The article presents results of the influence of the GMDH (Group Method of Data Handling) neural network input data preparation method on the results of predicting corrections for the Polish timescale UTC(PL). Prediction of corrections was carried out using two methods, time series analysis and regression. As appropriate to these methods, the input data was prepared based on two time series, ts1 and ts2. The implemented research concerned the designation of the prediction errors on certain days of the forecast and the influence of the quantity of data on the prediction error. The obtained results indicate that in the case of the GMDH neural network the best quality of forecasting for UTC(PL) can be obtained using the time-series analysis method. The prediction errors obtained did not exceed the value of ± 8 ns, which confirms the possibility of maintaining the Polish timescale at a high level of compliance with the UTC.
Statistical Process Control (SPC) based on the well known Shewhart control charts, is widely used in contemporary manufacturing
industry, including many foundries. However, the classic SPC methods require that the measured quantities, e.g. process or product
parameters, are not auto-correlated, i.e. their current values do not depend on the preceding ones. For the processes which do not obey this
assumption the Special Cause Control (SCC) charts were proposed, utilizing the residual data obtained from the time-series analysis. In the
present paper the results of application of SCC charts to a green sand processing system are presented. The tests, made on real industrial
data collected in a big iron foundry, were aimed at the comparison of occurrences of out-of-control signals detected in the original data
with those appeared in the residual data. It was found that application of the SCC charts reduces numbers of the signals in almost all cases
It is concluded that it can be helpful in avoiding false signals, i.e. resulting from predictable factors.
The aim of the paper was an attempt at applying the time-series analysis to the control of the melting process of grey cast iron in production conditions. The production data were collected in one of Polish foundries in the form of spectrometer printouts. The quality of the alloy was controlled by its chemical composition in about 0.5 hour time intervals. The procedure of preparation of the industrial data is presented, including OCR-based method of transformation to the electronic numerical format as well as generation of records related to particular weekdays. The computations for time-series analysis were made using the author’s own software having a wide range of capabilities, including detection of important periodicity in data as well as regression modeling of the residual data, i.e. the values obtained after subtraction of general trend, trend of variability amplitude and the periodical component. The most interesting results of the analysis include: significant 2-measurements periodicity of percentages of all components, significance 7-day periodicity of silicon content measured at the end of a day and the relatively good prediction accuracy obtained without modeling of residual data for various types of expected values. Some practical conclusions have been formulated, related to possible improvements in the melting process control procedures as well as more general tips concerning applications of time-series analysis in foundry production.
The purpose of this paper was testing suitability of the time-series analysis for quality control of the continuous steel casting process in
production conditions. The analysis was carried out on industrial data collected in one of Polish steel plants. The production data
concerned defective fractions of billets obtained in the process. The procedure of the industrial data preparation is presented. The
computations for the time-series analysis were carried out in two ways, both using the authors’ own software. The first one, applied to the
real numbers type of the data has a wide range of capabilities, including not only prediction of the future values but also detection of
important periodicity in data. In the second approach the data were assumed in a binary (categorical) form, i.e. the every heat(melt) was
labeled as ‘Good’ or ‘Defective’. The naïve Bayesian classifier was used for predicting the successive values. The most interesting results
of the analysis include good prediction accuracies obtained by both methodologies, the crucial influence of the last preceding point on the
predicted result for the real data time-series analysis as well as obtaining an information about the type of misclassification for binary data.
The possibility of prediction of the future values can be used by engineering or operational staff with an expert knowledge to decrease
fraction of defective products by taking appropriate action when the forthcoming period is identified as critical.