Improved Nonlinear PCA for Process Monitoring Using Support Vector Data Description

Nonlinear principal component analysis (PCA) based on neural networks has drawn significant attention as a monitoring tool for complex nonlinear processes, but there remains a difficulty with determining the optimal network topology. This paper exploits the advantages of the Fast Recursive Algorithm, where the number of nodes, the location of centres, and the weights between the hidden layer and the output layer can be identified simultaneously for the radial basis function (RBF) networks. The topology problem for the nonlinear PCA based on neural networks can thus be solved. Another problem with nonlinear PCA is that the derived nonlinear scores may not be statistically independent or follow a simple parametric distribution. This hinders its applications in process monitoring since the simplicity of applying predetermined probability distribution functions is lost. This paper proposes the use of a support vector data description and shows that transforming the nonlinear principal components into a feature space allows a simple statistical inference. Results from both simulated and industrial data confirm the efficacy of the proposed method for solving nonlinear principal component problems, compared with linear PCA and kernel PCA.

The inferential monitoring of the screw disturbance torque to predict process fluctuations in polymer extrusion

Polymer extrusion is one of the major methods of processing polymer materials and advanced process monitoring is important to ensure good product quality. However, commonly used process monitoring devices, e.g. temperature and pressure sensors, are limited in providing information on process dynamics inside an extruder barrel. Screw load torque dynamics, which may occur due to changes in solids conveying, melting, mixing, melt conveying, etc., are believed to be a useful indicator of process fluctuations inside the extruder barrel. However, practical measurement of the screw load torque is difficult to achieve. In this work, inferential monitoring of the screw load torque signal in an extruder was shown to be possible by monitoring the motor current (armature and/or field) and simulation studies were used to check the accuracy of the proposed method. The ability of this signal to aid identification and diagnosis of process issues was explored through an experimental investigation. Power spectral density and wavelet frequency analysis were implemented together with a covariance analysis. It was shown that the torque signal is dominated by the solid friction in the extruder and hence it did not correlate well with melting fluctuations. However, it is useful for online identification of solids conveying issues.

Monitoring and population estimation of the European badger (Meles meles) in Northern Ireland

The estimation of animal abundance has a central role in wildlife management and research, including the role of badgers Meles meles in bovine tuberculosis transmission to cattle. This is the first study to examine temporal change in the badger population of Northern Ireland over amedium- to long-term time frame of 14-18 years by repeating a national survey first conducted during 1990-1993. A total of 212 1-km2 squares were surveyed during 2007-2008 and the number, type and activity of setts therein recorded. Badgers were widespread with 75% of squares containing at least one sett. The mean density of activemain setts,which was equivalent to badger social group density, was 0.56 (95%CI: 0.46-0.67) active main setts per km2 during 2007-2008. Social group density varied significantly among landclass groups and counties. The total number of social groups was estimated at 7,600 (95%CI: 6,200-9,000) and, not withstanding probable sources of error in estimating social group size, the total abundance of badgers was estimated to be 34,100 (95% CI: 26,200-42,000). There was no significant change in the badger population from that recorded during 1990-1993. A resource selection model provided a relative probability of sett construction at a spatial scale of 25m. Sett locations were negatively associated with elevation and positively associated with slope, aspect, soil sand content, the presence of cover, and the area of improved grassland and arable agriculture within 300 m.

Technical Note: Assessing a 24/7 solution for monitoring water quality loads in small river catchments

Quantifying nutrient and sediment loads in catchments is dif?cult owing to diffuse controls related to storm hydrology. Coarse sampling and interpolation methods are prone to very high uncertainties due to under-representation of high discharge, short duration events. Additionally, important low-?ow processes such as diurnal signals linked to point source impacts are missed. Here we demonstrate a solution based on a time-integrated approach to sampling with a standard 24 bottle autosampler con?gured to take a sample every 7 h over a week according to a Plynlimon design. This is evaluated with a number of other sampling strategies using a two-year dataset of sub-hourly discharge and phosphorus concentration data. The 24/7 solution is shown to be among the least uncertain in estimating load (inter-quartile range: 96% to 110% of actual load in year 1 and 97% to 104% in year 2) due to the increased frequency raising the probability of sampling storm events and point source signals. The 24/7 solution would appear to be most parsimonious in terms of data coverage and certainty, process signal representation, potential laboratory commitment, technology requirements and the ability to be widely deployed in complex catchments.