Search

Complex processes often work with multiple operation regions, it is critical to develop effective monitoring approaches to ensure the safety of chemical processes. In this work, a discriminant local consistency Gaussian mixture model (DLCGMM) for multimode process monitoring is proposed for multimode process monitoring by integrating LCGMM with modified local Fisher discriminant analysis (MLFDA). Different from Fisher discriminant analysis (FDA) that aims to discover the global optimal discriminant directions, MLFDA is capable of uncovering multimodality and local structure of the data by exploiting the posterior probabilities of observations within clusters calculated from the results of LCGMM. This may enable MLFDA to capture more meaningful discriminant information hidden in the high-dimensional multimode observations comparing to FDA. Contrary to most existing multimode process monitoring approaches, DLCGMM performs LCGMM and MFLDA iteratively, and the optimal subspaces with multi-Gaussianity and the optimal discriminant projection vectors are simultaneously achieved in the framework of supervised and unsupervised learning. Furthermore, monitoring statistics are established on each cluster that represents a specific operation condition and two global Bayesian inference-based fault monitoring indexes are established by combining with all the monitoring results of all clusters. The efficiency and effectiveness of the proposed method are evaluated through UCI datasets, a simulated multimode model and the Tennessee Eastman benchmark process.

This work reported real-time generation of microbial natural gas (methane) from lignite. Lignite (brown coal) samples and formation water were collected from locations in Australia, Indonesia, and China with no history of natural gas production (zero gas baseline). These samples were set up in microcosms and stimulated for new gas production by amending with essential nutrients. Results show that the Indonesian lignite yielded the highest methane production rate along with the highest enumeration of total bacteria and methanogens. The Australian lignite also produced methane that was associated with a decrease in CO2 composition. The Chinese lignite generated methane, but showed a decrease in methanogen counts, presumably due to microbial community shift. Overall, the results from this preliminary study indicate that application of nutrients to lignite can realize real-time methane production. Further molecular biology analysis is required to determine the constitution and shift in microbial populations during the process of methane generation from lignite.

The development of a bioluminescent sensor organism (Shk1) that was created for assessing wastewater toxicity was reported several years ago. In order to establish a test battery to better characterize wastewater toxicity, additional luminescent sensor organisms were later created. The present study focused on one promising candidate (PM6), a Pseudomonas spp. strain, because of its high level of luminescence compared to that of other newly created organisms. Using a batch toxicity testing protocol, the toxicity of 7 metals and 25 organic compounds was evaluated with the PM6 and Shk1 assays. Results indicated that the toxicity data of the PM6 and the Shk1 assays were correlated, and no assay appeared to be particularly more sensitive to a group of toxicants than the other assay. The results of the PM6 and Shk1 assays were further evaluated by comparing with the results of the Vibrio fischeri luminescence inhibition assay and activated sludge inhibition assays. Data suggested that PM6 and Shk1 more closely represented activated sludge organisms than V. fischeri. The suitability of using PM6 and Shk1 for assessing wastewater toxicity on activated sludge, both individually and in a test battery, was discussed.

Reducing bioassay variability by identifying sources of variation and controlling important parameters in assay protocol was demonstrated in this study. The variability of a bioassay based on a luminescent bacterium was examined as an example. This assay involved the growth of cells, storage at a low temperature, activation, and exposure to a test sample, and the assay response was bioluminescence inhibition. After determining that measurement error was small and negligible, the total assay variability was decomposed in an initial variance components study into between-batch, between-vial, and between-tube variations. Results indicated that between-vial variations accounted for the majority of the total observed variability and that reducing this type of variation would be beneficial. Five parameters in the assay protocol were determined as factors that potentially affected assay variability significantly. A split-plot design was employed to investigate the effects of these factors and some of their interactions on the assay response. One of the five factors, i.e., activation temperature, turned out to have a significant effect. The variance components study was repeated with better control of activation temperature as well as other parameters. Results indicated that the total variability of the bioassay was reduced by approximately 85%.

