Improving the Coefficients of Proportional-Integral Controller Based On System Identification Process on Doosti Irrigation Network

The use of automatic control system for managing the conveyance and distribution of water in surface irrigation systems, as a means of improving the management and performance of these systems is essential. Automating networks using the programmable controller provide the implementation of a number of control ways according to different water delivery scenarios. Also the implementation of these systems using sensors and recording water levels by hydraulic devices can supply an accurate data collection of water distribution networks in the long term and can influence the decisions to manage the network. Control systems in irrigation canals include two parts: calculation for adjustment the structures (algorithms or control system software) and application the calculated settings on structures (hardware of control systems). The success of these parts is depending on the ability of control algorithm to determine the control parameters.

Doosti irrigation network in the plains of Sarakhs is located in the northeast of Iran. In the end of the main canal, two canals has branched which named EPC and WPC2. This study was performed on the EPC canal with the discharge and length of 4.43 m3/s and 18.7 km respectively. There are 15 duck-bill check structures along the EPC. Also 14 intake structures and secondary canals are responsible for water distribution between local water users. In this study, the performance of system identification process, SI, in estimation of coefficients of proportional-integral controller and improvement of adjusting the control algorithm was compared with trial and error process. This proportional-integral control algorithm was designed for EPC canal. The efficiency of algorithm was evaluated using the simulation results of several different choices of operating systems with SOBEK hydrodynamic model and computing evaluation indices of control systems. This model can simulate all kinds of structures in water conveyance and distribution networks. In addition to the integrity of this software, two proprietary approach of the model are ability to design the control algorithms for smart irrigation systems and the ability to integrate each of its modules with each other and increase compliance with real systems. Evaluation of designed control algorithm was done using the pattern of the one-month operation, with extreme daily variation to cover the most intense perturbations in canal. To evaluate the performance of the designed control algorithm for investigated scenarios in the canal, the performance evaluation indices of control systems were used.

The results of SI process show that the delay time for the longest reach is greater than the other. Also the values of delay time show that the smaller discharges produce more delay time. By comparing the storage area for applied discharges for SI process, it was concluded that the amount of storage area for low flows is higher. Because for low flows with fixed target level in upstream the check structures, the length of backwater curve is more and consequently the amount of storage area would be higher. By comparing the amount of resonance frequency and maximum resonance for the potential of 10%, 50% and 80% of the design flow for various scenarios, it is indicated that in low flows the resonant behavior has significant effects on canal reach. The results show that the adjusted control algorithm using identification process, rather than trial and error, has considerable accuracy and high potential to flow control and to damp the perturbation of hydraulic disturbances. Also this algorithm provides the on-demand distribution and promotes the efficiency of water control and distribution systems.

Although using automated control systems faced with the problem of high initial investment but as an effective tool, it improves the performance of water delivery and provides the optimal usage of irrigation network capacity. For this reason, it is recommended to custodians of irrigation networks to design and employment of these controllers in irrigation networks with more attention. In present control system IP controller was used. This controller requires the calculation of adjustment coefficients of control algorithm. For comparison between the calculation of these coefficients by trial and error and the identification system, the IP controller for EPC canal of Doosti Irrigation was designed. Then its performance was studied based on the method of canal operation and standard indices. The results indicated that the tuned algorithm control using SI has considerable accuracy and potential to control the flow and to damp the perturbations concluding from structural and hydraulic disturbances. Also this algorithm provides demand oriented distribution and performance promotion of water distribution system. Since in this study a relatively long canal includes multiple reaches and structures such as inverted siphon was studied, the capability of PI and SI process is also confirmed in these situations.