Scheduling and optimal delivery of water in irrigation networks by combining AquaCrop model and genetic algorithm

Abstract

Introduction

Lack of attention to the proper distribution of water in irrigation networks has caused their usefulness to decrease. In order to an irrigation network to function properly, in addition to reducing the losses of water transmission and distribution in canals, the allocation and distribution of water to agricultural units should also be in accordance with water requirements. On the other hand, because of the weakness of experimental methods in the delivery and distribution of water, most of these networks do not have optimal performance. Considering these problems and due to the limited water resources, it is necessary to use optimization methods for scheduling the delivery and distribution of water in irrigation networks. In recent years, optimization of irrigation scheduling has been developed using crop simulation models such as AquaCrop instead of extracting experimental production functions. The simulator models used in irrigation programs provide the possibility of checking and predicting the behavior of the system in the face of various variables and save costs and make more appropriate decisions. Therefore, linking the optimal scheduling of water allocation and delivery, and its integration with the crop model, is one of the issues that needs investigation and research. This research was done with the aim of optimal distribution and delivery of water in the distribution canals based on the water requirement of agricultural units, which is extracted from the crop model, and using the genetic algorithm method.

Materials and Methods

Delivery discharge to each outlet, the number of outlets grouped in a block and the sequence of water delivery to the outlets are the three main components of water delivery in irrigation networks. Therefore, scheduling the distribution and delivery of water in the network means determination of these three components. In this research, an optimal water delivery model was developed. The proposed approach was applied on the M canal of the Moghan Irrigation Network of Iran. By choosing a certain number of blocks, the outlets located in the distribution canal are distributed inside each block in such a way that all the outlets are placed inside the blocks and no block are left without outlets. In the following, the way of scheduling water delivery to each of the outlet located in each block and the flow rate delivered to them is such that the flow rate of the canal, the time required to complete the irrigation scheduling and also the difference in the volume of water delivered to each outlet are minimized. The limitations of the model are: a) the total flow entering the outlets that collect water at the same time should not exceed the capacity of the canal. b) the total water intake time of the outlets inside each block should not exceed the irrigation interval. In this research, the allocation of water and the cultivated area in the optimal water allocation model, which is linked to the crop model, were used to optimization water delivery.

Results and Discussion

By using the volume of water allocated to each crop and cultivated area, the water demand for each outlet in different time steps was obtained. Following and by determining the water demand of each outlet and also by knowing the physical characteristics of the canal, optimal water delivery modeling was done in the canal for each time step by using the optimal water delivery model and genetic algorithm. Water delivery factors are presented for the 17th time step (peak demand period). The results show that the optimal flow is not exceed the minimum and maximum flow of each off-take. Also, the shorter total duration of the exploitation process than the irrigation interval indicates compliance with the relevant condition in the model. The hydrograph of the inflow into the canal M showed that the maximum flow of the inflow into the canal is less than the capacity of the canal, which indicates compliance with the relevant condition in the model. The hydrograph of the inflow to the canal provides the number of settings of the main off-take of the distribution canal. During the completion of the water delivery program, the main off-take of the distribution canal, which receives water from the main canal A, is adjusted a total of 23 times by the network operator.

Conclusion

In this research, the optimization model of water distribution and delivery in distribution canals was developed using genetic algorithm in MATLAB software. With the aim of optimal allocation and delivery of water at different levels of the irrigation network and by combining the crop model, the results and output of the models were combined with each other so that the results of optimal water scheduling are more appropriate with the real conditions and water requirements of the network. For the canal in the peak demand period, the maximum and minimum canal capacity were calculated to be 2.573 and 0.590 m3/s, respectively, the maximum time to complete the irrigation program was 232 hours and the number of settings of the main off-take was calculated as 23 settings. The obtained results indicate that the developed models are useful for scheduling the optimal allocation and delivery of water and with its help different goals can be optimized simultaneously.