CLEAN MANUFACTURING:METHODS

METHODS

Traditional methods for clean manufacturing focus on waste or energy audits, which are summarized in Section 4.1. New methods focus on life cycle design, life cycle assessment, production planning models with environmental considerations, and environmental management systems, which are de- scribed in Sections 4.2, 4.3, 4.4, and 4.5, respectively.

Waste / Energy Audits for Waste / Energy Minimization

Waste and energy audits require a detailed inventory analysis of waste generation and energy con- sumption. The point of origin of each waste and the breakdown of the equipment energy consumption patterns must be determined. Audits are used to identify significant sources of waste and energy costs. Because some environmental impacts are interconnected, both individual source and system impacts must be considered. General guidelines are given for waste audits and energy audits in the next two subsections.

Waste Audits

Waste audits may be performed at the waste, product, or facility level. Waste-level audits simply require that each waste stream and its source be identified. Although this approach is the simplest, it ignores the implications and interactions of the waste stream as a whole. Waste audits performed at the product level are product life cycle inventory assessments, which are discussed in Section 4.3. Facility waste audits are the most common type of audit because most environmental laws require discharge reporting by facility. Estimating plant-wide emissions is discussed in Chapter 19.

Facility waste audits require process flow charts, product material information (commonly from the bill of materials), process material information (such as cutting fluids in a machine shop), and

environmental information (solid, liquid, and gaseous wastes). Waste auditing guides are available from state-funded programs (e.g., Pacific Northwest Pollution Prevention Resource Center 1999). Allen and Rosselot suggest that waste audits answer the following series of questions: What waste streams are generated by the facility? in what quantity? at what frequency? by which operations? under what legal restrictions or reporting requirements? by which inputs? at what efficiency? in what mixtures? (Allen and Rosselot 1997)

Waste audits require identification of solid wastes, wastewater, direct and secondary emissions. In the United States, solid wastes may be classified as nonhazardous or hazardous according to the Resource Conservation and Recovery Act (RCRA). In general, wastewater is the most significant component of total waste load Allen and Rosselot (1997). Several methods for estimating the rates of direct (fugitive) and secondary emissions are outlined with references for further information in Allen and Rosselot (1997).

Once companies have identified their major wastes and reduced them, they can turn their focus toward prevention. Pollution-prevention checklists and worksheets are provided in U.S. EPA (1992) and Cattanach et al. (1995). Process- and operations-based strategies for identifying and preventing waste are outlined in (Chadha 1994). Case studies sponsored by the U.S. Department of Energy NICE3 program detail success stories for cleaner manufacturing or increased energy efficiency for several major industries (see Office of Industrial Technologies, NICE3, www.oit.doe.gov/nice3/).

Energy Audits

Energy audits may be performed at either the facility or equipment level. Plant-wide energy audits are most common because utility bills summarize energy usage for the facility. Facility energy audits focus on characteristics of use such as occupancy profiles, fuel sources, building size and insulation, window and door alignment, ventilation, lighting, and maintenance programs. (Facility audit forms and checklists are available on the Web from the Washington State University Cooperative Extension Energy Program, www.energy.wsu.edu/ten/energyaudit.htm.) Some industries have developed spe- cialized audit manuals. For example, an energy audit manual for the die-casting industry developed with funds from the state of Illinois and the Department of Energy describes how to assess energy use for an entire die casting facility (Griffith 1997). In addition to industry-specific energy consump- tion information, the U.S. Department of Energy Industrial Assessment Centers provide eligible small- and medium-sized manufacturers with free energy audits to help them identify opportunities to save energy and reduce waste (Office of Industrial Technologies 1999). Energy management is described in Chapter 58.

At the equipment level, energy usage may be determined through engineering calculations or monitors placed on the equipment in question. Identifying equipment with significant energy con- sumption may lead to actions such as adding insulation or performing maintenance.

Waste and energy audits are performed to identify existing problems. In the next four subsections, new approaches are presented that focus on prevention through life cycle design, life cycle assess- ment, production planning models with environmental considerations, and environmental management systems.

Life-Cycle Design*

The design and implementation of manufacturing activities and products have environmental impacts over time. Thus, industrial ecology requires consideration of the materials and energy consumption as well as effluents from resource extraction, manufacturing, use, repair, recycling, and disposal. Environmental considerations for product design and process design are summarized in the next two subsections.

