Performance measurement techniques requirements

The performance measure data requires analysis. The analysis techniques are tied to the performance measures and the performance goals. The format of the performance measures indicates the types of data analysis techniques that can or cannot be used (see Chapter 3). The performance goals are also tied to the analysis techniques. If the performance goal is the reduction of back injuries over a period of time, the analysis technique used should be capable of identifying such a decrease, if it does exist. (See Chapters 3 and 4 for a discussion of analysis techniques.)... 

Sampling saturated reservoirs with this technique requires special care to attempt to obtain a representative sample, and in any case when the flowing bottom hole pressure is lower than the bubble point, the validity of the sample remains doubtful. Multiple subsurface samples are usually taken by running sample bombs in tandem or performing repeat runs. The samples are checked for consistency by measuring their bubble point pressure at surface temperature. Samples whose bubble point lie within 2% of each other may be sent to the laboratory for PVT analysis. 

Some measures of PSM and ESH performance are easy to identify, establish and track. These include accident rates, effluent tonnages and composition and number of days lost to illness. Almost all of these traditional performance measures are end-of-pipe that is, they measure the output of the management system and allow corrective action only after a failure has occurred. The ideal measurement system identifies potential problems ahead of actual failure allowing corrective action to be taken. This requires using techniques such as audits and hazard assessments. 

Comparison of the success of different classification methods requires a realistic estimation of performance measures for classification, like misclassification rates (% wrong) or predictive abilities (% correct) for new cases (Section 5.7)—together with an estimation of the spread of these measures. Because the number of objects with known class memberships is usually small, appropriate resampling techniques like repeated double CV or bootstrap (Section 4.2) have to be applied. A difficulty is that performance measures from regression (based on residuals) are often used in the development of classifiers but not misclassification rates. 

Source and detector selection are interrelated, where the output of the source is matched to the sensitivity range of the detector. However, the exact nature of the source is also dependent on the type of sample(s) under consideration, the intended optical geometry, the type of measurement technique, and the final desired performance. The bottom line is that adequate source energy must reach the detector to provide a signal-to-noise performance consistent with the required precision and reproducibility of the measurement. This assumes that the detector is well matched, optically and performance-wise, and is also capable of providing the desired performance. 

Determination of moisture content is described in Section 8, ANALYTICAL PROCEDURES, ETC", where are listed numerous US Military Specifications. These specs contain also a brief description of required physical tests. Before describing the specification requirement tests, a resume is given of "Measuring Techniques" of fuze explosive components, as discussed in NOLTR 1111(1952), pp 9-1 to 9-56 (Ref 11) The mea surement of fuze component performance consists mainly of a determination of (a) the input characteristics, and (b) the output characteristics... 

Robustness. Examples of typical possible sources of variation in automated methods are homogenization speed, homogenization time, age of sample, accuracy of solvent dispense, and temperature variation. If all studies described in the method development have been performed, the robustness of the sample preparation has been demonstrated and does not require additional testing. Parameters in relation to the measurement technique may need to be considered and are covered in the relevant chapter. 

Approaches based on parameter estimation assume that the faults lead to detectable changes of physical system parameters. Therefore, FD can be pursued by comparing the estimates of the system parameters with the nominal values obtained in healthy conditions. The operative procedure, originally established in [23], requires an accurate model of the process (including a reliable nominal estimate of the model parameters) and the determination of the relationship between model parameters and physical parameters. Then, an online estimation of the process parameters is performed on the basis of available measures. This approach, of course, might reveal ineffective when the parameter estimation technique requires solution to a nonlinear optimization problem. In such cases, reduced-order or simplified mathematical models may be used, at the expense of accuracy and robustness. Moreover, fault isolation could be difficult to achieve, since model parameters cannot always be converted back into corresponding physical parameters, and thus the influence of each physical parameters on the residuals could not be easily determined. 

Another source of deviations to the ideal behavior is the smoothness of the channel surface which, in reality, is hardly perfect. The surface quality affects substantially both retention and zone dispersion. Smith et al. [223] illustrated this fact experimentally for Th-FFF. Dilks et al. [458] studied experimentally the effect of sample injection and flow pattern on the zone shape inside the channel by performing measurements in a transparent channel and photographing the colored zones formed under various conditions of injection, flow, and geometric channel irregularities. One important result was that even apparently minor channel irregularities can give rise to considerable distortion of the zone formed. In Fl-FFF, the membrane is the critical parameter as ideally it has to fulfill the requirements of pressure and mechanical stability, even surface, uniform pore size, inert behavior with respect to solvent and samples and sufficient counter pressure to achieve smooth and uniform flow rates. A membrane fulfilling all the above requirements does not exist so that the choice of a membrane for Fl-FFF is always a compromise and depends on the analytical problem. In addition, for all other FFF techniques, the surface quality, in particular the smoothness of the channel accumulation wall, substantially affects both retention and zone dispersion. Smith et al. [223] illustrated this fact experimentally for Th-FFF. 

Table I compares results achieved when seven variables that may affect the performance of a particular catalyst were tested one-at-a-time with results from a statistical design (fractional factorial) approach. In this comparison, a shift in measured performance is assumed to be real if it represents at least twice the uncertainty of the measuring technique. The one-at-a-time strategy, still prevalent among many catalyst researchers, requires 48 experiments to determine with 95% confidence which variables significantly impact catalyst performance. Whereas, with the fractional factorial approach, this same information was obtained in only 16 experiments with a 98.5% confidence level. The fractional factorial approach also shows possible interactions among the variables the classical one-at-a-time approach does not.

Many different techniques have been developed for the measurement of contact angles 17.8). Of these, the three most useful methods are the Wilhelnty technique, the technique of capillary rise at a vertical plate, and the drop shape methods. These techniques require the solid surface to be flat and smooth. Direct measurement of contact angles on fibers (of uniform thickness) can also be performed using the Wilhelmy technique. For nonflat surfaces or particles, indirect methods such as capillary penetration into columns of powders, sedi-... 

To show activation of NF-kB, electrophoretic mobility shift assays are typically performed to look at the specific binding of activated NF-kB to DNA. This technique requires relatively large numbers of cells, is laborious, is not performed in intact cells, and is subject to artifacts. Another typical cellular assay measures translational regulation of gene reporter constructs in transfected cells occurring hours after cellular aetivation. 

---