Joint order and parameter estimation of multivariate autoregressive models using multi-model partitioning theory

In this paper the multi-model partitioning theory is used for simultaneous order and parameter estimation of multivariate autoregressive models. Simulation experiments show that the proposed method successfully selects the correct model order and estimates the parameters accurately, in very few steps, even with a small sample size. They also show that the proposed method performs equally well when the complexity of the model is increased. The results are compared to those obtained using well-established order selection criteria. Finally, it is shown that the method is also successful in tracking model order changes, in real time.     

Multi-model minimum variance control

A new method for simultaneously selecting the order and identifying the parameters of an ARX model and the control strategy design has been developed. The method is based on the reformulation of the problem in the standard state space form and the subsequent implementation of a bank of minimum variance controllers, each fitting a different order model. Thus, the problem is reduced to selecting the true model among a set of candidate models, using the well-known multi-model partitioning theory, for general (not necessarily Gaussian) data pdf's. Thus, the cumulative control is the average of the model-conditional minimum variance controls, weighted by the respective a posteriori probability that each particular model is the true model. Simulation experiments indicate that the proposed method is 100% successful in selecting the correct model order and that it accurately identifies the model parameters, in a sufficiently small number of iterations. Furthermore, the method is insensitive in variations of the used filters' variance, and it is adaptive, in the sense that it has the ability of successfully tracking changes in the model structure, in real time. The proposed algorithm lends itself to parallel and VLSI implementations. © 1998 John Wiley & Sons, Ltd.     

Generation and Deposition of Organic Nanoparticles in a Liquid Ring Pump

A new method for the production of organic nanoparticles or nanosuspensions in a liquid ring vacuum pump is presented. A hot carrier gas charged under vacuum conditions with poorly water‐soluble organic substances is compressed and simultaneously cooled by the partially evaporating sealant water of the liquid ring vacuum pump. During compression and quenching, the gas‐vapor mixture becomes highly supersaturated and aerosol formation initiated by homogeneous nucleation occurs in the compression chambers between the impeller blades of the pump. The thus generated gas‐borne droplets or solid particles are mainly trapped and enriched in the sealant ring liquid, which is discharged, separated from the carrier gas in an external separator and returned to the pump. Experiments yielded aqueous nanosuspensions of insoluble organic substances with typical particle sizes between 50 and 300 nm, as well as foam in the external gas‐liquid separator. The foam showed stable structures due to incorporated solid nanoparticles.     

A design-and-play approach to accessible user interface development in Ambient Intelligence environments

User interface development in Ambient Intelligence (AmI) environments is anticipated to be a particularly complex and programming intensive endeavor. Additionally, AmI environments should ensure accessibility and usability of interactive technologies by users with different characteristics and requirements in a mainstream fashion. Therefore, appropriate user interface development methods and tools are required, capable of both reducing development efforts and ‘injecting’ accessibility issues into AmI applications from the early design stages. This paper introduces two tools, named AmIDesigner and AmIPlayer, which have been specifically developed to address the above challenges through automatic generation of accessible Graphical User Interfaces in AmI environments. The combination of these two tools offers a simple and rapid design-and-play approach, and the running user interfaces produced integrate non-visual feedback and a scanning mechanism to support accessibility. AmIDesigner and AmIPlayer have been evaluated to assess their usability by designers, and have been put to practice in the redevelopment of a light control application in a smart environment as a case study demonstrating the viability of the design-and-play approach. The results confirm the usefulness and usability of the tools themselves. Overall, the proposed approach has the potential to contribute significantly to the development, up-take and user acceptance of AmI technologies in the home environment.     

A region dissimilarity relation that combines feature-space and spatial information for color image segmentation.

This paper proposes a methodology that incorporates principles from cluster analysis and graph representation to achieve efficient image segmentation results. More specifically, a feature-based, inter-region dissimilarity relation is considered here in order to determine the dissimilarity matrix in a graph-based segmentation scheme. The calculation of the dissimilarity function between adjacent elementary image regions is based on the proximity of each region's feature vector to the main clusters that are formed by the image samples in the feature space. In contrast to typical segmentation approaches of the literature, the global feature space information is included in the spatial graph representation that was derived from the initial Watershed partitioning. A region grouping process is applied next to form the final segmentation results. The proposed approach was also compared to approaches that use feature-based, or spatial information exclusively, to indicate its effectiveness.     

Adaptive algorithms for heading-assisted radar tracking of manoeuvering targets: A comparative study

Three different algorithms based on the multimodel partitioning approach are designed for the problem of tracking a manoeuvering target utilizing aircraft derived data (ADD) measurements of heading. These are subsequently evaluated with respect to their performance in the sense of the mean square estimation error by means of simulations in three realistic tracking scenarios. The extend Kalman filter (EKF) for this problem has been used as a measure of comparison for the performance of all algorithms. The results indicate the improvement in performance obtained when using adaptive techniques to track manoeuvering targets instead of the non-adaptive EKF.     

