Investigations

This section contains summaries of reports on drinking water
quality incidents and investigations.

Report on drinking water quality incident at:
Herricks Water Treatment Works in July 2006

Scottish Water Executive Summary

On Wednesday 5th July 2006 at approximately 1630hrs, failure of the duty polyaluminium chloride (PAC) dosing pump at Herricks WTW resulted in the loss of the coagulation process. Although an alarm was generated onsite at the time of the pump failure, the alarm wasn't received by the Scottish Water telemetry server until 2058hrs. This delay was due to a communications fault between the site remote telemetry unit (RTU) and the telemetry server, possibly associated with a call divert system in place at the time which has since been changed. However, once the alarm had been received by the Operations Management Centre, the standby Operator for the site was notified almost immediately resulting in the coagulation process being restored by 2130hrs. This required manual intervention by the Operator due to the lack of an auto-changeover facility for the PAC dosing pumps.

Although checks undertaken on site at the time together with the results of samples taken in response to the incident showed that the final water quality remained satisfactory, the full impact of the coagulation failure on the individual process units is unclear due to all but one of the turbidity meters malfunctioning. These instruments are currently being replaced. The one available meter at the time of the incident, relating to Filter No.2, highlighted a short exceedence of the Emergency Action Level, the alarm for which had cleared by the time the outstation connected to the telemetry server.

DWQR Assessment on the incident

This incident was initiated by the failure of a pump to supply polyaluminium chloride (PAC) for the coagulation process. Scottish Water's incident report together with additional information gained through correspondence and an on-site investigation identified two separate issues that contributed to the failure and the lack of appropriate alarms about the failure.

The first issue was the lack of appropriate turbidity meter readings/alarms and the delay in the telemetry system informing the Operations Management Centre's Flight Desk of the failure of the pump. Subsequent investigations revealed that of the three turbidity meters that should be operating on the three filters, two were out of action. From December 2005 it was acknowledged that there was no beneficial use regarding the turbidity meter monitoring filter No 3 and there were continuing questions over the reliability of the turbidity meter monitoring Filter No 1. There is a record of maintenance being requested on 9 December 2005 for the turbidity meters which resulted in a decision to replace all the turbidity meters being taken in February 2006. By 20 March 2006 a scope of work for the replacement was agreed and, following various revisions to the quotations and scope of work, the instruments were received on 14 June 2006 by the contractor employed to fit them. The contractor started work to fit the instruments on 26 June 2006 but that day a serious accident involving the contractor resulted in the work being put on hold until Scottish Water and the contractor had completed their investigations. The contractor returned to the site on 2 August 2006 to complete the installation of the turbidity meters. In summary, from February 2006 Scottish Water relied on a single turbidity meter (monitoring filter No 2) and spot checks to control turbidity. When the dosing pump system failed on 5 July no alarm from the single turbidity meter was sent to Scottish Water.Throughout the build-up to the incident the issue of the turbidity meter replacement and associated improvements were discussed and agreed between the operator and the team leader. During DWQR's investigation each of the decisions reached regarding the replacement of the turbidity meters and the operation of the plant were considered to be appropriate given the information available to the staff at the time. The significant delay in the purchasing of the replacement meters and the accident on-site which further delayed the installation of the meters could not have been foreseen by staff when the critical decisions were taken. While not ideal, the method of operation of the plant using the single turbidity meter and spot checks was appropriate given the circumstances and had the subsequent failure of the telemetry system not occurred, then the stop gap solution put in place would have been successful.

This leads to the second issue - the failure of the telemetry system to inform the world outside of the treatment works that a failure on the PAC dosing pump had occurred. The remaining turbidity meter (on filter No 2) detected a breach of the Emergency Action Limit and sent a signal to the works monitoring system.

In turn, it is believed that a signal was then sent to the telemetry system to be sent to the OMC but the configuration of the telemetry system was such that the "Call Divert" system did not operate as expected. During the on-site investigation by DWQR, the issue of the "Call Divert" was discussed and it transpired that although the telemetry system at Herricks had been updated, there were around 100 sites that used a similar telemetry configuration.

At this meeting DWQR requested a list of all affected sites and an action plan for the rectification of the issues highlighted by the incident at Herricks. This information has been slow in forthcoming and DWQR is very concerned over Scottish Water's delay in providing this information.

DWQR understands that Scottish Water has now instigated a wholesale review of its telemetry systems in the light of this and other failures. The outcome of that review will be subject to close scrutiny by DWQR.