Two-level factorial experiments were employed in this study for understanding and predicting the toxicity of binary and ternary metal mixtures. Toxicity of metal mixtures with concentrations between the respective EC10 and EC80 values was experimentally measured. Models were fit to the experimental data and the resultant models were of high quality as reflected by R2 (coefficient of determination). Interactions between mixture components were indicated by the existence of statistically significant interaction terms in the models. Toxicity predictions based on the models were compared with observed toxicity for binary and ternary metal mixtures. The models developed did not assume additivity between metals, were simple and interpretable, and gave satisfactory predictions of the toxicity of metal mixtures in aqueous solutions without requiring knowledge on synergism or antagonism.

In aquatic toxicity testing, no single test species responds appropriately to all toxicants. Therefore test batteries consisting of several individual assays are becoming more common. The species comprising a test battery should be representative of the entire system of interest. Each assay should be complementary to other components in the test battery and the test battery should not include redundant tests. We studied the selection of test battery components for the assessment of the toxicity of organic chemicals to activated sludge. The assays considered were the continuous Shk1, Microtox, Polytox, activated sludge respiration inhibition, Nitrosomonas, and Tetrahymena assays. The correlations between the toxicity data obtained from these assays were analyzed by examining the correlation matrix and by principal component analysis. These statistical methods showed that the Nitrosomonas assay should be included in test batteries plus one of the remaining five assays for assessing toxicity of organic compo...

In this study, the use of decision tree in classifying and predicting the aquatic toxicity mechanisms of phenols was investigated. Four mechanisms including polar narcosis, respiratory uncoupling, pro-electrophilicity, and soft electrophilicity were involved. Using molecular descriptors as splitting variables, a three level decision tree with six terminal nodes was obtained. The tree model first separated polar narcosis/pro-electrophilicity from respiratory uncoupling/soft electrophilicity by Elumo in the first level of the tree. In subsequent levels of the tree, polar narcosis was separated from pro-electrophilicity by Nhdon and Ehomo, and respiratory uncoupling was separated from soft electrophilicity by Elumo and log Kow. Validation of the decision tree approach indicated that the overall mechanism prediction accuracy was approximately 85%. The decision tree model had the advantage of easy interpretation.

A two-step classification algorithm was developed for the classification and prediction of the mechanisms of toxic action of phenols. The mechanisms considered were polar narcosis, weak acid respiratory uncoupling, pro-electrophilicity, and soft electrophilicity. The classification algorithm was based on linear discriminant analysis using molecular descriptors as discriminating variables. Polar narcosis/pro-electrophilicity was separated from weak acid respiratory uncoupling/soft electrophilicity in the first step. In the second step, polar narcosis was separated from pro-electrophilicity and weak acid respiratory uncoupling was separated from soft electrophilicity. Different molecular descriptors were involved as discriminating variables at different steps of the classification algorithm. This algorithm classified any given phenol as acting by one and only one of the four mechanisms considered. Compared to one-step classification models, the prediction accuracy of the two-step algorithm was higher for all four mechanisms involved.

Phenolic compounds are environmentally important due to their extensive use in various industries, presence in wastewaters, and potential toxicity. When phenolic compounds reach the activated sludge processes of publicly owned treatment works, they can cause upsets in the operations of the treatment plants. A continuous toxicity testing system was developed based on the bioluminescent bacterium Shk1. The toxicities of 23 phenolic compounds to Shk1 were studied and the 5 min EC50 values were obtained. Quantitative structure-activity relationship models based on the logarithm of the octanol-water partitioning coefficient [log(Kow)] were established for estimating the toxicity of phenolic compounds and the model validity was verified.

Bioluminescent bacteria-based assays can be used for influent wastewater toxicity monitoring for biological wastewater treatment systems. The most thoroughly studied bioluminescent bacteria-based test is the Microtox® assay. However, the response to toxicants of Photobacterium phosphoreum, the marine bacterial strain used in this assay, is different from that of the activated sludge microorganisms. We developed a continuous influent wastewater monitoring system based on the bioluminescent bacterium Shk1, a genetically modified Pseudomonad isolated from the activated sludge in an industrial wastewater treatment plant. The Shk1 toxicity data were correlated with the Microtox® toxicity data for 79 organic compounds and the two toxicity assays were compared. The Shk1 assay is less sensitive than the Microtox® assay and could therefore be more suitable for influent wastewater toxicity monitoring.

