A study by George Mason University demonstrated that automation is effective at reducing manual labor, but also introduces new risk.
“That isn’t to deny that warehouse robotics benefits workers,” said Brad Greenwood, Professor at the Costello College of Business at George Mason University. “But we need to think more carefully about how to use them, and what that means for the humans they work with.”
In modern industrial environments, traditional safety risk assessment methods often miss automation hazards. Structured methodologies must account for both conventional machine risks and the new hazard patterns that automation introduces into the industrial environment.
Beyond Hazard Checklists
Machine safety requires going beyond hazard checklists in risk assessment. While checklists can be valuable in operation, regular risk assessment is key to making sure these checklists are updated.
Generic assessments can fail because they treat machines as static systems. Equipment may operate across multiple modes with different hazard profiles for production, setup, maintenance, and fault recovery. Risk patterns also shift depending on whether workers share workspace with robots or maintain strict separation.
Checklists are often developed during installation but fail to account for changes in the environment, usage, or operator. A safety light curtain that functions perfectly during commissioning might develop blind spots from contamination or alignment shifts over months of operation.
You need machine-specific, task-based analysis that evaluates normal operation misuse scenarios, changes in energy state, and edge cases that can create hazards.
ISO 12100: Three-Step Process
The ISO 12100 standard includes a three-step process:
- Hazard identification
- Risk estimation
- Risk evaluation
Hazard Identification
Hazard identification must span mechanical, electrical, thermal, noise, and ergonomic categories. If you miss any categories during your risk assessment, it can create blind spots that remain after implementation.
Risk Estimation and Evaluation
Risk estimation combines severity and probability.
This includes estimating whether harm is reversible or irreversible, whether exposure affects single or multiple persons, and the frequency and duration of exposure. For example, a pinch point that workers pass twice per shift presents a different risk from one that might occur dozens of times per shift.
One thing to note here is that human factors, including stress, fatigue, and production demands, significantly influence probability but are often excluded from many assessments.
Documentation Requirements
A machine safety risk assessment requires documentation that establishes traceability from each identified hazard to the safeguards you put in place, and you have to be careful about assumptions. A risk assessment that assumes your workers follow lockout tagout procedures perfectly will fail if that assumption doesn’t match reality.
Mitigation Hierarchy Application
Machine safety starts with the engineering and design of systems as part of a tiered approach.
Inherently Safe Design
Inherently safe design measures eliminate hazards at the source, for example, designing systems that:
- Limit force and speed
- Eliminate pinch points
- Minimize stored energy
The right design is the most effective way to reduce risk, yet this is sometimes overlooked during retrofits, expansion, or system changes.
Engineering Controls
Engineering controls, including guards, presence-sensing devices, and safety-rated control systems, form the second tier. Fixed guards work well where access is rarely needed, while interlocked guards are best when frequent access is needed. Presence-sensing device applications have limitations in harsh environments or where nuisance trips would halt production excessively.
Administrative Controls
Administrative controls and personal protective equipment serve as last resort measures when engineering solutions are not enough. They’re appropriate for residual risks after implementing higher-tier controls, but not as compensation for poor design, as training effectiveness has limitations. Workers trained on procedures today may develop workarounds next month when production demands increase.
Testing and Validation
Validation is critical to ensure that these safeguards work in operations, and that includes testing at various states of use. For example, testing a safety interlock only with fresh batteries misses how it performs after months of use.
Emerging Challenges in Automated Systems
In automated systems or systems using advanced technology, a risk assessment has to account for more possibilities.
Cobots create new human-machine interactions and must be tested and validated. Machine learning and AI integration raise new safety risk assessment questions as well; establishing bounded operating envelopes becomes critical when systems can modify their own behavior. Cybersecurity also emerges as a safety issue when network-connected systems could be compromised.
From design to integration to operations, machine safety must be foundational. Contact the machine safety experts at Pacific Blue Engineering for a free consultation on your next project.
