Design and modeling of ann based automatic generation control and automatic voltage regulator for two integrated hydro power plants in Ethiopia

Automatic Generation Control (AGC) and Automatic Voltage Regulator (AVR) systems are crucial components in the operation and control of electric power to offer adequate and reliable electric power of high quality. An electric power system is constantly subjected to numerous forms of disruption, resulting in a loss of stability. Variations in active power and reactive power can lead to unpredictable operation, data loss, system crashes, computer equipment damage, lamp flashing, etc. To lessen frequency and voltage variations as well as the system's oscillation characteristics, the design and modeling of AGC and AVR systems are discussed in this paper. It has been done by designing and modeling each component of AGC and AVR systems, including Propertional-Integral-Derivative (PID) and Artificial Neural Network (ANN) controllers by using a mathematical modelling system on Tanabeless and Tis Abay II integrated hydropower plants. For instance, Tanabe's data analysis reveals that there was an outage on October 16, 2013 EC, between 4:00 and 9:00 local time. This was brought on by the system's high-frequency variation. According to Ethiopian standards, the generation system's electric power frequency must be maintained within the range of [49.5 - 50.5] Hz, or 1 % fluctuation, during normal steady-state operation and 48 to 52 Hz during any disruption scenario. MATLAB software was used to simulate the designed model. By applying a PID controller on the designed model of Tanabeless system, 0.12 sec rise time, 21 % overshoot, and 2.2 sec settling time have been obtained on the change in voltage response, while these values have been improved to 0.05 sec rise time, 9 % overshoot, and 0.5 sec settling time, respectively, using an ANN controller. On the other hand, 2.13 sec rise time, 3.1 % overshoot, and 21 sec settling time have been obtained on the frequency response of Tanabeless system using PID. However, the ANN controller has improved these characteristics to 1.02 sec rise time, 2.2 % overshoot, and 13 sec settling time. Similarly, the change in voltage response and the change in frequency response oscillation characteristics of Tis Abay II generation plant have been improved by using an ANN controller rather than PID. The change in tie-line power flow from Tanabeless to Tis Abay II has also been improved by 1.3 sec rise time, 1.5 % overshoot, and 17 sec settling time. It has been observed that the designed model of the system using an intelligent controller (ANN) has better technical characteristics (overshoot, rise time, and settling time) than a classical controller (PID).