One of the underlying issues raised during the initial investigation of the incidents was the response on the Complex to the handling of alarms. There was clear evidence from both the FCCU fire and the MP steam main rupture of operators in the control room experiencing significant "alarm flooding" during the incidents.
In order to specifically investigate this issue further an inspection visit was made by specialists from the HSE’s Human Factors team.
The investigation at Grangemouth took a wide view of alarm handling across the whole Complex. It included discussions with operators and management personnel from the different business streams and reviewed the status of current initiatives at the Complex aimed at improving alarm handling. Many of the issues had already been identified by management at the Complex and much work was currently on-going (but not completed) in preparing an alarm philosophy and in standardising alarm handling for the Complex.
The key findings of the specific investigation into alarm handling at the Complex were:
BP did not have adequate arrangements for alarm handling. Although the management understood the nature and broad extent of the problem, projects were stalled or solutions had not been implemented. Although BP reviewed the HSE Texaco major incident report in 1997-98 which made recommendations on alarm handling, the actions from this review had not been completed;
BP had not allocated sufficient priority or resources to plan, and because of the organisation structure at the Complex, coordination and direction was poor and tracking progress and tackling alarm handling issues was limited;
The alarm management review process had not been carried out to an agreed standard and staff at the Complex had not undergone specific training for tackling alarm handling problems;
The BP Task Force audit report and BP incident investigation team reports identified the alarm handling issues correctly.
BP recognised the problems caused by alarm flooding and work had been commenced at the site to address the problem. A project is now under way to put software in place to manage the prioritisation of alarms properly. The projected timescale of 5 years for completion has been accepted by the Competent Authority.