Functional safety: New challenges in smart factories

In simple terms, functional safety means that people, equipment and the environment are physically protected from malfunctions of a machine. In addition to appropriate design principles, this also includes automatic protection systems that can detect and remedy faulty behaviour, for example in machine tools or robots.
 

Various relevant standards

A number of standards exist to establish functional safety. For automation technology – more precisely for electrical, electronic and programmable electronic systems – IEC 61508 is the basic standard. It defines four safety requirement levels for safety functions. These levels are called Safety Integrity Levels (SIL) and range from Level 1 (lowest requirement level) to Level 4 (highest requirement level).

Further standards have been derived from IEC 61508. IEC 62061, for example, applies to industrial devices. It is supplemented by other, device-specific standards, such as IEC 61131 for programmable logic controllers (PLCs) or IEC 61800-5 for electrical power drive systems with adjustable speed. DIN EN ISO 13849-1 provides safety requirements and a guideline for the principles of design and integration of safety-related parts of control systems. For the safety of industrial robots, the European harmonised standards EN ISO 10218-1 and EN ISO 10218-2 apply. The technical specification DIN ISO/TS 15066 also defines specific safety requirements for collaborative industrial robot systems and the working environment.
 

Control systems and sensors

Embedded systems are the core of most automation systems and thus also of protection systems. In order to ensure functional safety, both hardware and software must be designed securely. Thus, special microcontrollers and processors can already be equipped with integrated safety mechanisms, for example with a lockstep architecture. This is a self-checking security technology in which two CPU cores permanently check each other when executing operations. This enables manufacturers of automation components and machines to develop and validate functionally safe systems much faster.

In addition to computing components, sensors are also potential sources of error. Undetected sensor errors can lead to serious risks. The probability of a functional error can be reduced by checking the plausibility of the measurement data. This can be done either in the sensor itself, which is proven by appropriate certification according to the relevant standard, or in a higher-level control system. In the latter case, two redundantly designed sensors are typically used.
 

Collaborative robots

A hallmark of the flexible factory according to the Industrie 4.0 concept is also closer cooperation between humans and machines. Thus, many companies are opting for collaborative robots, cobots, which can literally work hand-in-hand with humans. However, this requires sophisticated safety features: For example, safety torque sensors can detect when the robot arm encounters an obstacle such as a human colleague and reduce the speed and force of the movement. Other systems work with touch-sensitive surfaces that are directly connected to the motion control.

The role and need for functional safety changes depending on the degree of interaction between robots and humans. (Source: Intel)

Constantly changing cooperations

The connected production of the future is based on cooperation between a wide variety of flexible systems. This poses completely new challenges for functional safety: Because previous solutions are based on determining the risks for specific operations - at fixed positions in production. However, if the production processes and the cooperation partners are constantly changing, these arbitrary combinations do not have to be safe, even if each individual system is functionally safe on its own. The individual modules must therefore still recognise hazards in the smart factory and react accordingly when the production sequence, their position or the cooperation partner changes.