The parametric study and development of a permanent magnet wheel-based crawling robot for spherical tanks with curvature adaptability

To address the issues of unstable adhesion and poor curvature adaptability of rigidly connected climbing robots in large industrial spherical tanks, a novel climbing robot has been developed. This robot features a hinge that connects the two sides of the body to the chassis, allowing it to passively adapt to the curvature changes of the spherical tank surface with independent degrees of freedom. A mathematical and mechanical model of the robot’s working state was established to analyze the forces acting on various positions and the changes in the hinge rotation angles. Based on this analysis, a structural parametric study was conducted to enhance curvature adaptability—accommodating typical radii ranging from 2.88 to 8.9 m with a defined hinge flipping range of 5 ° inward and 25 ° outward—and reduce the required minimum adhesion force. The parametric analysis involved the magnetization direction of the permanent magnetic wheels (incorporating a Halbach array), the thickness of the yoke (identified optimally at 8 mm), and the arrangement of the magnetic wheels (utilizing a double-layer design with magnetic isolation). Three different magnetic circuit models were proposed and evaluated through simulation, quantitatively improving adhesion strength (achieving a high adhesion force fluctuating between 178 and 190 N per wheel) and magnetic efficiency. Finally, a prototype and experimental platform were built for testing. The results showed that the robot could stably crawl on both the inner and outer surfaces of spherical tanks with curvature at a speed of 3 cm/s. Furthermore, it demonstrated robust obstacle-crossing capabilities by traversing 5 mm high weld seams at 6 cm/s, and exhibited a reliable load-bearing capacity with a 20 kg payload without slipping, supported by a total magnetic adhesion force exceeding 712 N to ensure a substantial safety margin. This provides a basis for the intelligent development of spherical tank inspection.