South Hoy Water Treatment Works

Water Quality Incident Report – November 2009

DWQR’s assessment of this incident is that at 07:19 hrs on 30 October 2009 the operator received a low low final water chlorine alarm. The operator attended the site and raised the chlorine dose rate to the treated water. At 10:17 hrs the operator received a high treated water alarm but did not respond as he had just raised chlorine levels. At 08:08 hrs on 31^st October, and again at 18:57 hrs, the Team Leader received high treated water chlorine alarms but did not respond on either occasion because the operator had raised the chlorine dose the previous day.

At 23:48 hrs on 31^st October the Team Leader received a low low treated water chlorine alarm but did not respond as he thought it was because of the water treatment works shutting down due to the clear water tank being full. When this occurs it generates a temporary low treated water chlorine alarm. However, it later proved to be due to the chlorine dosing pump partially blocking with manganese dioxide deposit.

At 10:00 hrs 1^st November the relief operator attended the treatment works but failed to react to the low chlorine residual.

From the chlorine residual records, at about 09:40 hrs on 2^nd November the chlorine dosing pump completely blocked but the pump action continued to run. At 12:00 hrs on 2^nd November the relief operative visited the works. He noted that chlorine levels were low but did not intervene.

Sometime between 15:30 hrs and 16:00 hrs on 2^nd November the dosing pump blockage cleared by itself and normal chlorine dosing re-started

At 16:00 hrs on 2^nd November the Team Leader noticed that the final water chlorine levels were low and sent the operative to the works. The Public Health team were contacted at this point and discussions held with the CPHM. It was agreed that normal disinfection would be restored, manual dosing of chlorine to the final water tank would be carried out, and flushing of the distribution would be undertaken. The operative confirmed that the chlorine dosing pump was now operating as normal then he dosed the final water tank with a shock dose of sodium hypochlorite. Flushing and sampling of the distribution took place on the evening of 2^nd November.

From the telemetry data it seems that, during the incident, chlorine concentrations in the treated water fell to zero for a period of around 20 hours, while that of the final water fell to 0.05mg/l. Seven bacteriological samples of potable water, taken between the 2^nd and 4^th November, were all satisfactory

DWQR notes that the incident was further compounded by the final water telemetry alarm, which had already dialled out. This is coupled with the fact that the final water chlorine residual never recovered to a sufficiently high value to take it above the low low trigger point and was therefore still live, so no new alarms were generated.

DWQR is disappointed with the time it took Scottish Water to discover and deal with the events that led to this incident. The failure to act on repeated low chlorine alarms and the operative’s lack of intervention once he was on-site are unacceptable.

Scottish Water’s actions in response to this incident are noted, the use of distilled water for chlorine solution make-up to prevent the build up of manganese dioxide in the pump, the reconfiguration of the telemetry to stop it giving an alarm when the plant shuts down, and a range of staff training exercises on the management and operation of the works, are essential if a repeat of this incident is to be avoided.

DWQR also considers it critical that the lessons learned from this incident are fully disseminated to the whole of Scottish Water as a delay in, and/or inappropriate response to, alarms can result in minor events becoming major incidents when a fault occurs in part of the water supply process.