Facility Metrics & Database Evaluation
Summary Facility Level Metrics Development Evaluation of Databases for Calculating Sustainability Metrics along Supply Chains Concluding Remarks Staffing/Oversight	Summary Teams of students and faculty members employed by BRIDGES™ tested the sustainability indicators at the facility level and evaluated the usefulness of university and government databases for the calculation of sustainability metrics along supply chains. A team from Rice University developed metrics for Formosa Plastics, a producer of olefins and olefin derived products. Students and faculty from Georgia Tech calculated metrics for an Interface, Inc. facility, which manufactures carpet tile and is a customer for raw materials derived from olefins. The third team, from the University of Texas at Austin, evaluated metrics produced for a facility manufacturing asphalt roofing tiles. In addition, this team used the Environmental Input-Output Life Cycle Assessment (EIO/LCA) database developed by Carnegie-Mellon University and data from the DOE Industrial Assessment Center (IAC) waste and energy audits to calculate metrics for industrial sectors related to chemical or plastics manufacturing. Evaluation and testing indicators at the facility level focused on their technical feasibility, such as the required degree of precision and availability of data, the clarity of decision rules, definitions and compilation procedures, and on interpretation issues - the meaning that may be ascribed to the indicators by users. Issues that had to be addressed in formulating the facility metrics included:
	How far upstream and downstream in the life-cycle analysis should be explored for data, At what level of the business should the investigation take place (e.g. business unit, facility level, product), How much precision should be sought in data (or will models and assumptions replace precise data), Will equivalent types of data be sought from all participating companies? How to allocate impacts when evaluating integrated facilities?
	Through the process of practical testing, the project teams discovered which indicators were most useful, what data and resources were needed to compile them, how to interpret and apply them at different levels within each company, and how to assess their benefits compared to the costs of producing them.
	Facility Level Metrics Development Formosa Plastics The Formosa Plastics complex at Point Comfort, Texas, is a modern, horizontally integrated chemical plant. Products from the facility include polypropylene, ethylene glycol, linear low density polyethylene, high density polyethylene, ethylene dichloride, polyvinyl chloride, caustic soda and electricity. Inputs to the complex include natural gas that is burned for electricity production and for olefins production, naphtha that is the feedstock for the olefins plant and brine that is the feedstock for the Ion Exchange Membrane (IEM) plant. Additionally, both ethylene and propylene are imported to augment production of polypropylene and ethylene-derived products. The team from Rice University calculated four metrics for the Formosa complex. Metrics for toxics emissions, nitrogen oxide (NOx) emissions, carbon dioxide emissions and water consumption were calculated for seven different products on a mass basis (per pound of product). There were three distinct steps in the development of the metrics calculations. First, data associated with the metrics calculations had to be assembled. Second, decision rules had to be developed to allocate releases and resource use to the various products. Third, the actual calculation of pounds of release and/or resource use per pound of product was completed. Data for toxics emissions was compiled from the Toxic Release Inventory (TRI) prepared for the Point Comfort facility. Information on nitrogen oxide emissions was obtained from information submitted by Formosa Plastics to the Texas Natural Resource Conservation Commission (TNRCC) as part of the emissions inventory process. The carbon dioxide emissions from the Formosa facility were not assembled prior to the initiation of this project. At the team's request, Formosa assembled information regarding the combustion of natural gas in the utility co-gen units and in the olefins plant. The information given in metric tons of carbon consumption, was converted to carbon dioxide equivalents. Water usage at the facility had been previously evaluated, and a detailed water balance was available that identified the water coming into the facility, the distribution of water throughout the facility and the generation of wastewater associated with this facility. Decision rules were developed to identify how the emissions and resource usage from the intermediate production steps would be allocated within the facility. The olefin plant, for example, produces propylene and ethylene, which are feedstocks elsewhere within the plant. By weight, the products of the plant are 30 % propylene and 70 % ethylene. Therefore, 30 % of the resource use and emissions were allocated to propylene production and 70 % was allocated to ethylene-derived products. This 70 % was further allocated on the basis of ethylene usage for the ethylene-derived products. Decision rules were developed for each metric using a similar mass-based approach. Once the allocations for each product had been determined according to the decision rules, the metrics could be calculated with the data that had been gathered for the entire facility. The toxics, NOx, CO2 and water metrics were calculated on a per pound of product basis for the following seven products: polypropylene, linear low density polyethylene, high density polyethylene, ethylene glycol, ethylene dichloride, vinyl chloride monomer and polyvinylchloride combined, and caustic. Interface, Inc. The team of faculty and students from Georgia Tech calculated metrics for a carpet tile product manufactured by a facility of Interface, Inc. The manufacturing process at the Interface facility takes place in two phases: Plant A receives yarn, chemicals and backing from various suppliers and performs the first stage of the process. When the carpet is 70 % complete, it is shipped to Plant B for the final steps in the process. The metrics for the carpet product were developed taking into consideration the combined effect of both plants, and the effect of shipping the carpet from Plant A to Plant B. Data for the metrics calculations were obtained from annual production and sales reports, Toxic Release Inventory data and mass balance information from the facility's engineering department. Decision rules for allocation of the emissions and resource usage were based on the calculation of a product contribution index, which defined the product contribution to the total plant production processs, and was derived from the mass balance for the facility as a whole. The Georgia Tech team calculated five metrics for the Interface facility-material intensity, water consumption, energy intensity and toxics and pollutants dispersion using three output denominators (mass of product, revenue and value-added). The teams reached several conclusions regarding the development of metrics at the facility level. In regard to obtaining the necessary data, much of the information