Benchmarking data sources

Individuals who prepare benchmarkcalculations for code validation are well aware that the benchmark- data sources are voluminous. The reports collected for our bibliography on criticality experiments occupy three five-drawer file cabinets. Finding and selecting the right data can be a trying task. The difficulty is compounded because, in many cases, the same experimental data are reported in several publications, sometimes with corrections of earlier errors, or in an evaluated clean format. To do their jobs properly, the criticality analysts have to be aware of the best sources for their needs. This paper addresses how some tools that are being developed, especially computer data-base systems, can make the job easier. In this paper, we show wlut we have done to make this task easier, and how it may be of beneHt toothers. 

The PACE-T(H2) model analyses the world economy based on the GTAPinGAMS structure (Rutherford, 1998). It incorporates top-down benchmark data in the form of social accounting matrices to which the model is calibrated. The main data source for the calibration of national and international commodity flows is the GTAP5 database (GTAP, 2002). The trade shares of hydrogen and conventional cars are taken from the sector Motor vehicles and parts of the GTAP database. 

This step involves planning the process of using external potentially useful information sources, including so-called solution databases (such as compiled by Lyngby, the Danish EPA and the Dutch TNO), safety perfor-mance/benchmarking data, literature on process safety and reliability, literature on cleaner production/pollution prevention, academic experts/... 

Common Data Source Approach. Another approach is to use a common source of data for benchmarking a set of chemicals. Examples of potential data sources include Material Safety Data Sheets (MSDSs) from product manufacturers, the Hazardous Substances Data Bank (HSDB), the Integrated Risk Information System (IRIS), the International Uniform Chemical Information Database (lUClID), the High Rroduction Volume Information System (HRVIS), the Organisation for Economic Cooperation and Development (OECD) Screening Information Dataset (SIDS), and the Canadian Domestic Substances list... 

Uncertainty information provided with the benchmark data is often inadequate. In addition, recent sensitivity studies have highlighted the need for covariance data between the integral parameters. Detailed studies of uncertainties in experimental critical masses and reaction rate ratios have been made for several of the more recent benchmarks. The experimental data were reviewed and the individual sources of uncertainty reassessed. Frpm these data, the covariance matrix for a set of measurements has been developed. Appropriate weighting for several measurements of the same quantity, in some cases, hasle.d to significant revisions of the benchmark parameters. These revised data and covariances have been used in deveiopment of the ORACLE adjusted library for LMFBRs (Ref. 5). 

In addition to the insightful use of existing data, the acquisition of new chemical and physical property data continues to grow in importance— as does the need to retrieve data for future needs. Such efforts require careful experimental measurements as well as skilled evaluation of related data from multiple sources. It will be necessary to assess confidence with robust uncertainty estimates validate data with experimentally or calculated benchmark data of known accuracy and document the metadata needed for intapretation. 

A range of studies have now appeared modeling complete catalytic cycles based on an initial C-H bond activation and the subsequent functionalization steps. The functionalization processes involve reactions with an insertion partner (an alkyne, an alkene, or a carbene/nitrene source) and lead to the formation of new heterocychc rings or direct replacement of the C—H bond with a new C-R group. In many cases, the focus of the computational studies is on these functionalization processes as well as the reoxidation steps (particularly with internal oxidants). The C-H activation in these catalytic processes is typically of the intramolecular AML A-6 type, although exceptions do exist, as will be detailed later. Input from parallel experimental studies, including the determination of kfj/ko KIEs, the observation of H/D exchange, competition experiments, and the isolation of key intermediates both complement and provide benchmark data... 

FPEtool is described in Deal (1995). No formal benchmarking has been performed, nor has any source code validation or validation against experimental data but the code is based on a number of experiments Notarianni (1993), Walton (1993), Nelson (1992), Peacock (1993b), Heskestad (1986) and Hinkley (1975). 

The target level procedure was applied to 16 common air contaminants (Table 6.19). These are common contaminants in the industrial environment, and in many cases are the most critical compounds from the viewpoint of need for control measures. The prevailing concentration data as well as the benchmark levels were taken from Nordic databases, mainly the Finnish sources, and described elsewhere.In addition, a general model for assessing target values for other contaminants is presented in the table. 

Occupational and toxicological studies have demonstrated adverse health effects from exposure to toxic contaminants. Emissions data from stationary and mobile sources are used in an atmospheric dispersion model to estimate outdoor concentrations of 148 toxic contaminants for each of the 60,803 census tracts in the contiguous United States for 1990. Approximately 10% of all census tracts had estimated concentrations of one or more carcinogenic HAPs at a greater than l-in-10,000 risk level. Twenty-two pollutants with chronic toxicity benchmark concentrations had modeled concentrations in excess of these benchmarks, and approximately 200 census tracts had a modeled concentration 100 times the benchmark for at least one of these pollutants. This comprehensive assessment of air toxics concentrations across the United States indicates hazardous air pollutants may pose a potential public health problem (Woodruff et al., 1998). 

Diagnostic approaches typically involve collecting data from a variety of sources. In addition to fact-based analyses, such as competitive benchmarking or value driver analyses, surveys and interviews can help in gaining a better understanding of the company s workings and uncovering the complex roots of performance problems ... 

The backbone of the climate forecast, of course, is the operational model that links the short-term El Nino scale to the longer term. The observing system is the key challenge for testing the veracity of calculations. Carbon sources and sinks have been discussed. I believe that upper ocean observations, climate data records at the surface, and benchmark observation that establish the long-term evolution of the climate in an absolute sense constitute the centerpiece of what must be done. 

---