When an Incident Investigation Fails
What happens when an investigation reaches the wrong conclusion?
In process industries, incident investigations are a cornerstone of safety management, forming part of frameworks such as the US Center for Chemical Process Safety Risk-Based Process Safety model and the Energy Institute Process Safety Framework. But even well-structured systems can fail, particularly when culture and organisational behaviour are not properly examined.
The following real-world example illustrates how an inadequate investigation resulted in a repeat incident years later.
The Scenario: Chlorine Offloading
A site operated a manual chlorine tanker offloading system with detailed procedures, including pressure checks, purging and torque requirements for flanged connections. A new gasket was required for every operation.
Previous Incident
Five years earlier, a minor chlorine leak occurred. The investigation concluded that the operator had failed to correctly tighten flange bolts. The procedure was deemed adequate, and the operator was disciplined. Refresher training was provided.
Latest Incident
A second chlorine leak occurred during offloading.
This time, the operator followed the procedure exactly:
All required operations and checks were completed
A new gasket was installed
The correct torque was applied
Yet the system still leaked.
What the Investigation Found
A deeper investigation revealed two critical issues:
1. A hidden workaround.
Operators had been routinely applying extra torque beyond the procedure.
Why? Because the previous incident had resulted in disciplinary action and no one wanted to take that risk again.
This informal practice was:
Never reported
Never documented
Never incorporated into training
The new operator, unaware of this, followed the written procedure and experienced the failure.
2. A silent technical change
Five years earlier, the gasket supplier had changed. The original supplier stopped producing the gasket and the purchasing department had selected an alternative one, to the same specification, from a different supplier.
Purchasing department has informed the plant engineer of the proposed change and despite chasing them, they received no response. Since the usage rate of the gasket was high, the new supplier was selected.
Although within specification, the new gasket had slightly different compressibility. Under operating conditions, the specified torque was no longer sufficient.
This change:
Was not formally agreed with engineering, despite the efforts of the purchasing department.
Fell outside the site’s formal Management of Change (MoC) process
Was never assessed for operational impact
The Real Root Causes
The original investigation had focused on operator error.
The real issues were:
A failure of Management of Change, particularly across departments
A culture that discouraged speaking up
The emergence of undocumented workarounds to manage risk
The operator blamed five years earlier was later formally exonerated.
Lessons Learned
This case highlights a common failure mode in incident investigation:
Stopping at the immediate cause
Failing to explore systemic and cultural factors
Over-reliance on procedures as proof of control
When this happens, organisations don’t just fail to learn; they actively embed risk into their systems.
What Changed
Following the second incident:
Informal workarounds were actively identified and reviewed
Disciplinary responses to incidents ceased
Reporting increased significantly
Maintenance workload rose, but so did risk visibility
The management of change system was revised to improve inter-departmental involvement
And Finally
A strong safety system is not just built on procedures and compliance.
It depends on a speak-up culture, where people feel safe to share how work is really done; not just how it is written.
Without that, even the best-designed systems can fail.
Effective investigations go beyond identifying “what happened”; they uncover why the system allowed it to happen. That’s where real risk reduction lies.
Comments