Data collection generally

The results from each study are considered on a case by case basis and the subsequent analysis and interpretation will be dependent on the type of study and the data collected. Generally, data comparison is made between test and control substances. Standard analysis for the majority of skin irritation studies includes a breakdown of the range of assessment grades elicited by each substance tested, a summary of subjective comments and some form of statistical analysis. 

The same general comments hold as for Unit 3. Figure 7 provides an example of the AE monitoring data collected from 19.06.97 to 16.07.97, in terms of the main plant parameters vs time (fig. 7a), as well as of the AE RMS values (fig. 7b). 

In many molecular dynamics simulations, equilibration is a separate step that precedes data collection. Equilibration is generally necessary lo avoid introducing artifacts during the healing step an d to en su re th at the trajectory is aciii ally sim u laiin g eq u i librium properties. The period required for equilibration depends on the property of Interest and the molecular system. It may take about 100 ps for the system to approach equilibrium, but some properties are fairly stable after 1 0-20 ps. Suggested tim es range from. 5 ps to nearly 100 ps for medium-si/ed proteins. 

Area Detectors. A two-dimensional or area detector attached to a powder diffractometer can gready decrease data collection time. Many diffraction appHcations require so much time with a conventional detector that they are only feasible if an area detector is attached to the iastmment. The Siemens General Area Detector Diffraction System (GADDS) uses a multiwire area detector (Fig. 17). This detector measures an x- and ajy-position for each x-ray photon detected. The appHcations foUow. 

A powerhouse (thermal) application is the most stringent application, as discussed in Section 7.19. Based on field data collected from various installations by different agencies the general insulation failures observed may be attributed to the following. 

The use of a measurement generally dictates the circumstances of data collection. For example, to provide a best estimate of plume transport direction, hour by hour, of a release from a 75-m stack, a wind vane at the 100-m level of a tower will probably provide more representative wind direction measurements than a vane at 10 m above ground. If the release has buoyancy so that it rises appreciably before leveling off, even the 100-m measurement may not be totally adequate. 

The overall conclusion that can be drawn from a survey of CPI data collection systems is that the better systems do attempt to address the causes of human error. However, because of the lack of knowledge about the factors which influence errors, the causal information that is collected may not be very useful in developing remedial strategies. General information in areas such as severity, work control aspects and the technical details of the incident will be required in all data collection systems. However, in almost all cases a structured process for causal analysis is lacking. Some of the requirements for causal analysis are set out in the following sections. 

As mentioned earlier some measures will be chosen because improvements in these areas were part of the project justification. It is most likely that these will be efficiency measures. Calculation of these measures generally requires analysis of data or specific data collection exercises. There is a relatively high cost associated with preparing these measures so they should be used prudently. In choosing efficiency measures, you should use only those where you have comparative data about the current management systems. For example, if there is no information on the number of hours dedicated to PSM and ESH, don t use this to try to demonstrate the improvement in efficiency. 

Chapter 8—Supplemental References A collection of references that describe data collection, analysis, and application techniques but, in general, do not contain reliability data. 

WASH-1400 is a fundamental document for PRA methodology. The data appendixes contain a great deal of useful information on methods of data assessment. A large number of sources for data are considered, and very general failure rate estimates will produce only gross approximations. Since the advent of data collection schemes across and within plants, the WASH-1400 data are solely useful as a constituent to a data aggregation process or as widely bounded figures that provide a basis for comparison. 

Let us assume that we have collected a set of calibration data (X, Y), where the matrix X (nxp) contains the p > 1 predictor variables (columns) measured for each of n samples (rows). The data matrix Y (nxq) contains the q variables which depend on the X-data. The general model in calibration reads... 

Models are generally built from either fundamental knowledge about a system or empirical data collected from a system. Models based on fundamental knowledge attempt to directly predict actual plant behavior. Therefore, they can be especially useful for those operating situations that have not been previously observed. However, accurate fundamental models are... 

---