Townley, L.R., Barr, A.D., Johnston, C.D., and Salama, R.B. (1990), Hydrologic and process modelling of solar salt production at Useless Inlet and Useless Loop, W.A., CSIRO Division of Water Resources, Consultancy Report, 125pp. [Confidential]

This is a report on the production of solar salt by Shark Bay Salt Joint Venture at Useless Inlet and Useless Loop, Western Australia. The study was carried out by the CSIRO Division of Water Resources between August 1989 and May 1990.

A major outcome of the study has been the development of a computer simulation model capable of predicting the movement of salt through SBSJV's ponds at Shark Bay. This model can simulate long sequences of historical data and simulate the behaviour of modified configurations of ponds. The model does not allow the development of control strategies to allow target concentrations to be reached in individual ponds, other than by trial and error.

Another outcome has been the development of a calibration aid, using Lotus 1-2-3 Release 3, which uses historical data to infer likely long-term average flows through the channels and pipes connecting ponds in Useless Inlet. This method indicates quite clearly that significant leakage from P1A and P1B to P3B acts to short-circuit the intended flow path through P2. This explains the significant jump in density from P1B to P2 and the levelling off in density beyond P2.

The study reached the following conclusions, of which some are confirmation of beliefs already held by Shark Bay Salt Joint Venture.

1.Potential production of salt is reduced by leakage in Useless Loop. In Production Year 1988, we believe that about 1,800,000 tonnes of salts, of which 85% is NaCl, passed into Useless Loop. Since the annual production of NaCl is only about 600,000 to 700,000 tonnes, there is significant loss of salt by leakage in the Loop.

Leakage occurs both in the Inlet and the Loop, but by different mechanisms. In the Inlet, the pond bottom elevations are below sea level, and we believe that there is a density-driven exchange of brine in the ponds with older seawater below. This results in a net loss of salt, but no net loss of water. In the Loop, pond bottoms are above sea level and leakage results in loss of both water and salt. Since brine densities increase progressively through the Loop, loss of equal volumes of brine result in progressively larger losses of salt near the feed channel. The largest leakages of salt appear to be in Ponds Cl and C2 which are unlined.

2.Production is affected by the limited capacity of the flume connecting Useless Inlet and Useless Loop. There are several options by which the production of salt could be increased, and these are currently being investigated by Clough Engineering for SBSJV. These include: i)developing Useless Inlet to increase the concentration of brine entering the flume; and (ii) feeding brine to Useless Loop for a larger proportion of the year and creating new ponds or storages adjacent to existing ponds at the Loop. The model produced as part of this project is the primary tool being used for investigation of these options.

3.The accuracy of any model of historical behaviour of the ponds at Shark Bay is severely limited by: (i) uncertainty in the level-depth-area-volume relationships for all ponds; (ii) the almost complete lack of any measurements of flow rates between ponds; (iii) the fact that connections between ponds do not act as hydraulic controls, in the sense that there is no unique relationship between levels and flows; and (iv) inaccuracies in records of gate openings and closings (in the Inlet, but especially in the Loop where each gate change creates a surge which rapidly affects levels in ponds of small surface area). All of these difficulties have made calibration of the model an extraordinarily difficult task.

Further surveying of the ponds has been undertaken by SBSJV during this study. Acoustic measurements have also been made to verify that the daily flume flow is about 650 ha-cm/day. These additional data have helped us to calibrate a model of historical behaviour, but it is recommended that SBSJV should take steps to progressively create hydraulic controls which will allow series flow throughout all ponds and facilitate flow measurements.

4.The development of any new ponds in Useless Inlet will ultimately increase the density of brine entering the flume. Leakage out of the bottom of ponds in Useless Inlet is not large enough relative to evaporation rates for additional ponds to have a negative effect. Increasing the duration of flow in the flume will require significant expansion of pond areas in Useless Loop in order to allow sufficient evaporation to produce pickle.

5.Optimal control of a solar salt facility requires good knowledge of rates of evaporation. Evaporation rates depend on the difference between vapour pressures in the atmosphere and at the surface of the ponds. These depend on relative humidity of the atmosphere and chemical activity of the brine, but whereas we have been able to determine a relationship between activity and density, records of relative humidity are not available. We have therefore calculated evaporation by scaling daily pan evaporation by activity and by a pan-to-lake coefficient. By measuring wet and dry bulb temperatures, water temperature in a number of ponds, and net incoming radiation, two independent methods could be used to calculate evaporation with greater precision than is now possible. Measuring temperatures alone would be a simpler alternative. The primary reason for seeking to enhance the accuracy of evaporation estimates is to assist in efforts to calculate leakage by water and salt balance methods.

6.Leakage out of the bottom of ponds in Useless Inlet probably occurs by a large-scale fingering mechanism which can not easily be quantified. There is some evidence that a salinity anomaly observed near the levee between P1C and P2C (our terminology) is the result of upward leakage of older seawater. The upward leakage is caused by a nearby downward leakage of brines. There is good reason to believe that the rate of loss of salt by fingering should progressively decrease with time, regardless of any sealing of the bottom of the ponds. A programme of research on fingering processes is now being undertaken by several CSIRO scientists in different locations.

7.Leakage through levees in Useless Inlet has a significant effect on the efficiency of the ponds. It is believed that leakage in the past from Ponds lA and 1B to Pond 3B has significantly reduced the potential maximum density in the flume. SBSJV have recently taken steps to reduce leakage from P1A by constricting a new levee adjacent to the old one.

8.An increase in the concentration of brines in ponds P3A and P3B could increase the sealing of the bottoms of these ponds by gypsum deposition. This would be an advantageous side effect of aiming to increase salt flow in the flume by this method. A disadvantage would be the likelihood of increased gypsum deposition in the flume, requiring more regular flume maintenance.

9.Leakage from Useless Loop now appears to be the limiting factor which reduces salt production from SBSJV's ponds. We recommend that careful experiments should be carried out to determine relative leakage rates from different ponds. These experiments would be most effective if carried out in non-rainy periods to remove the confounding effect of rainfall. Evaporation and leakage can be separated, especially if water levels are allowed to drop by 25 or 50%, if water chemistry (density, conductivity, composition) is measured throughout the experiments.

10.Use of the simulation model SALTPROD to simulate new pond configurations or changed operating procedures requires the specification of gate opening and gate closing data for all ponds in Useless Loop. This is a very time consuming procedure for each new scenario. The alternative of simulating operating rules would facilitate the evaluation of alternatives, but the resulting sequences of gate changes would need to be modified to obtain an operating procedure which could be implemented in practice.

It is very difficult with SALTPROD to simulate a series flow situation in the Loop, because the typical connections (culverts with weir controls) do not allow the flow between ponds to be regulated by downstream levels. Our experience with SALTPROD leads us to believe that this is a fundamental design problem which will make series flow difficult to achieve in the field.