Seasonal agriculture is increasingly becoming incapable of providing sufficient food for the growing population and as a result irrigation is necessary. There have been developments in irrigation using traditional methods. Traditional methods are not effective since farmers do not know how much water should be applied and the time of its application. The aim of this project was to develop a system that detects changes in soil moisture levels, quantify the soil moisture deficit and triggers instant irrigation for a specified period of time. A low-cost sensor data-driven water-use efficient automation system was required to precisely schedule and meter irrigation. The main objective of this project was to design, construct and evaluate an automated drip irrigation control system that would increase water use efficiency in irrigated agriculture. The specific objectives were to create the system AI algorithm in C++ that will control the system, to design a core circuitry to house the algorithm from the environmental conditions and to test and evaluate the system functionality. The system AI algorithm was developed using C++ program. This algorithm was the set of instructions written on a micro controller to be executed in a certain sequence. The algorithm checked the tank level and sensed the soil moisture content and triggered the pump and solenoid valve. The program contained irrigation specific information, irrigation trigger commands and irrigation timing formulae. The C++ program consisted of the equations for irrigation scheduling and for tank refill pump running time. The system electro-ware was constructed. This comprised of capacitors, transistors, resistors and integrated circuits (ICs) which were chosen to check the tank level, sense soil moisture variations, trigger the pump and auto open and close the solenoid valve. The tank level sensor was evaluated by comparing the readings measured manually using a tape measure and those displayed by the kit screen. The pump was tested for auto switching on and off when the tank level was below 20% and above 80% respectively. The C++ program designed was able to detect, quantify the changes in the soil moisture, schedule irrigation and trigger the irrigation solenoid valve and automatically run the pump. The electro-ware system was able to house the program and interact effectively. The tank level sensor values obtained from measuring using the tape measure corresponded with the values read by the system. The pump turned on automatically when the tank level decreased to 200litres (20%) and off when the water level in the overhead tank reached 800litres (80%). The soil moisture sensors were validated by recording the sensor readings and comparing with the values from gravimetric method. A paired t-test was carried out on the readings from the kit and those from gravimetric method and the mean difference was found not to be zero. The system was able to apply a specified amount of water each time the soil moisture deficit occurred for a corresponding period of time. The system was found to be efficient in scheduling irrigation.

show more