1 Copyright © 2004 by ASME Proceedings of IMECE2004: 2004 ASME International Mechanical Engineering Congress & Exposition November 13-19, 2004, Anaheim, California IMECE2004-62599 LIFE CYCLE ANALYSIS OF CONVENTIONAL MANUFACTURING TECHNIQUES: SAND CASTING Stephanie Dalquist Massachusetts Institute of Technology Timothy Gutowski Massachusetts Institute of Technology ABSTRACT Conventional manufacturing techniques have not been subject to much scrutiny by industrial ecologists to date. Many newer techniques and products draw more attention as they rise quickly from research to global scales, amplifying their environmental consequences. Despite the presence of new technologies and increased overseas production, casting activity continues to have a strong presence in the US, and represents a stable component in the national economy. Data from the US government, US industry groups, and UK mass balance profiles facilitate an understanding of sand casting and comparison across manufacturing processes. The figures in the US and UK are similar in terms of diversity of metals (where the US is 72%, 13%, 10% and the UK 76%, 13%, 8% for iron, aluminum, and steel, respectively), energy per ton of saleable cast metal (10.1 and 9.3 million Btu/ton in the US and UK), and overall emissions, with notable similarities in benzene and particulate emissions. One notable discrepancy is in sand use, where the US sends to waste 0.5 tons of sand per ton of cast metal, whereas the UK sends 0.25 tons to landfill. INTRODUCTION Although we live in an age where new technologies demand exotic manufacturing techniques, most products still require traditional manufacturing processes and carry along their inherent environmental ramifications. Developing countries entering mass production, in particular, are taking on an increased environmental burden in manufacturing. Complex products like semiconductors (Williams, 2002) and cars are frequently subjected to life cycle assessments as a part of or in conjunction with environmental impact analyses. For conventional processes like sand casting, such an evaluation is uncommon. Although sand casting has reached a stable market size in the United States, international production is growing (Modern Casting, 2003). China alone increased shipments 60% over the years 1997 to 2002, and in 2002 shipped 17.8 million tons of cast metal. Global casting production showed a 3% overall mass increase in 2002. Cast material requires substantial energy, often in the form of fossil-fuel generated electricity or direct firing of coke or natural gas. Most of this energy is used to melt the metal, but increasing quantities of energy and materials are required to meet customer specifications. The highest material demand, besides the metal forming the product, is the sand used to create the mold. Organic compounds are used as binders, and burned out as gaseous releases during mold formation. More organic compounds are used in cleaning and finishing. In casting, like other sectors of private industry, there is little consensus on the magnitude of the impacts, in part due to the lack of published data. Some firms keep relatively good information, but publicly available aggregate data and sector analyses are limited to self-reported surveys from the DOE and EPA. The lack of evidence that the environmental impact is well understood or well addressed exemplifies the suitability of this sector and need for life cycle research. SYSTEM BOUNDARIES: PROCESS MATERIALS AND ENERGY USE The system boundary outlines the sand casting manufacturing process and the boundaries of this inventory. The resources considered (Figure 1) are the material and energy inputs and outputs for mold preparation, metal preparation, casting, and finishing stages and their subprocesses.2 Copyright © 2004 by ASME Figure 1. System boundaries of analysis and materials flow of sand casting. Note that the manufacture of materials (e.g., extraction of metals or sand and water purification) is not within the system boundary. Also, equipment manufacture is not included. Furnace service life is long enough that manufacturing impacts are in the distant past and end-of-life impacts are yet in the distant future. Due to their long life, the environmental cost of machines is amortized over numerous products and the environmental impact is insignificant on a per part basis. The limited body of literature available suggests greater concern about energy implications of sand casting than the material issues of sand and water consumption. The US Department of Energy (DOE) has published an energy and environmental profile of the casting industry, as has the US Environmental Protection Agency (EPA). Some sand casting foundries file a Toxic Release Inventory (TRI) with the EPA. TRI applies to companies with manufacturing operations in Standard Industrial Classification (SIC) codes 20 through 39 – casting is in 33. TRI applies to companies with more than 10 employees that use 25,000 pounds of approximately 600 designated chemicals or use more than 10,000 pounds of any designated chemical or chemical category (EPA, 2004). 654 of some 2,800 US foundries were required to file a TRI in 1995 (EPA, 1998a). Of the remaining foundries, 33% had fewer than 10 employees, and the other 44% had more than 10 employees but did not use enough regulated material to be required to file. Energy data is available from the Energy Information Administration (EIA) Manufacturing Energy Consumption Survey (MECS). Although the results are obtained from self-reporting questionnaires, the numbers tend to agree with other surveys and with approximations calculated from basic information. Even though sand casting is relatively uniform in concept, the myriad parameters complicate the derivation of accurate systematic estimations on a national scale from individual foundries and equipment. Consequentially, determining concordance between individual foundry data and reported data is an exercise of coincidence. At least three gaps in the known literature on sand casting must be addressed. Two of these are in common with those which have not been addressed in semiconductors (Williams, 2002), demonstrating an issue in life cycle analysis which extends beyond this process. There is a noticeable lack of process data, like input and output materials. This is particularly true of the cleaning process and the components in and
View Full Document
Unlocking...