How often should risk assessments be reviewed?

A common question associated with machine safety is, how often does the machine’s risk assessment need to be reviewed? In the past each state and territory had its own Act, Regulations and Code of Practice, which had slightly different requirements. Some legislation, such as NSW Regulations 2001, states that some risk assessments should be reviewed every 5 years. Another document released by the Queensland government titled “Guide to Safeguarding common Machinery & Plant” indicated that risk management should be reviewed once every 12 months.

The national Work Health and Safety (WHS) legislation was planned to remove these differences by providing a uniform set of Act, Regulations and Code of Practice for all states and territories. Unfortunately this hasn't been adopted by all states this year. The following guidance comes from the Safe Work Australia Code of Practice, “How to Manage Work Health and Safety Risks.”

The process of Risk Management, which includes risk assessment, is an ongoing process that may be triggered when:

• Starting a new business
• Changing work practices, procedures
• Purchasing new equipment
• New information on risks becomes available
• An incident occurs
• Workers, health and safety representative or others raise concerns
• WHS regulation specify the need for specific hazards

My question would be; if none of the above events occur, when should the existing risk controls be reviewed?

Reliability Data for Safety Devices

With the transition from Safety Categories to Performance Levels (PL) or Safety Integrity Levels (SIL) in international machine safety standards, product reliability has become a quantifiable requirement. Both the methods for design to PL and SIL require reliability to be calculated for devices used in the safety control system.

Guidance is given in ISO 13849-1:2006 clause 4.5.2, which states the following hierarchy of obtaining reliability values:

1. Manufacturer’s data
2. Use Annex C and D in the standard
3. Choose 10 years

Most manufacturers are now supplying Mean Time To Dangerous Failure (MTTFd) or B10d for their safety components and the tables supplied in the standard offer good guidance for components where manufacturer data isn’t available. However both these data sources require the components to be used in the correct environmental conditions.

How can designers obtain reliability data for real world conditions where temperature, dust, vibration, direct sunlight, etc may adversely affect the lifetime of their components?

Searching on the internet I can find many references for process safety components and installations. Organisations such as Exida publish handbooks for reliability of different components in varying installations, based on real world results. These would be useful for machine safety but are biased towards process components and conditions. Are there such publications specific for machine safety? This would be a useful resource for designers of machine safety systems in harsh or extreme conditions.


For more information be sure to contact our team of safety specialists.