Helical spring design optimization for endoscopic devices using a design-of-experiments approach

摘要

Introduction: Endoscopes are long thin tubular devices for non-invasive visualization of the interior of cavities, canals and vessels. They can be inserted through natural orifices or small surgical incisions. A typical endoscope outer diameter is 10mm and length can vary from 70 to 180mm. While some rigid endoscopes are used, the flexibility of articulated endoscopes is important in many clinical applications. Articulated endoscopes are generally controlled by cables and levers which bend the distal extremity of the device to enable route selection or to change the field of view. Challenges of flexible endoscopes include reliable advancement and steerability of the device tip. Reliable advancement can be addressed by utilization of relatively rigid helical springs within the device shaft. However, steerability requires that the tip must be flexible to accommodate different bend radii. The aim of this study is to understand the main effects (geometric and physical properties of the tip and their interactions) that impact the radius of curvature governing steerability in traditional endoscopes. Methods: Using the Abaqus Scripting Interface (ASI), a helical coil and cable assembly were modeled with the cable connected to the coil proximal circumference (Figure 1). The cable was placed inside the helical spring to represent a common endoscope design configuration. A two-level factorial design-of-experiments methodology was utilized to understand the main geometric and physical effects and their interactions on the attainable radius of curvature (outcome). The effects of five coil parameters were assessed: pitch, height, Young's modulus (E), applied force, and wire width. Abaqus dynamic/explicit was utilized to analyse 32 models generated using ASI. The results were analyzed using a commercial design-of-experiments-specific statistical package (Design-Ease). Based on the Box-Cox plot, a Log transformation of the output response was used to stabilize the variance. A Sum-of-Squares chart with a 2% weighted contribution threshold was utilized to determine the model inclusion criteria. Analysis of variance (with a Bonferroni correction) was used to determine the significance of the model, effects, and interactions. Results: The model, four of the input parameters (height, E, force, and wire width) and two interactions (between pitch/wire width and pitch/height/E/wire width) were found to be statistically significant (P＜0.0001). The largest contribution to the model was wire width (effect=1.39,58.33% contribution). The force and pitch/wire width interaction each had a negative influence, reducing the radius of curvature (allowing for tighter bends), whereas the height, E, wire width and pitch/height/E/wire width interaction showed a positive influence. Pitch and all other interaction effects below the 2% contribution threshold were applied as model error. Using this model, geometric and physical properties can be optimized to meet specific endoscopic tip design criteria. A small pitch and wire width yield a radius of curvature optimal for acute bending radii, however, if due to design restrictions, a larger wire width is necessary, an increase in pitch may accommodate to yield a sufficiently small bending radius. Conclusion: A robust computational model was developed using a design-of-experiments approach that allows parametric optimization of endoscopic tip design parameters for radius of curvature minimization.