Optimum Length of lining to Reduce Losses in Watercourses by Using Advanced Non Linear Modelling

Abstract

Life on earth is dependent on water. It is very valuable natural resource. Distribution of water on earth is not uniform in its both forms i.e. surface water and ground water. The quality of groundwater is varying from fresh to brackish in many areas. So it may not be fit for drinking and irrigation purposes. Surface water has fresh quality water and mostly used for irrigation purposes by diverting water from rivers and stream into canals. As this precious and scare resource moves into the irrigation system, certain part of water is lost. There is shortage of canal water supply in the dry spell and mostly during the winter season due to less availability of water and comparatively significant losses in the system. The losses in the watercourses is much more than those in the main canal and distributaries. So farmer’s face acute shortage of irrigation water issues. In addition to this ground water pumping is also increasing that is decreasing ground water table. The conveyance losses in the watercourses can be minimised by applying some lining techniques. To overcome this problem, there is a need to work out optimum length of lining that ensures maximum water saving. The estimation of water conveyance losses in watercourses is very important. A detail study has been carried out in the work to calculate conveyance losses using operational inflow and outflow approach. The losses from both lined and unlined watercourses of a similar geographical area have been calculated and used to compute the percentage saving of water. In this research work too from the field measurements, huge losses in unlined watercourses and a less losses in the lined watercourses have been observed. In this research work, the lining percentage is also authenticated numerically evaluated and authenticated. The determination of this struggle is to safeguard both water and economy at the same stint. A total of 32 ANN models of actual and predicted water saving against the percentage length of lining are developed and analysed against the 8 polynomial regression forecasting method. The performance of TLBP and BFGS models is studied in terms of variance, R Square, Root mean square error. As anticipated, TLBP performed better as compare to BFGS for respective length of data. TLBP models have better values of random and systematic error than BFGS. It also shows better R square values in most of the cases. The percentage of water saving against increase in percentage lining were modelled using polynomial regression and optimum lining length for unlined water courses. The optimum percentage length of lining has been evaluated as 50% that ensures maximum economic benefits and 80% saving of water.