A Follow-Up on the Methodical Framework for the Identification, Analysis and Consideration of Uncertainty in the Context of the Integration of Sensory Functions by Means of Sensing Machine Elements

Abstract

When integrating sensing machine elements for in-situ measurements in technical systems, special attention must be paid to uncertainty to ensure the reliability of the provided information. Therefore, a methodical framework for the identification, analysis and consideration of uncertainty was already developed in prior research, which still offers room for improvement regarding the included methods and tools. Therefore, in this contribution, the initially proposed methods and tools are adapted and extended to enhance their efficiency and applicability and to reduce their error proneness in order to increase the acceptance of the framework in practice. First, the identification of uncertainty is improved by means of an extended effect graph for an automated identification of disturbance factor induced data and model uncertainty. Second, the significance of the subsequent evaluation of uncertainty is enhanced by replacing the initially proposed local sensitivity analysis with a global sensitivity analysis. Finally, a flowchart is proposed that supports the identification of applicable and promising strategies for the development of measures to consider critical disturbance factor induced uncertainty.

Evaluation of the methodical framework for the management of uncertainty in the context of the integration of sensory functions

“…Therefore, a filter option can be used to discard all results that include physical effects that do not occur in the considered system. (Welzbacher et al, 2022;Welzbacher et al, 2023) In the next step, identified data and model uncertainty is evaluated in a modified Failure-Mode and Effects-Analysis (FMEA) to determine its criticality for the reliability of the data provided by the sensory function. Uncertainty is evaluated in this context using three different criteria:…”

“…The evaluation schemes range from 1 to 5, where 1 means a small and 5 a significant severity, deviation or impact. (Welzbacher et al, 2022;Welzbacher et al, 2023) Based on the results of the evaluation, a reasoned decision can be made whether or not an uncertainty is critical for the reliability of the data provided by the sensory function. Therefore, rigid threshold values are used that are oriented at, e.g., the maximum permissible measuring error of the sensory function.…”

“…Therefore, rigid threshold values are used that are oriented at, e.g., the maximum permissible measuring error of the sensory function. (Welzbacher et al, 2022;Welzbacher et al, 2023) In the final step of the methodical framework, measures are developed to eliminate or reduce critical uncertainty. In order to methodically support the user in this step, Welzbacher et al (2023) adapted the Robust Design strategies by Mathias et al (2010) for mechanical load bearing systems and refined them for sensory functions.…”

“…(Welzbacher et al, 2022;Welzbacher et al, 2023) In the final step of the methodical framework, measures are developed to eliminate or reduce critical uncertainty. In order to methodically support the user in this step, Welzbacher et al (2023) adapted the Robust Design strategies by Mathias et al (2010) for mechanical load bearing systems and refined them for sensory functions. To enable the selection of suitable refined Robust Design strategies, a flowchart was developed to empower the user to identify suitable strategies for the considered application case.…”

“…To answer the specific user questions, information from preceding steps of the methodical framework is used. (Welzbacher et al, 2023)…”