Forecasting production economics using ergonomics data

Full text

HUMAN FACTORS IN ORGANIZATIONAL DESIGN AND MANAGEMENT – XI
NORDIC ERGONOMICS SOCIETY ANNUAL CONFERENCE – 46
285
Optimizing production cost using ergonomic data!
W. Patrick NEUMANN, Ahmad SOBHANI, M.I.M WAHAB
Department of Mechanical and Industrial Engineering, Ryerson University, Toronto,
Canada
Keywords. Occupational Human Factors, production economics forecasting, work-related
ill health, productivity loss
1. Introduction
This paper presents a case study application of a novel Operations Research (OR)
approach to forecasting production system level financial performance using Human
Factors (HF). OR models rarely include HF aspects in predicting financial outcomes, which
may lead to overestimating the profitability of the system. They usually assign a fixed value
to employee performance and do not investigate Work-related ill health (WIH) risk factors
and their financial and operational consequences when forecasting financial results.
2. Methods
We demonstrate a hierarchical, 3-level approach with worker, workstation, and system
level sub-models). At the individual level; workers are assumed to have three health states:
healthy, in-pain but at work (presenteeism), or injured and absent - in which case a
replacement worker is assumed and claim costs are added. Individual health states are
predicted over time based on the quantified physical risk factor exposures and risk ratios
obtained from epidemiological studies. A Markovian modelling method was used to
determine the probability of an employee being in each of the possible health states in a
given period. At the ‘Workstation Level’ the productivity and quality of an operator will
depend on their health state, as well as the design of the workstation itself. Here
conventional production economic costs are included such as the cost of the materials,
labour time, overhead costs. Each worksation’s performance is passed on to the ‘System
Level’ model which accounts for the configuration of workstations including aspects like
bottlenecks. Then a mathematical model was developed for the production system to
replicate its production behavior and incurred total cost. This system level mathematical
model was optimised to calculate the best-case (lowest) total production cost and
production profile for the system based on in-data provided.
We present a proof of principle demonstration based on a seven workstation auto parts
production facility. Seven physical risk factors for WIH of the shoulder and back were
quantified in the field for each workstation. Since this system had no recent sickness
absence reports the resulting model had only normal and presenteeism health staes.
Optimal production costs were estimated without HF (providing a referent index at 100)
and with HF WIH effects. A one-factor analysis was conducted examining the effects of
varying a cumulative spine load in a single station – the system bottleneck which also had
the highest measured workload. Production volume was held constant in all models.
Neumann, W.P., Sohbani, Wahab, M.I.M. (2014) Forecasting production economics using ergonomics data.
In the proceeding of Human Factors In Organizational Design And Management (ODAM) - XI and Nordic
Ergonomics Society (NES) Annual Conference - 46 O. Broberg, N. Fallentin, P. Hasle, P.L. Jensen, A.
Kabel, M.E. Larsen, T.Weller (Editors), Copenhagen, DK. August 18-20.

O. Broberg, N. Fallentin, P. Hasle, P.L. Jensen, A. Kabel, M.E. Larsen, T. Weller (Editors) 2014
286
3. Results
When all seven risk factors were considered at their quantified levels the total cost of
production was 1.23% greater then when the costs were calculated without including HF
effects. Figure 1 illustrated the effects of varying the single Cumulative Spinal Load risk
factor for the bottleneck workstation. This showed that the underestimation of costs,
compared to non-HF assessement, could be as low as 0.7% and as high as 1.5% extra as
this single risk factor was changed. For comparison, the observed cumulative compression
as quantified was 20.6 Mega Newton seconds per full time shift (MNs/shift) – the mid
point on Figure 1. The calculated cost was aligned with the historical data of the facility
according to managers of the facility.
4. Discussion and Conclusions
This study has demonstrated the feasibility of predicting total production costs based
on conventional economic data supplemented with WIH risk factor data that can be
quantified in the field. This is a unique capability as conventional economic models
usually ignore HF aspects and health economic models do not account for system level
production costs. Furthermore this system allows potential improvements in HF to be
evaluated in terms of total production costs for the system.
In this particular case, there were no recent lost time accidents so the impact of HF in
economic terms was limited to presenteeism effects. While it is technically possible to
include changes in quality levels, we were unable to isolate the human factors related
fraction of quality deficits in the system in this case Furthermore the available research
quantifying the mathematical relations between HF and production quality is sparse.
Future work must focus on broadening the application of the technique to more sites as well
as followup studies to examine the actual financial costs incurred by the company. While
validity concerns remain, this technique may provide a valuable tool to help managers
understand the impact of poor HF in terms of the performance of the company.
Acknowledgements
This work was made possible with financial support from the Workplace Safety and
Insurance Board of Ontario. The participation of company personnel is also acknowledged.
Figure 1: Total cost index change versus cumulative compression level variations in
a single workstation while all other risk factors are kept constant
Cumulative Compression (MN.s/shift)