Product Design

Product design guidelines for clean manufacturing are scattered throughout the industrial, mechanical, environmental, and chemical engineering, industrial design, and industrial ecology literature with labels such as ‘‘life-cycle design,’’ ‘‘design for environment (DFE),’’ ‘‘environmentally conscious design,’’ and ‘‘green design.’’ Traditionally, product design and materials selection criteria included geometric, mechanical, physical, economic, service environment, and manufacturing considerations. Industrial ecology introduces criteria such as reducing toxicity, avoiding resource depletion, increasing recyclability, and improving product upgradeability. The product design criteria in Table 4 are cate- gorized by component and assembly level. As design functions for complex products are increasingly distributed, it is important to recognize product level considerations so that local, component design efforts are not cancelled out. For example, if a simple repair module is inaccessible, the design efforts for easy maintenance will be lost.

* This section has been adapted and reprinted with permission from Stuart and Sommerville (1997).

Because design decisions may result in environmental burden transfers from one stage in the life cycle to another or from one medium to another, it is important to recognize life cycle environmental impacts. Therefore, Table 4 is also categorized by five life-cycle stages: process, distribution, use, refurbishment, and recycling. Note that the process design stage for clean manufacturing is detailed separately in Section 4.2.2 and in Table 5.

One of the emerging themes in Table 4 is dematerialization. Dematerialization focuses on using fewer materials to accomplish a particular task (Herman et al. 1989). For example, consumers may subscribe to periodicals and journals on the Web rather than receive printed paper copies. Clearly, miniaturization, information technology, and the World Wide Web are increasing the potential for

dematerialization. Evaluating the criteria in Table 4 to avoid resource depletion or to use renewable materials requires assumptions to be made regarding the uncertainties in technological improvement, material substitutability, and rates of extraction and recycling (Keoleian et al. 1997).

The design criterion to increase product life reduces the number of products discarded over time. In the mid-1970s, DEC VT100 terminals could be disassembled quickly without tools to access the processor for maintenance and upgrades (Sze 2000). In 1999, the Macintosh G4 was introduced with a latch on the side cover that provides quick access to upgrade memory and other accessories (Apple 1999). Product life extension is especially important for products with short lives and toxic materials. For example, battery manufacturers extended the life of nickel–cadmium batteries (Davis et al. 1997). An example of product toxicity reduction was the change in material composition of batteries to reduce mercury content while maintaining performance (Tillman 1991). In another example, popular athletic shoes for children were redesigned to eliminate mercury switches when the shoes were banned from landfills (Consumer Reports 1994). The criteria related to recyclability may apply to product material content as well as the processing materials described in the next section.

Process Design

The criteria for process design for clean manufacturing focus on minimizing pollution, energy con- sumption, water consumption, secondary processes, or redundant processes. Table 5 provides a sum- mary of suggested guidelines for materials selection and process design.

Careful process design can reduce environmental impacts and processing costs. For example, many companies eliminated the cleaning step for printed circuit card assembly by changing to low- solids flux and controlled atmospheres. These companies eliminated the labor, equipment, materials, and waste costs as well as the processing time associated with the cleaning step (Gutierrez and Tulkoff 1994; Cala et al. 1996; Linton 1995). Another example of reducing processing material consumption is recycling coolants used in machine shops. Recycling coolant reduces coolant con- sumption as well as eliminates abrasive metal particles that can shorten tool life or scar product surfaces (Waurzyniak 1999).

Product Life-Cycle Assessment*

Life-cycle assessment (LCA) is a three-step design evaluation methodology composed of inventory profile, environmental impact assessment, and improvement analysis (Keoleian and Menerey 1994). The purpose of the inventory step is to examine the resources consumed and wastes generated at all stages of the product life cycle, including raw materials acquisition, manufacturing, distribution, use, repair, reclamation, and waste disposal.

Materials and energy balance equations are often used to quantify the inputs and outputs at each stage in the product life cycle. Vigon et al. (1993) defines multiple categories of data for inventory analysis, including individual process, facility-specific, industry-average, and generic data. The most desirable form of data is the first data category, data collected from the process used for a specific product. However, this data category may require extensive personnel, expertise, time, and costs.