Bioluminescent bacteria are widely used in the toxicity assessment of aqueous solutions of heavy metals. They have also been used to screen wastewater treatment plant influent for toxicity resulting from the presence of heavy metals in the influent. However, some studies have indicated that certain strains of bioluminescent bacteria are not appropriate for influent toxicity screening for wastewater treatment plants. Shk1 is a bioluminescent bacterial strain genetically engineered for the specific purpose of monitoring influent wastewater to wastewater treatment plants for toxicity. Heavy metals at sufficient concentrations are toxic to Shk1 cells as they are to activated sludge microorganisms, and the exposure of Shk1 cells to heavy metal ions results in bioluminescence repression. The kinetics of the toxic effects of the heavy metals to Shk1 can be mathematically described in a manner similar to the non-competitive inhibition of enzymes. We determined the inhibition coefficients K i of seven heavy metals. We arranged K i and EC 50 values (a frequently used indicator of toxicity) of the seven heavy metals in increasing order and found that the sequences were in good agreement. We also show that under appropriate conditions, predictions of toxicity can be made based on EC 50 values that contain kinetic information similar to that contained in predictions using K i .

Luminescent bacterium Shk1 was created for the purpose of testing and screening the toxicity of activated sludge wastewater treatment plant influent to avoid toxic shock to the wastewater treatment plant microorganisms. The toxicity of a number of organic compounds was tested using an assay employing Shk1. Because these compounds exhibit toxicity by mechanisms of both polar and nonpolar narcosis, their toxicity cannot be properly modeled together using a quantitative structure-activity relationship model based on the logarithm of the octanol-water partition coefficient (log K(ow)). A solvation parameter model was developed to describe and predict the nonspecific (i.e., polar and nonpolar narcosis) toxicity of organic compounds to Shk1, which does not depend on the discrimination between polar and nonpolar narcotic compounds. The statistically significant model descriptors were the McGowan's characteristic volume (V(x)) and the hydrogen-bond basicity (sigmabetaH). The model was similar to the solvation parameter model developed for Vibrio fischeri, but it did not include an excess molar refraction (R) term.

Phenols are widely used in agriculture and various industries. Many quantitative structure-activity relationships (QSARs) have been developed for phenols. The most toxicologically meaningful QSARs have been established byseparating compounds by their mechanisms of action (MOAs). However, correctly determining the MOA of a compound is not easy. Discriminant analysis was employed in this study to separate selected phenols by three MOAs (polar narcosis, weak acid respiratory uncoupling, and soft electrophilicity) using molecular descriptors as discriminating variables. Results showed that quadratic terms of several molecular descriptors were needed in addition to their linear terms as discriminating variables. Cross-validation of the linear discriminant functions showed that a small total error rate for mechanism classification was achieved.

In this study, the use of solute descriptors (the McGowan's characteristic volume V(X), the excess molar refraction R, the dipolarity/dipolarizability pi(H), the effective or summation hydrogen-bond acidity summation operatoralpha(H) and the effective or summation hydrogen-bond basicity summation operatorbeta(H)) in classifying and predicting the non-polar, polar, and ester narcosis toxicity mechanisms for organic compounds was investigated. Discriminant analysis was performed and the significant discriminating variables were found to be R, pi(H), sigma alpha(H), and (sigma beta (H))(2), the latter of which was created to aid the mechanism classifications. Cross-validation of the non-linear discriminant functions showed a small total error rate of approximately 5% which was reduced to approximately 2% when seven compounds with uncertain a priori mechanism designations were removed. Compared with other methods for toxicity mechanism classification and prediction, the method developed in this study has advantages. It relies on the use of objective numerical values of molecular descriptors that can be calculated and does not require additional experimental work when used for new compounds. The descriptor values can also aid the interpretation of the toxicity mechanism classifications and predictions. Because of the possibility of making incorrect mechanism predictions using a single method, it was recommended that several methods be used together to make the most appropriate mechanism designations and to increase the confidence level associated with the mechanism designations.