FORMATION OF HYDROBROMIC AND HYDROCHLORIC ACID AEROSOLS IN WET FLUE GAS CLEANING PROCESSES

The formation and behavior of hydrobromic and hydrochloric acid aerosols in a wet flue gas cleaning pilot plant were investigated. The optical three-wavelength extinction (3-WE) method was used to determine mean aerosol droplet diameters and droplet number concentrations. The experimental data are compared with theoretical results of the simulation tool AerCoDe (aerosol formation in contact devices). Results are presented for a raw gas temperature of 200°C and raw gas concentrations up to 260 mg/m³(STP) for HBr, and 2500 mg/m³ (STP) for HCl. Under these conditions aerosol formation for both species is initiated by heterogeneous nucleation. It is shown that during absorption processes HBr is forming essentially higher supersaturated gas phases in comparison to HCl, resulting in higher droplet number concentrations and smaller droplet sizes. For both species the number concentration is a strong function of the maximum degree of saturation, which corresponds to the classical theory of heterogeneous nucleation. HBr aerosol droplets cannot be detected by the 3-WE method directly after the first stage of the flue gas cleaning plant (quench column) because they are smaller than the detection limit ( ≈ 500 nm) of the 3-WE method for nonabsorbing particles. In this case the droplets are enlarged in a second step, and the number concentration is determined after enlargement. Using the actual number concentration as an input parameter, the mean diameter after the first stage can be calculated with the simulation tool AerCoDe.     

Effect of substituting fresh‐cut perennial ryegrass with fresh‐cut white clover on bovine milk fatty acid profile

BACKGROUND

Including forage legumes in dairy systems can help address increasing environmental/economic concerns about perennial ryegrass monoculture pastures. This work investigated the effect of substituting fresh‐cut grass with increasing quantities of fresh‐cut white clover (WC) on milk fatty acid (FA) profile and transfer efficiency of dietary linoleic (LA) and α‐linolenic (ALNA) acids to milk fat. Three groups of three crossbred dairy cows were used in a 3 × 3 crossover design. Dietary treatments were 0 g kg^?1 WC + 600 g kg^?1 grass, 200 g kg^?1 WC + 400 g kg^?1 grass, and 400 g kg^?1 WC + 200 g kg^?1 grass. All treatments were supplemented with 400 g kg^?1 concentrates on a dry matter basis. Cows had a 19‐day adaptation period to the experimental diet before a 6‐day measurement period in individual tie stalls.

RESULTS

Increasing dietary WC did not affect dry matter intake, milk yield or milk concentrations of fat, protein or lactose. Milk polyunsaturated FA concentrations (total n‐3, total n‐6, LA and ALNA) and transfer efficiency of LA and ALNA were increased with increasing dietary WC supply.

CONCLUSION

Inclusion of WC in pastures may increase concentrations of nutritionally beneficial FA, without influencing milk yield and basic composition, but any implications on human health cannot be drawn. © 2018 The Authors. Journal of the Science of Food and Agriculture published by JohnWiley & Sons Ltd on behalf of Society of Chemical Industry.

     

A simulation tool for aerosol formation during sulphuric acid absorption in a gas cleaning process

A simulation tool has been developed to predict sulphuric acid aerosol formation in typical industrial absorption processes for gas cleaning. The underlying model comprises homogeneous nucleation and the growth of a polydisperse droplet collective under the special circumstances of a gas–liquid contact device where heat and mass transfer processes between the bulk phases take place simultaneously. The model is applied to a hot flue gas (200 °C) with sulphuric acid concentrations between 5 and 100 mg m⁻³ (STP) (STP: standard temperature and pressure). The simulation yields high droplet number concentrations up to 10¹⁶ m⁻³ especially for low gas inlet concentrations of sulphuric acid (5 mg m⁻³ (STP)), and very small droplet sizes in the range 20–100 nm. The droplet number concentrations decrease and the droplet sizes increase with increasing sulphuric acid inlet concentrations. It is shown that small droplets (<20 nm) need relatively high supersaturation for growing. If the saturation in the absorption equipment is not high enough the droplets partially re-evaporate but do not vanish due to the extremely low vapor pressure of concentrated sulphuric acid. The resulting size distributions of the aerosol droplets are not very sensitive with respect to the nucleation model used. This is demonstrated by comparing nucleation models with and without hydrate formation. The new simulation tool allows an estimate of the true sulphuric acid removal efficiency of absorption processes which is often not more than 50% due to aerosol formation. In general, the simulation results enable a deeper insight in the mechanisms of aerosol formation and behavior in absorption processes.