Multi-Objective Design Optimization of the Leg Mechanism for a Piping Inspection Robot

This paper addresses the dimensional synthesis of an adaptive mechanism of contact points ie a leg mechanism of a piping inspection robot operating in an irradiated area as a nuclear power plant. This studied mechanism is the leading part of the robot sub-system responsible of the locomotion. Firstly, three architectures are chosen from the literature and their properties are described. Then, a method using a multi-objective optimization is proposed to determine the best architecture and the optimal geometric parameters of a leg taking into account environmental and design constraints. In this context, the objective functions are the minimization of the mechanism size and the maximization of the transmission force factor. Representations of the Pareto front versus the objective functions and the design parameters are given. Finally, the CAD model of several solutions located on the Pareto front are presented and discussed. INTRODUCTION In a nuclear power plant, there are many places that the human workers cannot reach due to the high level of irradiation (which can be deadly). However, for safety reasons inherent to a nuclear power plant, the pipe-line equipments (which are large in such a plant) require periodic and rigorous inspections. In this context, the development of robotic system suitable to finding and to repaire a failure in such an environment is essential. It is for this reason, since many years, numerous articles appeared on this subject. In [1], the key issues raised by the design and the control of a piping inspection robot are presented. More generally, in the field of pipe inspections, four following major issues have to be faced: 1) how to move in a pipe; 2) how to locate in a pipe; 3) how to inspect a pipe area; 4) how to repair a default. Let us note that the work presented in our paper only deals with the first issue which falls within the locomotion issue. To achieve locomotion in such highly constrained pipe environment, we can classify the robot designs in two categories depending whether one takes inspiration from animals living in narrow spaces or from engineering knowledge and nine associated subcategories [2]. In the first category, we can imagine designs inspired from the earthworms [3], the snakes [4], the millipedes [5], the Lizards [6] and even from soft animals as the octopus [7]. For the second one, we find designs based on the using of wheels and pulleys [8], the telescopic [9], the impact [10] and the natural peristalsis [11]. The main problem of all these solutions is that each designed robot has a specific architecture for a given specification which is different in each case. Thus it is difficult to find the best architecture that responds to an user request. Despite this fact, the common denominators of all these systems are the mechanisms to adapt the contact points of the robot on the pipe surface and those to generate the expected contact forces required by the desired motion in the pipe. It is in this context that this paper takes place. Thus, the aim of the paper is to design a mechanism able to adapt the contacts on the inner surface of a pipe under constraints inherent to the environment. More precisely, before defining the complete architecture of our piping robot, we want to find by an optimization into a mechanism selection, the best design minimizing the bulk volume of the robot and maximizing the transmission factor between the embedded motor and the contact points. To simplify, in this article, we call now such a system : the legs such as they are depicted in figure 1. In the context of our study, the considered pipe has a variable diameter between 28 mm and 58 mm with bends.