Process Safety, HAZOP, Risk Assessment and Loss Prevention
Scope
This is a high-yield domain. It tests structured hazard identification, consequence and frequency thinking, and safeguard independence. Expect HAZOP guide words, LOPA, risk matrices, relief and flare basics, flammability, and the difference between hazard and risk.
Core rules
- HAZOP examines deviations from design intent at nodes using guide words: NO or NOT, MORE, LESS, AS WELL AS, PART OF, REVERSE, and OTHER THAN. For each deviation, record causes, consequences, existing safeguards, and recommendations.
- A safeguard must be independent, auditable, and effective. An alarm without operator response time is not a safeguard. A control loop is not an independent protection layer if it is also the cause of the deviation.
- LOPA: mitigated frequency = initiating frequency x product of PFDs of independent protection layers. Compare with risk tolerance; do not add layers until the number looks small.
- Risk = frequency x consequence. ALARP requires demonstration that further risk reduction is grossly disproportionate, not that risk is zero.
- Relief design must quantify credible relief loads from fire, runaway reaction, blocked outlet, or utility failure. Check inlet and outlet line losses and back pressure.
- Flammability limits, flash point, autoignition temperature, and dust explosibility are material properties that must be checked against operating conditions.
Worked example
A feed line to a reactor has the HAZOP deviation MORE flow. Causes include a control valve failing open, an operator opening a bypass, and upstream pressure rising. The consequence is reactor overfill, exothermic runaway, and relief demand. Safeguards are a high-flow trip with PFD 0.1, an independent high-temperature trip with PFD 0.01, and a relief valve with PFD 0.01. Initiating frequency is 0.1/yr. Mitigated frequency = 0.1 x 0.1 x 0.01 x 0.01 = 1e-6/yr. The numbers are illustrative; the method matters. If the trip shares a sensor with the control loop, it is not independent and its PFD cannot be claimed.
Traps
- Running HAZOP as a checklist rather than a systematic deviation study.
- Listing operator training as a safeguard without a defined response and evidence of effectiveness.
- Ignoring common-cause failure, especially shared instrument air or power.
- Confusing a relief valve's set pressure with its capacity. Both must be checked.
Record the design intent and node boundaries. A HAZOP without a defined intent cannot judge deviation.