Gastrointestinal (GI) symptoms, the primary acute effect of the essential micronutrient copper, paradoxically occur at lower exposure levels than hepatotoxicity, the primary chronic effect. We developed a remedial action criterion (RAC) for copper to protect against GI symptoms, which primarily relate to the stomach copper concentration, and subside within an hour. Using Monte Carlo methods, we generated a distribution of RACs protective against GI symptoms for a 1 h exposure (hourly RACs) based on soil ingestion rate, volume of liquid and food in the stomach, and bioaccessibility. We then generated a distribution of daily RACs, selected as the minimum hourly RAC for each day over a year, constrained by total daily soil ingestion. Next, we identified a percentile of the distribution of daily RACs, and associated RAC, that would result in a high probability of having a minimal number of GI symptom episodes per year. Our analysis indicates that a copper concentration of 3600 mg/kg would result in a 95% probability of having fewer than five episodes of GI symptoms per year, for a child ingesting outdoor soil 180 days per year. Children residing near copper smelters are most likely to experience GI symptoms from ingestion of copper in soil.

Influent toxicity can be a critical problem for publicly owned treatment works that use the activated sludge process as part of their treatment regime. In this project, the researchers developed two protocols for the screening of wastewater treatment plant influent for toxicity. Both protocols are based on a genetically engineered bioluminescent bacterium designated Shk1. Scientists at the University of Tennessee’s Center for Environmental Biotechnology constructed Shk1 from a host Pseudomonas strain isolated from an industrial wastewater treatment plant (WWTP). The first of the two Shk1-based assays that were developed utilizes a batch-wise sampling technique for analyzing grab-samples from industrial effluent, WWTP influent, and the various operations in an activated sludge WWTP (aeration basin, clarifier, etc.). The second method utilizes a continuous sampling technique and is designed for continuous monitoring of the wastewater treatment plant influent upstream of the activated sludge process. The researchers used the batch Shk1 assay to test the influent, activated sludge, and clarifier supernatant in a bench-scale wastewater treatment plant subjected to shock loads of metals (zinc, copper, nickel, and cadmium) for toxicity. They compared data on the repression of bioluminescence to activated sludge respirometry and conventional measures of plant performance (effluent ammonia and chemical oxygen demand, COD). In general, they found that the Shk1 assay indicated toxicity at levels similar to that indicated by activated sludge respirometry. However, no trend could be established between increasing Shk1 toxicity response and effluent quality. Effluent COD and NH3 data (when available) showed little or no significant effect or were highly varible. The researchers used the Shk1-based continuous toxicity screening method to generate toxicity data for a large suite of metals and synthetic organic compounds. They compared these results to literature data for toxicity as indicated by activated sludge respirometry and by the P. phosphoreum-based assay. In general, the Shk1 system gave EC50 values similar to those found in the literature for activated sludge respirometry for 102 organic compounds and to concentrations found to affect activated sludge for seven metals.The researchers adapted the continuous monitoring system for field application and installed it immediately downstream of the effluent from the primary clarifier in a municipal wastewater treatment system. They compared data from the Shk1-based system to plant performance data. During the time of the field study, no significant event occurred during which the operation of the plant was seriously impaired. Therefore, the researchers compared the Shk1 signal to the operations data provided by the plant personnel to determine if any correlation existed between the signal from Shk1 and minor fluctuations in the operations data. They found no simple quantitative relationship between the signal from the toxicity monitoring system and the plant performance data. They applied principal component and factor analysis to the Shk1data and 20 additional plant variables. The results of these analyses showed that 10 principal components were needed to account for 90% of the variability of the data and that the signal from Shk1 was therefore not sufficient to predict the system state in the absence of a major toxic event without knowledge of the values of other operating variables. In summary, these analyses indicated that the Shk1 signal would be a valuable addition to models to predict the future system state from the influent, operating, and effluent variables but it is not a sufficient variable by itself. This title belongs to WERF Research Report Series ISBN: 9781843396314 (eBook)

A three-stage system was developed to automate a batchwise toxicity testing protocol designed for assessing wastewater toxicity to activated sludge. The three-stage system used the luminescent bacterium Shkl. The three stages were cell storage, cell activation, and continuous toxicity testing. Shkl cells were stored in a bioreactor at 4 degrees C when the system was not in use and activated in another bioreactor for use in toxicity tests conducted in a continuous manner. The system could quickly be switched between the "off" and "on" modes, and operation of the system was easy. The stability of the system, in terms of cell density and bioluminescence in the storage and activation bioreactors, and the response of the activated cells to a metal and an organic toxicant were studied. The feasibility of the system design was demonstrated by simulating zinc toxicity episodes in synthetic wastewater. The needs for further modifications and improvements of the system were discussed.

