APLIKASI NEURAL NETWORK PADA SYSTEM CONTROL TURBIN MIKRO HIDRO

Abstract

The flow of water that flows in the river at an altitude of a certain level can be used to rotate turbines
that can produce rotary energy. When this rotary we used to rotate a generator it will generate
electrical energy. Micro-hydro power plant generating electricity in a small capacity that could be
built in location depends on its environmental conditions and capable of generating electricity in a
limited capacity. The problem arises is that the frequency resulting from micro-hydro plants can not
be stable at around 50 Hz, this will be very influential with the voltage drops due to the addition of
the load. To stabilize the turbin at a certain round, so far only use the height of water level inside the
tank as the input, so that the turbine is expected to generate a fixed rotation. This standard method
will extremely unstable when there is an increase in load. To overcome this problem, the volume of
water entering the turbine is adjusted by using the Application Control Neural Network (ANN) with
feedback from the output frequency of the generator. ANN will be able to produce output according
to the learning process based on the weight of the input neurons of the frequency, rotation, and height level to be able to control the microcontroller to drive the water turbine governor, so that can
keep pace with the changing load on a stable frequency at 50 Hz and a voltage of 220 VA.