Modeling Sand-Shoveling Related Pain Risks with Fuzzy Logic
This study developed a fuzzy linguistic model to predict work-related pain in sand shovelling. The primary objective was to develop a knowledge-based economic tool for ergonomics risk assessment capable of predicting same opinions of injury as obtainable with workers’ self narrated. The model used 81 possible “IF THEN” linguistic rules fired into Mamdani inference engine to make decisions about the rank of risk associated with sand shovelling task variables. Scoop per minute, length of scoop, shovel/load mass and throw span were the four inputs variables used with “Sand-Shoveling Pain (SSP) risks’ as the output. Validation result shows that 70% of the model predictions opinions corresponded to that of the opinion interpretation of the self-narrated numeric pain rating (SNNPR) of the affected 120 workers. The model generated risk (MGR) values had statistically significantly higher level of predicted risk (mean=3.91, SEM=0.47) compared to SNNPR (mean= 3.6, SEM = 0.50), with t(38) = -0.449, p = 0.656 with 95% confidence interval for the difference (-1.71, 1.09). Pearson correlation coefficient of the MGR values and the workers’ SNNPR values was found to be 0.73. The independent sample t-test result (p = 0.667) also indicated a no significant difference of means. The model which could be applied in any workplace where it is necessary to consider ergonomics of work method and/or workplace design for manual shovelling tasks, was able to achieve the targeted objectives hence quality of reality was attributed to it.