Toxicants in municipal sewage treatment plant (STP) influent wastewater may inhibit the biological activity of the activated sludge and cause treatment plant process upsets. Such process upsets may be avoided if influent wastewater is monitored for toxicity and protective actions are taken when toxicity is detected. A comprehensive review of the methods that can be used for assessing wastewater toxicity to biological treatment systems was conducted several years ago and the resultant report was published in 2000 by Water Environment Research Foundation (WERF). The WERF report also specified the criteria for influent wastewater toxicity monitoring methods and suggested research needs to be addressed. A significant amount of effort was made since the publication of the WERF report to develop new assays or devices and to improve existing ones. In this manuscript, recent research and developments in methods for assessing wastewater toxicity to activated sludge were reviewed. The literature indicates that bioluminescence- and respirometry-based methods received much attention in recent research. A comparison of the new/improved methods with the criteria described in the WERF report reveals that none of these methods has been shown to meet all the specified criteria. The present review also indicates that research efforts since 2000 have not fully taken into account the criteria for influent wastewater toxicity monitoring methods and have not addressed the research needs proposed in the WERF report. Keywords: Activated sludge, Sewage treatment plant, Influent, Wastewater, Toxicity

Statistical models have long been used for reliability analysis and risk assessment. In the present study, an accelerated life-test model was used to analyze a set of dose-time-response data obtained with the terrestrial isopod Porcellio scaber. Survival data were experimentally obtained by exposing P. scaber to diazinon (a nonpersistent insecticide) at six concentrations between 2 and 11.31 microg/g (toxicant/soil). Survival data are presented on a weekly basis. The accelerated life-test model assumed a log-normal distribution and constant variance across all diazinon concentrations. Model parameters were obtained by maximum likelihood estimation. The accelerated life-test model was compared to a toxicokinetics-based model reported in the literature. Survival predictions made by both models were compared with the observed data. Both the accelerated life-test model and the toxicokinetics-based model underestimated toxicity at a diazinon concentration of 8 microg/g. Overall, however, the accelerated life-test model outperformed the toxicokinetics-based model, with survival predictions closer to the observed data in most cases and a stronger correlation between predicted and observed survivals. However, as a statistical model, the accelerated life-test model did not reveal mechanistic information, and only statistical and distributional interpretations of its model parameters could be made.

The use of regression methods for classifying and predicting the mechanisms of toxic action of phenols was investigated in this study. Multiresponse regression was conducted using a total of six linear and nonlinear regression methods: simple linear regression (LinReg), logistic regression (LogReg), generalized additive model (GAM), locally weighted regression scatter plot smoothing (LOWESS), multivariate adaptive regression splines (MARS), and projection pursuit regression (PPR). A database containing phenols acting by four mechanisms (polar narcosis, weak acid respiratory uncoupling, proelectrophilicity, and soft electrophilicity) was used to assess the performances of the six regression methods in the multiresponse regression approach. For comparison purposes, traditional linear discriminant analysis (LDA) was also conducted as a baseline method to study the potential improvement of prediction accuracy by the multiresponse regression approach. Results showed that compared to LDA, the overall mechanism prediction error rate could be reduced to below 10% by multiresponse regression based on PPR. In addition to prediction accuracy, interpretability of the resultant models was discussed.

The most successful quantitative structure–activity relationships have been developed by separating compounds by their mechanisms of toxic action (MOAs). However, to correctly determine the MOA of a compound is often not easy. We investigated the usefulness of discriminant analysis and logistic regression in determining MOAs. The discriminating variables used were the logarithm of octanol–water partition coefficients ( log K ow ) and the experimental toxicity data obtained from Pimephales promelas and Tetrahymena pyriformis assays. Small total error rates were obtained when separating nonpolar narcotic compounds from other compounds, however, relatively high total error rates were obtained when separating less reactive compounds (polar, ester, and amine narcotics) from more reactive compounds (electrophiles, proelectrophiles, and nucleophiles).

