Progressing errors

The number of neurons to be used in the input/output layer are based on the number of input/output variables to be considered in the model. However, no algorithms are available for selecting a network structure or the number of hidden nodes. Zurada [16] has discussed several heuristic based techniques for this purpose. One hidden layer is more than sufficient for most problems. The number of neurons in the hidden layer neuron was selected by a trial-and-error procedure by monitoring the sum-of-squared error progression of the validation data set used during training. Details about this proce-... 

During the selection of the number of hidden layer neurons, the desired tolerance should also be considered. In general, a tight tolerance requires that the selected network be trained with fewer hidden neurons. As mentioned earlier, cross-validation during training can be used to monitor the error progression, which subsequently serves as a guideline in the selection of the hidden layer neurons. 

Figure 16 Root-mean-squared error progression plot for Fletcher nonlinear optimization and back-propagation algorithms during training.

Fisher polynomials can be used only within the T range for which they were created. Extrapolation beyond the T limits of validity normally implies substantial error progression in high-F entropy and enthalpy calculations. For instance, figure 3.4 compares Maier-Kelley, Haas-Fisher, and Berman-Brown polynomials for low albite. As can be seen, the first two interpolants, if extended to high T, definitely exceed the Dulong and Petit limit. The Berman-Brown interpolant also passes this limit, but the bias is less dramatic. 

The derivation of the resonance energies for 27 and 10 reveals that (a) homodesmotic reactions are well suited to compensate for the different electronic effects that hinder the calculation of pure homoconjugated resonance energies, (b) use of a homodesmotic reaction such as 24 requires the inclusion of many reference compounds, which of course can lead to considerable error progression in the calculated reaction energy, and (c) the... 

Camino, A. 1989. Preventing human errors progress made in this field. In Proceeding of a Conference on Human Reliability in Nuclear Power. London. 

Much progress has been made ia understanding how to create and use catalysts, but the design and preparation of practical catalysts stUl rehes on a substantial amount of art that is, the appHcation of known facts and iatuition to trial and error methods. General principles are described ia a number of texts (18—21). Very few completely new catalyst systems have been designed from first principles or completely theoretical considerations. New catalysts are much more likely to be discovered as a result of an adventitious observation than designed by iatent. 

As microprocessor-based controls displaced hardwired electronic and pneumatic controls, the impac t on plant safety has definitely been positive. When automated procedures replace manual procedures for routine operations, the probability of human errors leading to hazardous situations is lowered. The enhanced capability for presenting information to the process operators in a timely manner and in the most meaningful form increases the operator s awareness of the current conditions in the process. Process operators are expected to exercise due diligence in the supervision of the process, and timely recognition of an abnormal situation reduces the likelihood that the situation will progress to the hazardous state. Figure 8-88 depicts the layers of safety protection in a typical chemical jdant. 

Measurement Error Uncertainty in the interpretation of unit performance results from statistical errors in the measurements, low levels of process understanding, and differences in unit and modeled performance (Frey, H.C., and E. Rubin, Evaluate Uncertainties in Advanced Process Technologies, Chemical Engineering Progress, May 1992, 63-70). It is difficult to determine which measurements will provide the most insight into unit performance. A necessary first step is the understanding of the measurement errors hkely to be encountered. 

An example adapted from Verneuil, et al. (Verneuil, V.S., P. Yan, and F. Madron, Banish Bad Plant Data, Chemical Engineeiing Progress, October 1992, 45-51) shows the impact of flow measurement error on misinterpretation of the unit operation. The success in interpreting and ultimately improving unit performance depends upon the uncertainty in the measurements. In Fig. 30-14, the materi balance constraint would indicate that S3 = —7, which is unrealistic. However, accounting for the uncertainties in both Si and S9 shows that the value for S3 is —7 28. Without considering uncertainties in the measurements, analysts might conclude that the flows or model contain bias (systematic) error. 

There are two main approaches to its solution. Traditional approach is based on preliminary separation of UGC samples to gaseous and liquid phases and their subsequent analyses [1]. This approach is well-developed and it allows obtaining quite precise results being used properly. However, this method is relatively complicated. Multi-stage procedure is a source of potential errors, then, it makes the analyses quite time consuming. More progressive approach is based on the direct analysis of the pressurized UGC samples. In both cases the determination of heavy hydrocarbons (up to C ) is made by capillary gas chromatography. 

Quality goals should be set for all aspects of the operation, and there should be evidence that progress is being made towards those goals. There should be a system for error cause removal. 

Late drawings/technical queries/ tooling delays/errors/shop overload/ out of sequence/overtime/concessions/ shortages/excess work in progress/excess inventory/poor supplier quality/penalty clauses/ lost sales/poor management... 

Bridges, W. G., Kirkman, J. Q., Lorenzo, D. K. (1994). Include Human Errors in Process Hazard Analyses. Chemical Engineering Progress, May. 

Geyer, T. A., Bellamy, L. J., Astley, J. A., Hurst, N. W. (1990). Prevent Pipe Failures Due to Human Errors. Chemical Engineering Progress, November. 

---