will be able to make the vision of a robot workmate reality on a step-by-step and application-by-application basis.