Heavy metals are known to be inhibitory and toxic to the activated-sludge microbial community in biological wastewater treatment plants. Toxicity screening of aqueous mixtures of these heavy metal ions in plant influent could use both chemical and biological methods. As a biological method, luminescent bacterial bioreporters offer the advantages of a simple test procedure and rapid response. Current biologically based methods for screening aqueous streams for toxicity are labor-intensive, inaccurate, or difficult to use in continuous monitoring applications. In the present study, a system was developed that is simple and easily automated. This system is based on the bacterium Shk1, a genetically engineered bioluminescent Pseudomonad whose parent strain was originally isolated from activated sludge. Compared with other bioluminescence-based systems (specifically, the Microtox assay), the system of the present study more accurately reflects the effects of the toxicity of common metal ions on activated-sludge respirometry without being overly sensitive to typical constituents of wastewater. The use of Shk1 as a bioluminescent reporter for heavy metal toxicity testing for the application of wastewater treatment influent toxicity screening is presented in this study.

In aquatic toxicity testing, no single test species is sensitive to all toxicants. Therefore, test batteries consisting of several individual assays are becoming more common. The organisms in a test battery should be representative of the entire system of interest. The results of the assays should be complementary to other components in the test battery to avoid redundancy. With the aid of multidimensional scaling (MDS), a multivariate statistical method, we examined the toxicity data of five bioassays (the continuous Shk1, Polytox®, activated sludge respiration inhibition, Nitrosomonas, and Tetrahymena assays) that could serve as test battery components for the assessment of wastewater toxicity to activated sludge. MDS mapped the five assays into a two-dimensional space and showed that the Nitrosomonas assay should be included in test batteries plus one of the remaining four assays for assessing wastewater toxicity to activated sludge.

Luminescent bacterium Shk1 was created for the purpose of testing and screening the toxicity of activated sludge wastewater treatment plant influent to avoid toxic shock to the wastewater treatment plant microorganisms. The toxicity of a number of organic compounds was tested using an assay employing Shk1. Because these compounds exhibit toxicity by mechanisms of both polar and nonpolar narcosis, their toxicity cannot be properly modeled together using a quantitative structure-activity relationship model based on the logarithm of the octanol-water partition coefficient (log K(ow)). A solvation parameter model was developed to describe and predict the nonspecific (i.e., polar and nonpolar narcosis) toxicity of organic compounds to Shk1, which does not depend on the discrimination between polar and nonpolar narcotic compounds. The statistically significant model descriptors were the McGowan's characteristic volume (V(x)) and the hydrogen-bond basicity (sigmabetaH). The model was similar to the solvation parameter model developed for Vibrio fischeri, but it did not include an excess molar refraction (R) term.

Toxicity assays based on bioluminescent bacteria have several advantages including a quick response and an easily measured signal. The Shk1 assay is a procedure for wastewater toxicity testing based on the bioluminescent bacterium Shk1. Using the Shk1 assay, the toxicity of 98 organic chemicals were measured and EC 50 values were obtained. Quantitative structure–activity relationship (QSAR) models based on the logarithm of the octanol–water partition coefficient (log( K ow )) were developed for individual groups of organic chemicals with different functional groups. The correlation coefficients for different groups of organic compounds varied between 0.69 and 0.99. An overall QSAR model without discriminating the functional groups, which can be used for a quick estimate of the toxicities of organic chemicals, was also developed and model predictions were compared to experimental data. The model accuracy was found to be one order of magnitude from the observed values.

The most successful quantitative structure-activity relationships (QSARs) have been developed by separating toxicants by their mechanisms of action (MOAs). However, since the activity of a chemical compound on an organism is dependent upon several physical, chemical and biological factors, among which interactions may also exist, the MOA of a compound is not easily determined. In this study, the use of discriminant analysis and logistic regression in distinguishing between narcotic and reactive compounds was investigated. The discriminating variables included hydrophobicity (log(K(ow))) and electrophilicity descriptors (S(av)(N), E(HOMO), and E(LUMO)). Classification results showed that logistic regression gave a smaller total error rate compared to discriminant analysis. Since the value of the descriptors can be calculated, the classification methods can be used in predictive toxicology.