Rekayasa Mesin (Jun 2025)

DESIGN BACK PROPAGATION NEURAL NETWORK (BPNN) – PID OF ACTIVE AIR SUSPENSION BASED ON HALF CAR MODEL AT PLUG-IN HYBRID ELECTRIC VEHICLE (PHEV)

  • Bambang Sampurno,
  • Kyla Anisa Windarta,
  • Mohammad Berel Toriki,
  • Liza Rusdiyana,
  • Dika Andini Suryandani

DOI
https://doi.org/10.21776/jrm.v16i1.1919
Journal volume & issue
Vol. 16, no. 1
pp. 455 – 468

Abstract

Read online

A Plug-In Hybrid Electric Vehicle (PHEV) is a car with a combination of an electric motor and an internal combustion engine (ICE). The implementation of active air suspension in this research uses a half car model. Mathematical modeling is used to obtain system responses such as body displacement, body acceleration, rear wheel displacement, and rear wheel acceleration using MATLAB software. There are 3 test modes, namely passive suspension, active suspension, and implementation using a neural network-based control system. Based on these 3 test modes in 3 conditions, the use of passive suspension for body displacement produces a maximum overshoot of 133% and a settling time of 2.15 seconds. Meanwhile, the active suspension produces 43.33% and a settling time of 0.7 seconds. When using a neural network, it produces 50% and a settling time of 2.14 seconds. Some while, the use of passive suspesion foor body acceleration produces a maximum overshoot of 133%, arms of 124,2, and a settling time of 2.15 seconds. Meanwhile, the active suspension produces maximum overshoot of 43.33% , arms of 2.92, and a settling time of 0.7 seconds. When using a neural network, it produces maximum overshoot of 50%, arms of 2.92 and a settling time of 2.14 seconds.

Keywords