Noise control in industrial workplaces is enforced by health and safety regulations in order to prevent or reduce risks to personnel. Apart from compliance with rules, the adverse effects of noise on productivity have always been a challenge for industry. As a consequence, practical solutions, ranging from protection aids to acoustic damping and isolation, have occasionally been employed. These unplanned remedies do not necessarily aim at higher risk locations and hence may impose significant and unjustified expense on the company. In this paper, the optimum combination of treatments is investigated using binary integer programming with objective cost function. The model constraints include recommended noise doses for highly exposed operators as well as budget limits. In addition, sound specification of the sources, treatment effects and relevant production information are incorporated into the model through structured databases. Then a genetic algorithm is utilized in a Matlab environment and final results are obtained. The procedure is applied to an example of a press shop and the validity of the results is approved.
Noise nuisance in industry, particularly in manufacturing plants, is more crucial than in other working areas. According to NIOSH statistics, 14% of the world’s working population is exposed to a noise level of about 90dBA. In manufacturing, this rate increases to 35% and hence becomes even more difficult to tolerate (NIOSH, 1998). Numerous complaints of noise, received constantly from industrial workers, give clear evidence for this argument. Although the reported discomforts are mainly focused on direct hearing impairment, noise impression on human health is not confined to hearing loss or shifts in sensing thresholds (Irle et al., 1998). It can also have serious physiological and psychological consequences, including heart disease, drowsiness and lack of concentration (Szalma and Hancock, 2011 and Arezes and Miguel, 2005). Long exposure to noise can even generate acute disorders in sleeping and learning and cause poor reactions to warnings, which in turn can lead to more hazardous situations (Eleftheriou, 2010 and Gramopadhye and Wilson, 1997).
Review of the literature reveals that plenty of research studies on noise control are focused on urban areas and its technical issues. Specifically, noise produced by construction activities and transportations, e.g. airport neighborhoods are broadly investigated (Black et al., 2007, Ballesteros et al., 2010, Dekkers and Straaten, 2009 and Henrique and Zannin, 2008). However, economic outcomes of noise control, especially in industrial environments have rarely been discussed (Beevis, 2003 and Walker and Tait, 2004).
In this research, various noise treatments in industries are identified and the related costs are evaluated. The cost function of the model is defined by the sum of the cost reductions both from noise levels and the exposure times. The constraints include the allowable time of exposure and the budget limit. Due to the nature of the model, conventional optimization techniques cannot be applied. Therefore, a genetic algorithm is used, and a computer program is developed. The solution gives the optimum combination of the options, i.e. an optimum policy for noise control taking account of the total cost of its implementation. The model is then verified by sample data and proved to be informative and applicable.
Despite specific regulations for allowable noise doses and various control options available for industrial areas, the cost of the control solutions discourages the managers from implementing them. In order to find an economic policy, a mathematical model based on a genetic algorithm was described in this paper. It aimed to minimize the total cost of the strategies which are combinations of noise control methods allocated to individual sources, transmission and receivers. The noise status of the workplace including sound pressure levels at machineries and the specifications of the control methods form the input to the model. Safety measures, including noise levels and working hours, are defined as constraints, and budget limit is added to assure the feasibility of the solutions. Each chromosome of the model represented a combination of the control options for operators and machineries by a proper number of genes. For enhancing the accuracy of the model, decibel addition technique was used for evaluating the resultant pressure levels from various sources at each location. In addition, the model could identify those operators at higher risk due to noise exposure and eliminate the rest of the personnel for assigning personal equipment. Therefore, the model was downsized, and worthless computations were avoided. A computer program for the model was written in the Matlab environment and executed with sample data.
The results of the program presented the least cost policy which could be directly implemented without any further interpretation. An investigation for a press shop revealed that by spending a typical employee cost for the firm in one year, a practical noise control strategy can be established. Comparing this cost with the risk of the health and safety for the personnel can easily convince any manager to reconsider it. In the future study, the parameters of the model can be taken as fuzzy numbers so that the accuracy will be enhanced.
