major retail outlet currently offers a very cute robot manikin labelled “RoboSapiens —A fusion of technology and personality” with a $244 price tag. The reality in the manufacturing world is vastly different.
The manufacturing robots that will be discussed below are equally distant from those whirling dervishes one sees spewing pyrotechnics on motor vehicle production lines.
Just how far away from “cute” these manufacturing robots are can be is illustrated by a criminal prosecution by the Victorian Workcover Authority against a Geelong-based business.
Reported by Worksafe, September 4, 2003, the matter came before a Geelong magistrate who imposed a penalty of $200,000 after an employee died as a result of poorly maintained interlock switch.
The employer had identified the risk of a fast-moving robotic arm causing serious injury. The “control” installed was the interlock switch, which received no maintenance and therefore allowed the death of the worker when it failed.
Manufacturing in Australia is undergoing a tectonic shift and, in order to survive, it has to. The move is away from high-volume, low-value production toward both higher technology product manufacturing and an automated production line. Factors driving the shift include a diminishing skills pool, increasing labour costs and an ever-increasing cost of doing business, generally.
The trend is chiefly towards automation and emerging technologies. Process lines that were once labour-intensive are being closed down altogether in favour of machines that perform the task from start to finish, whether that requires bagging foam bed pillows in plastic covers or more complex assembly of multiple components previously handled by several teams of operators working on progressive stages of a production line.
The other new manufacturing trend is toward the use of emerging technologies such as that exemplified by manufacturers using exotic new carbon fibres instead of aluminium for goods from surf-boards to the Eagle Scout drone or the aircraft wings for the Boeing Dreamliner 787.
Automation has its own perils
One might think that installing an automated line would do away with OH&S issues, but these automated lines have brought their own peculiar OH&S issues, which are often fatal.
For example, a pharmaceutical company had to protect its equipment from unauthorised access. Forget the “primitive” machine guarding set down in AS4024.1: 2006. The new benchmark is to use biometric sensors needed to protect goods from tampering or theft by operators.
Another example illustrates a quite different aspect of robotics. An ergonomic issue arose where an operator periodically monitored and adjusted a robot with a moving arm.
Through an accidental design fault the operator had to perform an adjustment — from a position that brought his left hand directly into the arc of swing of the robot arm. He was naturally right-handed.
Because of restricted access to the robot plant, the company had to decide between redesigning the machine or relying on the operator’s agility in dodging the swinging robot arm. These examples illustrate the new hazards that can be created by this machinery, often without adequate guarding and in situations little considered by the creators of AS4024.1.
Safety issues are complicated by the fact that the robots may be powered by electro-mechanical energy, hydraulics and pneumatics. Robots typically have one or more arms inter-connected by sets of links and powered joints.
Robot arms comprise manipulators which support or move end effectors such as grippers, spot-weld guns or spray paint guns. All of this equipment occupies what is known as the robot operating work envelope. During operation, this envelope must be inaccessible.
OH&S problems arise when the robot has to be “tweaked”, for example by performing speed adjustments, making grip corrections and repairing breakdowns. Recent studies show that most accidents occur during robot programming, adjustments, testing, cleaning, inspections, repair and servicing. These are the times when the robot work envelope has to be entered, sometimes with the power on.
Developing an effective robotic safety management system (SMS) is the key to eliminating or minimising the foreseeable hazards, as required by Australian OH&S Law.
The first step is to consider all foreseeable hazards, which include the robot start-up sequence, program procedures, adjustments made during “normal” operations, human errors, robot malfunctions, control errors, unauthorized access to the envelope by a contractor, mechanical hazards and hydraulic or pneumatic hazards. Companies should also perform a failure mode effect analysis (FMEA), a test where a part of the system fails-and reveals possible safety hazards. This is a generic list and the reader must beware that specific robot applications may introduce hazards not listed above.
After identifying hazards the risk management principle requires the employer to move onto a risk assessment stage before identifying the appropriate control as per the Standard AS4360.
Designers of robotic plants would be well-advised to acquaint themselves with their “duties” as set down in the National Standard for Plant [NOHSC: 1010 (1994)].
However Australian law and conformance requirements are far behind the conformance requirements in America. An enquiry with the Australian Safety and Compensation Council (www.ascc.gov.au) which superceded the NOHSC regarding the current Australian Standard covering Robotics, found that “there is no such Standard.”
In America back in September 1987 the Federal Government issued Directive number STD 01-12-002: Guidelines for Robotics Safety. The directive sets out the conformance requirements of the American National Standards Institute (ANSI). Additional conformance requirements are set down in the NSW OH&S Regulations Part 3.2. While the Victorian Law is to be found in the OH&S Act 2004 at section 27.
While some designers of robotic plant are aware that their creations may only reveal “design faults” down the track, perhaps years after the plant was sold to a manufacturer, few realise that they may be held liable for incidents and injuries that may eventuate 5,10 or 20 years after sale.
Since the goods have been sold to the consumer and the manufacturer may have completely lost contact with the buyer, the problem becomes how the designer can know about the design fault.
Of course, provided the legal consequence are serious enough, the end user will find their way back to the designer in order to mount a criminal prosecution and a civil suit measured hundreds of thousands or even millions of dollars.
To minimise the probability of this, the designer needs to develop a “due diligence” management system that will cause “faults” to be reported back to the designer.
The latter can then alert customers to any design faults that may occur and set in place a call-back type announcement published in an appropriate technical or trade publication. This allows the designer to show due diligence, which may minimise their liability.
Incidents, injuries and compensation claims occurring from collisions between robots plant and workers can only increase.
Designers of such plants need to take steps to protect operators who may be exposed to design faults. Employers need to better analyse the foreseeable hazards of robots on the shop floor in order to better control them and protect themselves and their businesses from potentially huge.
* Ray Schaffer principal consultant with RMH Schaffer and Co. Health, Safety and Environment Consultants, 02 9878 0613. Visit www.environmentdiy.com.au, pose a question and receive an answer at no cost. The author would like to recognise the input of George Issa of NJ Phillips in this article.