A three-step methodology for activity-based environmental inventory allocation is useful in cal- culating data for the first step of life cycle assessment. First, the process flow and system boundary

* The following section is adapted and reprinted with permission from Stuart et al. (1998). Copyright MIT Press Journals.

are determined. Then the activity levels and the activity percentages of the inputs and outputs are identified. Finally, the activity percentages are used to determine the actual quantities of the inputs and outputs and assign them to the product and process design combination responsible for their generation. A detailed example of this method is given in Stuart et al. (1998). As industrial engineers assist companies in calculating the allocation of their wastes to the product responsible, they can help managers make more informed decisions about product and process design costs and environ- mental impacts.

Industry-average and generic data must be used with caution because processes may be run with different energy requirements and efficiencies or may exhibit nonlinear behavior (Barnthouse et al. 1998; Field and Ehrenfeld 1999). For example, different regions have different fuel-producing in- dustries and efficiencies that will have a significant effect on the LCA if energy consumption is one of the largest impacts (Boustead 1995).

Once the inputs and outputs are determined, the second and third steps of LCA, impact analysis and improvement analysis, can be pursued (Fava et al. 1991). For impact analysis, the analyst links the inventory of a substance released to an environmental load factor such as acid deposition, which is defined in Table 6 (Potting et al. 1998). Environmental load factors are a function of characteristics such as location, medium, time, rate of release, route of exposure, natural environmental process mechanisms, persistence, mobility, accumulation, toxicity, and threshold of effect. Owens argues that because inventory factors do not have the spatial, temporal, or threshold characteristics that are inherent to the environmental load, other risk-assessment tools should be used to evaluate a local process (Owens 1997).

LCA software tools and matrices may be used to estimate environmental load factors for impact analysis (Graedel 1998; Graedel et al. 1995). Alting and Legarth (1995) review 18 LCA tools for database availability, impact assessment methodology, and complex product capability.

The results of life-cycle impact assessment (LCIA) provide relative indicators of environmental impact. Eight categories for LCIA are defined in Table 6. Details regarding how to use life cycle impact assessment categories are provided in Barnthouse et al. (1998) and Graedel and Allenby (1995).

Life cycle assessment is a comprehensive, quantitative approach to evaluate a single product. ‘‘An extensive survey of the use of mathematical programming to address environmental impacts for air, water, and land is given in [Greenberg (1995)]. A review of applied operations research

papers on supply chain analysis and policy analysis with respect to environmental management is in [Bloemhof-Ruwaard et al. (1995)]’’ (Stuart et al. 1999). Models for production planning with envi- ronmental considerations are summarized in the next section.

Production Planning with Environmental Considerations

‘‘Introduction of product designs and process innovation requires a company to evaluate complex cost and environmental tradeoffs. In the past, these have not included environmental costs’’ (Stuart et al. 1999). In this section, production planning models are described for the entire product life cycle as well as for different stages of the product life cycle.

Models for Production Planning over the Product Life Cycle

Stuart et al. (1999) developed the first mixed integer linear programming model ‘‘to select product and process alternatives while considering tradeoffs of yield, reliability, and business-focused envi- ronmental impacts. Explicit constraints for environmental impacts such as material consumption, energy consumption, and process waste generation are modeled for specified assembly and disassem- bly periods. The constraint sets demonstrate a new way to define the relationship between assembly activities and disassembly configurations through take-back rates. Use of the model as an industry decision tool is demonstrated with an electronics assembly case study in Stuart et al. (1997). Man- ufacturers may run ‘‘what if’’ scenarios for proposed legislation to test the effects on design selection and the bottom line cost impacts.’’ The effects over time of pollution prevention or product life extension are analyzed from a manufacturer’s and potentially lessor’s perspective. Several new models explore the relationship between product and component reuse and new procurement. These models include deterministic approaches using mixed integer linear programming (Eskigun and Uzsoy 1998) and stochastic approaches using queueing theory (Heyman 1977) and periodic review inventory mod- els (Inderfurth 1997; van der Laan and Salomon 1997). Location of remanufacturing facilities are analyzed in Jayaraman (1996); Bloemhof-Ruwaard et al. (1994, 1996); Fleischmann et al. (1997). Scheduling policies for remanufacturing are presented in Guide et al. (1997).

Production Planning Models for the Manufacturing and Assembly Stage Early models focused on reducing the environmental impacts concurrent with process planning for continuous processes in the petroleum and steel industries (Russell 1973; Russell and Vaughan 1974). Recently, models with environmental considerations focus on process planning for discrete product manufacturing (Bennett and Yano 1996, 1998; Sheng and Worhach 1998).

