A Cascade BP Neural Network Tuned PID Controller for a High-Voltage Cable-Stripping Robot.

Area of Science:

• Electrical Engineering
• Robotics
• Control Systems

Background:

• Manual cable stripping in 10 kV power grids poses risks to workers.
• Existing automated grid maintenance solutions are often multi-functional and bulky, not specialized for cable stripping.
• There is a need for specialized, safe, and efficient automated solutions for cable stripping in power distribution networks.

Purpose of the Study:

• To introduce a novel live working cable-stripping robot designed for 10 kV overhead power lines.
• To develop and implement an advanced control system for the robot, addressing nonlinear system dynamics.
• To validate the robot's structural feasibility and the controller's performance in a practical cable-stripping scenario.

Main Methods:

• Design of a specialized cable-stripping robot considering the constraints of 10 kV overhead line environments.
• Development of a cascade PID control technology integrated with a back propagation neural network (BPNN) for enhanced adaptability and robustness.
• Conducting a 95 mm² cable-stripping experiment to evaluate the robot and controller performance.

Main Results:

• The developed BPNN-PID controller demonstrated superior speed tracking and system stability compared to traditional PID controllers.
• The cable-stripping robot proved structurally feasible for live working on 10 kV overhead cables.
• The robot successfully replaced manual stripping procedures in experimental conditions.

Conclusions:

• The novel cable-stripping robot is a viable and safer alternative to manual methods for 10 kV power grid maintenance.
• The BPNN-PID control strategy offers significant advantages in robustness and adaptability for nonlinear robotic systems in power infrastructure.
• This technology has the potential for practical implementation in routine power maintenance operations, improving worker safety and operational efficiency.