Variable test inputs

The test schematic is sketched in Figure 21.3, which describes the evolution of some of the variable test inputs and of the ceU voltage from the test steps of start-up and conditioning to the beginning of the on/off cycle measurements [9]. 

I 27 Quality Assurance for Characterizing Low-Temperature Fuel Cells Table 21.7 Variable test inputs. 

Table 21.9 Variable test inputs the static inputs are given in Table 21.6.

The input disturbance for a pulse test begins and ends at the same value. In terms of perturbation variables, the input m ) is initially zero and is returned to zero after some time, ... 

The aim of the Greek test site was to study the possibility for hydrogen to become an alternative product for wind-park developers, in case electricity transmission lines are saturated, studying at the same time the performance of hydrogen production and storage technologies under variable power input. 

To develop a test plan the program only requires information on the independent variables. After input of the variables (experimental parameters) and their levels, the program calculates those points in a multidimensional space that provide the best predictions in a subsequent modeling process. The input levels of a variable determine the range of values, which should have been determined in preliminary investigations and catalyst screening. In general at least 3-5 levels per variable should be chosen. 

Although this is an old and conceptually straightforward idea, it has not been widely used (except in some recent studies of chaotic dynamics of autonomous systems, where no input variable exists) because several important practical issues must be addressed in its actual implementation, for example, the selection of the appropriate coordinate variables (embedding space) and the impracticality of representation in high-dimensional spaces. If a low-dimensional embedding space can be found for the system under study, this approach can be very powerful in yielding models of strongly nonlinear systems. Secondary practical issues are the choice of an effective test input and the accuracy of the obtained results in the presence of extraneous noise. 

A pulse generator that can simulate detector pulses is perhaps not essential but is certainly desirable. It should provide pulses with variable rise time (perhaps 10 to 500 ns) and variable fall-time (perhaps 10 to 500 j,s). The output from the pulser is put into the TEST INPUT of the preamplifier. It allows a distinction to be made between detector problems and pulse processing problems. Some laboratories routinely use a pulser for dead time and random summing correction purposes. For systems with TRP preamplifiers, a simpler pulser providing square pulses would be adequate. 

At this point one has a compartmental model structure, a description of the measurement error, and a numerical value of the parameters together with the precision with which they can be estimated. It is now appropriate to address the optimal experiment design issue. The rationale of optimal experiment design is to act on design variables such as number of test input and outputs, form of test inputs, number of samples and sampling schedule, and measurement errors so as to maximize, according to some criterion, the precision with which the compartmental model parameters can be estimated [DiStefano, 1981 Carson et al, 1983 Landaw and DiStefano, 1984 Walter and Pronzato, 1990]. 

Multiple linear regression analysis is a widely used method, in this case assuming that a linear relationship exists between solubility and the 18 input variables. The multilinear regression analy.si.s was performed by the SPSS program [30]. The training set was used to build a model, and the test set was used for the prediction of solubility. The MLRA model provided, for the training set, a correlation coefficient r = 0.92 and a standard deviation of, s = 0,78, and for the test set, r = 0.94 and s = 0.68. 

Laboratory experiments using rodents, or the use of gas analysis, tend to be confused by the dominant variable of fuel—air ratio as well as important effects of burning configuration, heat input, equipment design, and toxicity criteria used, ie, death vs incapacitation, time to death, lethal concentration, etc (154,155). Some comparisons of polyurethane foam combustion toxicity with and without phosphoms flame retardants show no consistent positive or negative effect. Moreover, data from small-scale tests have doubtful relevance to real fine ha2ards. 

In principle, the step-response coefficients can be determined from the output response to a step change in the input. A typical response to a unit step change in input u is shown in Fig. 8-43. The step response coefficients are simply the values of the output variable at the samphng instants, after the initial value y(0) has been subtracted. Theoretically, they can be determined from a single-step response, but, in practice, a number of bump tests are required to compensate for unanticipated disturbances, process nonhnearities, and noisy measurements. 

This construction works because G is free and we are dealing with free interpretations. When we come to a test we can nondeterministically select either path - seme free interpretation will take either path. The final output under any free interpretation must be either f. -.f x for some input variable x or else for some constant o. In the first case, G guesses this will... 

The transformation t we saw at the end of the last section, which changes liberal schemes into free schemes, is such a canonical transformation. The corresponding canonical class of schemes is the class of schemes such that tests are applied initially on the input variables and are applied after assignment statements on the program variables involved, and at no other time. This transformation t is clearly recursive and equivalence preserving. The class of free schemes is not a canonical form class, since, as we saw, there are schemes not strongly equivalent to any free scheme. 

Thus the computation record of a computation contains the successive listings of the values of the variables - it is the listing of computation states minus the name of the instruction executed at tine i. The test added computation record contains this list of values plus the formal record of tests applied. If test T is applied at step i to, say, values 1 and 5, the test added computation record contains for step i the formal statement "T(l,5)", identifying the particular test applied and the values to which it is applied. The computation record or test added computation record relative to some subset of variables considers only values of those variables and tests applied to those variables. It is useful when comparing programs with different program variables but the same input and output variables. 

We assume that our program scheme P is written in linear sequential form with the first instruction in line 1. We use the convention that in a test statement the first transfer address is the transfer for positive (TRUE) outcome of the test. We can assume that P has input variables x, ...,x disjoint from program variables, ...,ym and an output variable z which is neither an input for a program variable and only appears in a halting sequence z + u STOP. 

---