Disassembly Planning Models

Based on graph theory, Meacham et al. (1999) present a fast algorithm, MAXREV, to determine the degree of disassembly for a single product. They are the first to model selection of disassembly strategies for multiple products subject to shared resource constraints. They use their MAXREV algorithm to generate maximum revenue disassembly configurations for their column generation procedure for multiple products. Other disassembly models based on graph theory approaches focus on determining economic manual disassembly sequences for a single product (Ron and Penev 1995; Penev and Ron 1996; Zhang and Kuo 1996; Johnson and Wang 1995; Lambert 1997). A process planning approach to minimize worker exposure hazards during disassembly is given in Turnquist et al. (1996). Disassembly may be economically advantageous for module and component reuse. How- ever, for material recovery, escalating labor costs favor bulk recycling.

Production Planning Models for Bulk Recycling

Production planning for bulk recycling is in the early stages of development. Models include a macro- level transportation model for paper recycling (Glassey and Gupta 1974; Chvatal 1980) and a goal- programming model for recycling a single product (Hoshino et al. 1995). Sodhi and Knight (1998) develop a dynamic programming model for float–sink operations to separate materials by density. Spengler et al. (1997) present a mixed integer linear programming model to determine the manual disassembly level and recycling quantity. Stuart and Lu (2000) develop a multicommodity flow model to select the output purity by evaluating various processing and reprocessing options for bulk recy- cling of end-of-life products. Isaacs and Gupta (1998) use goal programming to maximize disassem- bler and shredder profits subject to inventory balance constraints for the automobile recycling problem. Krikke et al. (1998) propose dynamic programming for disassembly planning and an al- gorithm to maximize revenue from material recycling. Realff et al. (1999) use mixed integer linear programming to select sites and determine the amount of postconsumer material collected, processed, stored, shipped, and sold at various sites.

Environmental Management Systems

An environmental management system (EMS) is a management structure that addresses the long- term environmental impact of a company’s products, services, and processes. An EMS framework should include the following four characteristics:

1. Environmental information collection and storage system

2. Management and employee commitment to environmental performance

3. Accounting and decision processes that recognize environmental costs and impacts

4. Commitment to continuous improvement of environmental performance

Federal U.S. EMS guidelines and information are documented in (Department of Energy 1998). International standards for EMS will be discussed in Section 4.5.3. Environmental management sys- tems (EMS) include environmental policies, goals, and standards, which are discussed in the next three subsections.

Corporate Environmental Policies

Corporate environmental policies require the commitment and resources of senior management. These policies often focus on actions that can prevent, eliminate, reduce, reuse, and recycle, respectively. These policies should be incorporated into all employees’ practices and performance evaluations. Communication of environmental policies and information is integral to the success of the policies. Setting viable goals from corporate environmental policies is the subject of the next section.

Environmental Goals and Metrics

Traditional environmental metrics often focus on compliance with legislation. Goals may concentrate on state-dependent benchmark metrics such as reducing emissions, reducing the volume or mass of solid waste, or reducing gallons of waste water to a specified level. On the other hand, goals may focus on non-state-dependent improvement metrics such as reducing the percentage of environmental treatment and disposal costs. It is also important to distinguish between local and aggregate data when developing goals.

Metrics may focus on local product or process goals or system-wide facility or company goals. An example of a local goal might be to lengthen tool life and reduce cutting fluid waste disposal costs. Sometimes local goals may translate to system goals. One machinist’s use of a new oil-free, protein-based cutting fluid that eliminates misting and dermatitis problems but provides the necessary lubricity and cooling may be a candidate for a system-wide process and procurement change (Koelsch 1997). With local goals, it is important to investigate their potential positive and negative impacts if implemented throughout the system. An example of a system-wide goal might be to reduce the percentage of polymer sprues and runners discarded at a particular facility by implementing regrinding and remolding or redesigning the mold. For system goals, it is important to identify the most sig- nificant contributors through Pareto analysis and target them for improvement.

ISO 14000 Series Standards

ISO 14001, ‘‘an international standard describing the basic elements of an environmental management system, calls for identification of the environmental aspects and impacts of a company’s products, processes, and services [Block 1997]. Industrial engineers may develop the information systems to quantify environmental aspects such as input materials, discharges, and energy consumption [Alex- ander 1996]’’ (Stuart et al. 1998).