which is needed for developing allocation rules and computing the metrics can be obtained from mass balance information which is usually readily available at a facility. TRI information, which is another source of data, is also accessible. The metrics appear to be most useful as a basis for comparison across products. In order for comparisons to be meaningful, however, a clear understanding of the methodology and the decision rules used in calculating them is critical. Evaluation of Databases for Calculating Sustainability Metrics along Supply Chains Environmental Input-Output Life Cycle Assessment Carnegie-Mellon University has a tool for assessing emissions, waste generation and energy use along supply chains (www.eiolca.net). The tool, an Environmental Input-Output Life Cycle Assessment (EIO/LCA), uses data on emissions, waste generation and energy use in approximately 500 economic sectors, identified by Standard Industrial Classification (SIC) codes. Supply chains are identified by tracking the flow of dollars between the 500 economic sectors. By assuming that the emissions, waste generation, and energy use scale with dollars of economic activity, the model is able to predict the emission, waste and energy impacts of an increase in economic activity in any of the 500 sectors, along the supply chain. The team from the University of Texas examined the supply chain impacts of 14 industrial sectors using the EIO/LCA model. These sectors were in the broad category identified by the EIO/LCA model as chemical or plastics manufacturing. For each of these sectors, the EIO/LCA model was used to estimate energy use, carbon dioxide emissions, RCRA wastes generated, and TRI releases along the supply chain. Results of the analyses showed significant variation in the environmental metrics for the various chemical manufacturing industries. Supply chain contributions were significant for all of the metrics except TRI releases, which may change when TRI reporting is expanded to include facilities. The EIO/LCA tool was also used to move further down the supply chain, to incorporate customers of the chemical industry. Environmental performance metrics along the life cycles of three products were calculated-floor covering products, miscellaneous plastic products and roofing materials. The results for roofing materials were quite different from those of floor covering and plastics products. As might be expected, asphalt roofing products have a much greater reliance on petroleum processing, and the total energy use along the supply chain was approximately double that observed for the plastic and floor covering products. Conversely, the RCRA waste generation rates and TRI releases along the supply chain are lower for roofing products than for plastics or floor coverings because the roofing products have less of a reliance on industrial organic chemicals. For all three products, the data indicated that customer (product manufacturer) contribution to the supply chain impacts is relatively small in each of the categories. To assess the accuracy and relevancy of the EIO/LCA data in evaluating environmental performance metrics, metrics calculated for the plastics and resins category were compared to those computed from life cycle inventory data collected in a project sponsored by the European Plastics Industry. The agreement between the two sources of information was remarkably good, lending support to the validity of the EIO/LCA data as a source of rough benchmarking data for environmental metrics. When the same metrics were compared to those computed for the Formosa Plastics facility, the comparison suggested that the Texas facility performs substantially better that industry averages, although the comparisons must be viewed with caution. The EIO/LCA and life cycle inventory data include contributions from feedstock production, which are not included in the data for the Texas facility. Even taking this difference into account, however, it appears as though the Texas facility's environmental performance metrics are significantly better than sector norms. The team concluded that overall the EIO/LCA database was easy to use and contained data necessary for calculating many of the core and supplementary sustainability metrics, using dollars of sales to normalize the data. Drawbacks are that some sectors are not sufficiently represented and that it is difficult to quality assure the data. DOE Industrial Assessment Center The DOE Industrial Assessment Center (IAC) information was also used to evaluate metrics. The Department of Energy funds Industrial Assessment Centers (IACs) that periodically perform waste and energy audits at industrial sites. The data collected during these site visits include electricity use, fuel oil use, hazardous waste generation, non-hazardous waste generation, air emissions, and water use. Data are also collected on sales and production activity. Metrics calculated with this information are for the manufacturing step only and can be compared indirectly to the metrics calculated from the EIO/LCA data, which integrates along the life cycle of products. Using information contained in the IAC database, electricity use per unit of sales, water use per unit of sales and waste generation per unit of sales were calculated. These metrics were examined because they could be calculated directly from the IAC data and because they closely matched metrics that could be calculated using the EIO/LCA data. The values for the metrics, based on the IAC data, showed substantial variability, ranging over several orders of magnitude, for facilities with similar descriptions (SIC codes, lists of products). In general, variability was greatest in water use and least in electricity use. The variability and the uncertain quality make the IAC data of little value in establishing benchmarks for environmental performance metrics (except for electricity use). The team noted, however, that the IAC data were not collected for the purpose of evaluating environmental performance metrics. The fact that the IAC data could be used at all in this project is an indication that the information contained in the proposed environmental performance metrics is valuable and that the information necessary to evaluate the metrics is accessible. Concluding Remarks The results of this project are not expected to measure and communicate all aspects and details of industrial performance, whether at corporate, division, facility or product level - other indicators and data such as absolute quantities will also be necessary. The indicators, however, will allow companies to establish overall goals for "roadmapping" and implementation phases of the vision projects for each industry segment. The indicators should assist in management of companies, their boards of directors and eventually external stakeholders in tracking progress toward performance targets, and facilitate meaningful comparison of performance between companies and across sectors. Staffing/Oversight The principal investigators were Beth R. Beloff and Earl R. Beaver of BRIDGES to Sustainability in Houston. Other Investigators were: Tom Graver at Georgia Tech, David Allen at the University of Texas in Austin, Jim Blackburn at Rice University, and the teams of students at these universities.