Control rules for Tayside's linked system of reservoirs supplying two demand zones

Backwater and Lintrathen reservoirs are major sources of raw water which together supply Tayside with 102 Ml/d of potable water, following treatment. There are two large supply zones, the City of Dundee and parts of Angus; Dundee is normally supplied from Lintrathen reservoir via the Clatto treatment works and the remainder is normally supplied from Backwater reservoir via the Lintrathen treatment works. Exceptionally the raw water sources can be interchanged, at the cost of some extra pumping and loss of some hydropower generation, or either reservoir can be used on its own to supply both demand zones.

In order to comprehend the action of this complex system, WRc created a computer model and set up a hydrological basis for simulating the system's operation. Sixteen years of monthly inflows, compensation releases, spills and draw-offs were simulated, based on the period 1968–1983. The reliable yield was re-assessed, revealing the nearness of present demands to the system's ultimate capabilities.

Control rules were tried, based on setting a limiting contents level in each reservoir, below which draw-off is curtailed and taken over by the other reservoir. A rationing rule was also postulated, to cut back on draw-off rates in a balanced fashion, when both reservoirs fall below critical levels.

The simulation program has been set up on a PC for day-to-day use by the Tayside Regional Council Water Services Department, who have made progressive improvements to the control rules and have had experience of using them in severe drought situations during the autumns of 1989 and 1990.

To make the system more readily used as an operational tool, the PC package was configured to read dates from and output results to the standard spreadsheet used by the Water Department, for ease of display and assimilation.

Extensive refurbishment of the flow measuring stations was undertaken, along with automation of data gathering, so that frequent checks could be made on the model to allow for its use for predictive analysis of possible operating plans.

• INTRODUCTION

• OPERATIONAL CONSIDERATIONS

• CONCLUSIONS AND RECOMMENDATIONS