No abstract

Four options for modeling and forecasting time series data containing increasing seasonal variation are discussed, including data trans formations, double seasonal difference models and two kinds of transfer function-type ARIMA models employing seasonal dummy variables. An explanation is given for the typical ARIMA model identification analysis failing to identify double seasonal difference models for this kind of data. A logical process of selecting one option for a particular case is outlined, focusing on issues of linear versus non-linear increasing seasonal variation, and the level of stochastic versus deterministic behavior in a time series. Example models for the various options are presented for six time series, with point forecast and interval forecast comparisons. Interval forecasts from data-transformation models are found to generally be too wide and sometimes illogical in the dependence of their width on the point forecast level. Suspicion that maximum likelihood estimation of ARIMA models leads to excessive indications of unit roots in seasonal movingaverage operators is reported. KEY WORDS ARIMA Box-Jenkins modeling Maximum likelihood estimation Prediction intervals Seasonality Time series forecasting TransformationsFigures 1 through 6 illustrate six monthly time series that contain increasing seasonal variation of some kind. In a time-series modeling context, one of the earlier suggestions for modeling increasing seasonal variation was Winter's (1960) three-parameter multiplicative exponential smoothing model. In the ARIMA modeling context, most modelers have used data transformations, particularly natural logarithms, and built ARIMA models for the transformed data. Another suggestion that has arisen in the ARIMA modeling context is the utilization of double seasonal difference models. When Chatfield and Prothero (1973a) modeled an increasing seasonal variation sales time series using the natural log transformation, three discussants, Akaike (1973), Priestley (1973 and Wilson (1973), suggested that double seasonal difference models of various forms appeared to model the increasing seasonal variation more accurately. In a subsequent reply, Chatfield and Prothero (1973b) said 'we are now of the opinion that the logarithmic transformation should generally be avoided', and

Gas hydrate formation and corrosion within gas pipelines are two major flow assurance problems. Various chemical inhibitors are used to overcome these problems, such as monoethylene glycol (MEG) for gas hydrate control and methyl diethanolamine (MDEA) and film formation corrosion inhibitor (FFCI) for corrosion control. As an economical solution, MEG is regenerated due to the large volume required in the field. MEG regeneration involves thermal exposure by traditional distillation to purify the MEG. During this process, MEG is subjected to thermal exposure and so might be degraded. This study focuses on evaluating six analytical techniques for analyzing the degradation level of various MEG solutions consisting of MDEA and FFCI that were thermally exposed to 135 °C, 165 °C, 185 °C, and 200 °C. The analytical techniques evaluated are pH measurement, electrical conductivity, change in physical characteristics, ion chromatography (IC), high performance liquid chromatography–mass spectroscopy (HPLC-MS), and gas hydrate inhibition performance (using 20 wt % MEG solutions with methane gas at pressure from 50 to 300 bar). Most of the analytical techniques showed good capability, while electrical conductivity showed a poor result for solution without MDEA and IC showed poor results for solution exposed to 135 and 165 °C. The primary aim of this paper is thus to provide the industry with a realistic evaluation of various analytical techniques for the evaluation of degraded MEG solutions and to draw attention to the impact of degraded MEG on gas hydrate and corrosion inhibition as a result of the lack of quality control.

ABSTRACT:The pharmacokinetics and dose proportionality of fexofenadine, a new non-sedating antihistamine, and its enantiomers were characterized after single and multiple-dose administration of its hydrochloride salt. A total of 24 healthy male volunteers (31 98 years) received oral doses of 20, 60, 120 and 240 mg fexofenadine HCl in a randomized, complete four-period cross-over design. Subjects received a single oral dose on day 1, and multiple oral doses every 12 h on day 3 through the morning on day 7. Treatments were separated by a 14-day washout period. Serial blood and urine samples were collected for up to 48 h following the first and last doses of fexofenadine HCl. Fexofenadine and its R(+) and S(− ) enantiomers were analysed in plasma and urine by validated HPLC methods. Fexofenadine pharmacokinetics were linear across the 20-120 mg dose range, but a small disproportionate increase in area under the plasma concentration-time curve (AUC) ( B 25%) was observed following the 240 mg dose. Singledose pharmacokinetics of fexofenadine were predictive of steady-state pharmacokinetics. Urinary elimination of fexofenadine played a minor role (10%) in the disposition of this drug. A 63:37 steady-state ratio of R( +) and S( − ) fexofenadine was observed in plasma. This ratio was essentially constant across time and dose. R( +) and S( −) fexofenadine were eliminated into urine in equal rates and quantities. All doses of fexofenadine HCl were well tolerated after single and multiple-dose administration.