Most hazards analyses focus on single, discrete systems. For example, a refinery hazards analysis team may carry out its review on just the catalytic cracking unit; a pipeline company may analyze just the marine loading operations; or an offshore team may analyze just one platform in a larger complex. Yet these sub-systems are part of larger systems; which means that hazards can be transferred to or from the other units across the interfaces.
One large oil production facility, for example, had both onshore and offshore operations. An operator was carrying out a routine pigging operation on a line that came from an offshore platform to the onshore gas processing plant. He inadvertently misaligned the valves around the pig trap and caused a high pressure surge to flow back along the line coming from offshore. This mishap had no significant effect on the onshore operations themselves, but the pressure surge caused the offshore platform to shut down, which triggered a chain reaction that caused many other offshore platforms in the complex to shut down in sequence. In the end, many millions of dollars of production were lost, and the company was lucky not to have had a safety or environmental incident.
Another example of interface operations concerns truck operations. Many process facilities use trucks from third party companies to bring in chemicals and to export products and waste streams. It is generally a good idea to invite a representative of the trucking company to the pertinent process hazards analysis. That way each party can assure itself that the chances of a mishap are small. The process facility, for example, can evaluate the procedures to make sure that delivered chemicals are what they should be; the trucking company representative can check for the possibility of reverse flow of process chemicals on to their truck.
An Interface Hazards Analysis (IHA) can usually be structured into three areas:
- Process fluids (wrong hazards analyses / reverse flow / wrong composition);
- Instrument signals;
- People interfaces.
No established methodology exists for analyzing system connectivity. However such a system can be viewed as being a collection of black boxes where each black box represents an operating unit, each of which has been thoroughly analyzed individually. These black boxes are like nodes in a HAZOP.
The sketch shows a system consisting of four operating units, each of which can be connected to each of the others in some manner, except that there is no link between Block 2 and Block 4. (All the arrows are two-way meaning that connectivity problems can flow in either direction.) There are ten interfaces. If the analysis uses the three categories shown above that means that there could be as many as 30 interface discussions.
. . . . .