Integrated structure selection and parameter optimisation for eng-genes neural models

The eng-genes concept involves the use of fundamental known system functions as activation functions in a neural model to create a 'grey-box' neural network. One of the main issues in eng-genes modelling is to produce a parsimonious model given a model construction criterion. The challenges are that (1) the eng-genes model in most cases is a heterogenous network consisting of more than one type of nonlinear basis functions, and each basis function may have different set of parameters to be optimised; (2) the number of hidden nodes has to be chosen based on a model selection criterion. This is a mixed integer hard problem and this paper investigates the use of a forward selection algorithm to optimise both the network structure and the parameters of the system-derived activation functions. Results are included from case studies performed on a simulated continuously stirred tank reactor process, and using actual data from a pH neutralisation plant. The resulting eng-genes networks demonstrate superior simulation performance and transparency over a range of network sizes when compared to conventional neural models. (c) 2007 Elsevier B.V. All rights reserved.

How much gas can we get from grass?

Grass biomethane has been shown to be a sustainable gaseous transport biofuel, with a good energy balance, and significant potential for economic viability. Of issue for the designer is the variation in characteristics of the grass depending on location of source, time of cut and species. Further confusion arises from the biomethane potential tests (BMP) which have a tendency to give varying results. This paper has dual ambitions. One of these is to highlight the various results for biomethane potential that may be obtained from the same grass silage. The results indicated that methane potential from the same grass silage varied from 350 to 493 L CH4 kg−1 VS added for three different BMP procedures. The second ambition is to attempt to compare two distinct digestion systems again using the same grass: a two stage continuously stirred tank reactor (CSTR); and a sequentially fed leach bed reactor connected to an upflow anaerobic sludge blanket (SLBR–UASB). The two engineered systems were designed, fabricated, commissioned and operated at small pilot scale until stable optimal operating conditions were reached. The CSTR system achieved 451 L CH4 kg−1 VS added over a 50 day retention period. The SLBR–UASB achieved 341 L CH4 kg−1 VS added at a 30 day retention time.

Use of a rotating disc reactor to investigate the heterogeneously catalysed oxidation of cinnamyl alcohol in toluene and ionic liquids

To evaluate the effect of mass transfer limitations in the three-phase oxidation of cinnamyl alcohol carried out in toluene and an ionic liquid (1-butyl-3-methyl-imidazolium bis(trifluoromethylsulphonyl)imide), studies have been performed in a rotating disc reactor and compared with those carried out in a stirred tank reactor where mass transfer effects are considered negligible. High catalyst efficiencies are found in the stirred tank reactor with the use of both ionic liquid and toluene, although there is a decrease in rate for the ionic liquid reactions. In contrast, internal pore diffusion limits the reaction in both solvents in the rotating disc reactor. This mass transfer resistance reduces the problem of overoxidation of the metal surface when the reaction is carried out in toluene, leading to significantly higher rates of reaction than expected, although at the cost of decreased selectivity.

Comparison of mass transfer effects in the heterogeneously catalysed hydrogenation of phenyl acetylene in heptane and an ionic liquid

The selective heterogeneous catalytic reduction of phenyl acetylene to styrene over palladium supported on calcium carbonate is reported in both an ionic liquid and a molecular solvent. By using a rotating disc reactor in conjunction with results from a stirred tank reactor it is possible, for the first time, to disentangle the mass transfer contributions in the ionic liquid system. For both heptane and 1-butyl-3-methyl imidazolium bis{(trifluoromethyl)sulfonyl}imide, the reaction in the rotating disc reactor is dominated by reaction in the entrained film on the disc compared with very limited reaction in the bulk liquid. The lower reaction rate obtained in the ionic liquid compared with the organic solvent is shown to be due to the slow transport of the hydrogen dissolved in the liquid. It is clear from the results presented herein that, although the hydrodynamics of similar reactors used for biological treatment of wastewater are well understood, on using a more viscous fluid and higher rotation speeds necessary for fine chemical catalysis these simple relationships breakdown.

Constructing networks of continuous-time velocity-based models

A conventional local model (LM) network consists of a set of affine local models blended together using appropriate weighting functions. Such networks have poor interpretability since the dynamics of the blended network are only weakly related to the underlying local models. In contrast, velocity-based LM networks employ strictly linear local models to provide a transparent framework for nonlinear modelling in which the global dynamics are a simple linear combination of the local model dynamics. A novel approach for constructing continuous-time velocity-based networks from plant data is presented. Key issues including continuous-time parameter estimation, correct realisation of the velocity-based local models and avoidance of the input derivative are all addressed. Application results are reported for the highly nonlinear simulated continuous stirred tank reactor process.

