Novel application of continuously variable transmission system using composite recurrent Laguerre orthogonal...

Area of Science:

• Control Engineering
• Artificial Intelligence
• Mechanical Systems

Background:

• V-belt continuously variable transmission (CVT) systems driven by permanent magnet (PM) synchronous motors exhibit complex nonlinear and time-varying characteristics.
• Designing linear controllers for these systems is challenging and time-consuming due to inherent system uncertainties.

Purpose of the Study:

• To develop an advanced control system capable of online learning to address the nonlinear and time-varying nature of PM synchronous motor-driven V-belt CVT systems.
• To improve control performance in the presence of lumped nonlinear load disturbances.

Main Methods:

• A composite recurrent Laguerre orthogonal polynomials neural network (NN) control system was developed.
• The system incorporates an inspector control, a recurrent Laguerre orthogonal polynomials NN control with an adaptation law derived from Lyapunov stability theorem, and a recouped control with an estimation law.
• Modified particle swarm optimization (PSO) was employed to determine optimal learning rates for NN parameter adaptation, enhancing convergence.

Main Results:

• Experimental results demonstrated the effectiveness of the proposed control scheme.
• The developed control system showed significant improvements in managing the nonlinear and time-varying dynamics of the V-belt CVT system.
• The online learning capability allowed the system to adapt to lumped nonlinear load disturbances.

Conclusions:

• The proposed composite recurrent Laguerre orthogonal polynomials modified PSO-NN control system offers a robust solution for controlling PM synchronous motor-driven V-belt CVT systems.
• The integration of online learning and adaptive control strategies significantly enhances system stability and performance.
• This approach overcomes the limitations of traditional linear control design for complex dynamic systems.