Chemometric techniques in multivariate statistical modelling of process plant

The techniques of principal component analysis (PCA) and partial least squares (PLS) are introduced from the point of view of providing a multivariate statistical method for modelling process plants. The advantages and limitations of PCA and PLS are discussed from the perspective of the type of data and problems that might be encountered in this application area. These concepts are exemplified by two case studies dealing first with data from a continuous stirred tank reactor (CSTR) simulation and second a literature source describing a low-density polyethylene (LDPE) reactor simulation.

Mechanochemical synthesis of TiO2/NiFe2O4 magnetic catalysts for operation under RF field

Composite NiFe2O4–TiO2 magnetic catalysts were prepared by mechanochemical synthesis from a mixture of titania supported nickel ferrite nanoparticles and P25 titania (Evonic). The former provides fast and efficient heating under radiofrequency field, while the latter serves as an active catalyst or catalyst support. The highest heating rate was observed over a catalyst prepared for a milling time of 30 min. The catalytic activity was measured over the sulfated composite catalysts in the condensation of aniline and 3-phenylbutyric acid in a stirred tank reactor and in a continuous RF heated flow reactor in the 140–170 °C range. The product yield of 47% was obtained over the sulfated P25 titania catalyst in the flow reactor.

Remediation of oily wastewater from an interceptor tank using a novel photocatalytic drum reactor

A novel photocatalytic reactor has been developed to remediate oily wastewaters. In the first instance degradation rates of model organic compounds, methylene blue (MB) and 4-c hlorophenol (4-CP) were determined. The experimental set-up investigated a 1:10 w/v catalyst to organic solution volume, 30 g catalyst, 300 mls MB (10 μM) or 4-CP (100 μM). The catalyst investigated was a pellet catalyst to improve separation of the remediated volume from the catalyst following treatment. MB concentration decreased by 93% after 15 mins irradiation whilst 4-CP concentration decreased by 94% following 90 mins irradiation. Oily waste water (OWW) from an interceptor tank typically containing diesel oils was obtained from Sureclean, an environmental clean-up company. The OWW was treated using the same conditions as MB and 4-CP, the model organic compounds. Levels of total organic carbon (TOC) and total petroleum hydrocarbon (TPH) were used to monitor the efficacy of the photocatalytic reactor. TOC reduced by 45% following two 90 mins treatment cycles. TPH reduced by 45% following 90 mins irradiation and by a further 25% during a second stage of treatment. This reactor can be used as a polishing technique assembled within a wastewater treatment plant. Allowing for more than one pass through the reactor improves its efficiency. 

Thin-zone TAP reactor as a basis of �state-by-state transient screening

Thin-zone TAP reactor is presented as a basis of the new state-by-state transient screening approach which has been proposed by the authors for non-steady-state kinetic characterization of industrial catalysts. The general thin-zone TAP reactor model is described, and its mathematical status is justified analytically. It is shown that this model provides high enough accuracy to be applicable in the wide conversion interval (up to 90%), which is an important advantage of this approach compared with the traditional differential reactor.

Characterization of a dielectric barrier discharge operating in an open reactor with flowing helium

A dielectric barrier discharge (DBD) generated by flowing helium between the parallel-plate electrodes of an open air reactor has been characterized using time resolved optical and electrical measurements. A sinusoidal voltage of up to 5 kV (peak to peak) of frequencies from 3 to 50 kHz has been applied to the discharge electrodes. The helium flow rate is varied up to 10 litre min(-1). The adjustment of flow rate allows the creation of uniform DBDs with optimized input power equal to 120 +/- 10 mW cm(-3). At flow rates from 4 to 6 litre min(-1) a uniform DBD is obtained. The maxima in the line intensities of N-2(+) and helium at 391.4 nm and 706.5 nm, respectively, 2 under those conditions indicate the importance of helium metastables and He-2(+) in sustaining such a discharge. The power efficiency and discharge 2 current show maxima when the DBD In He/air is uniform. The gas temperature during the discharge has been measured as 360 +/- 20 K.