Toward Robust Fault Identification of Complex Industrial Processes Using Stacked Sparse-Denoising Autoencoder With Softmax Classifier.

This article proposes a robust end-to-end deep learning-induced fault recognition scheme by stacking multiple sparse-denoising autoencoders with a Softmax classifier, called stacked spare-denoising autoencoder (SSDAE)-Softmax, for the fault identification of complex industrial processes (CIPs). Specifically, sparse denoising autoencoder (SDAE) is established by integrating a sparse AE (SAE) with a denoising AE (DAE) for the low-dimensional but intrinsic feature representation of the CIP monitoring data (CIPMD) with possible noise contamination. SSDAE-Softmax is established by stacking multiple SDAEs with a layerwise pretraining procedure, and a Softmax classifier with a global fine-tuning strategy. Furthermore, SSDAE-Softmax hyperparameters are optimized by a relatively new global optimization algorithm, referred to as the state transition algorithm (STA). Benefiting from the deep learning-based feature representation scheme with the STA-based hyperparameter optimization, the underlying intrinsic characteristics of CIPMD can be learned automatically and adaptively for accurate fault identification. A numeric simulation system, the benchmark Tennessee Eastman process (TEP), and a real industrial process, that is, the continuous casting process (CCP) from a top steel plant of China, are used to validate the performance of the proposed method. Experimental results show that the proposed SSDAE-Softmax model can effectively identify various process faults, and has stronger robustness and adaptability against the noise interference in CIPMD for the process monitoring of CIPs.

Complete Nutrient Removal Coupled to Nitrous Oxide Production as a Bioenergy Source by Denitrifying Polyphosphate-Accumulating Organisms.

Coupled aerobic-anoxic nitrous decomposition operation (CANDO) is a promising emerging bioprocess for wastewater treatment that enables direct energy recovery from nitrogen (N) in three steps: (1) ammonium oxidation to nitrite; (2) denitrification of nitrite to nitrous oxide (N2O); and (3) N2O conversion to N2 with energy generation. However, CANDO does not currently target phosphorus (P) removal. Here, we demonstrate that denitrifying polyphosphate-accumulating organism (PAO) enrichment cultures are capable of catalyzing simultaneous biological N and P removal coupled to N2O generation in a second generation CANDO process, CANDO+P. Over 7 months (>300 cycles) of operation of a prototype lab-scale CANDO+P sequencing batch reactor treating synthetic municipal wastewater, we observed stable and near-complete N removal accompanied by sustained high-rate, high-yield N2O production with partial P removal. A substantial increase in abundance of the PAO Candidatus Accumulibacter phosphatis was observed, increasing from 5% of the total bacterial community in the inoculum to over 50% after 4 months. PAO enrichment was accompanied by a strong shift in the dominant Accumulibacter population from clade IIC to clade IA, based on qPCR monitoring of polyphosphate kinase 1 (ppk1) gene variants. Our work demonstrates the feasibility of combining high-rate, high-yield N2O production for bioenergy production with combined N and P removal from wastewater, and it further suggests a putative denitrifying PAO niche for Accumulibacter clade IA.

Removal of cefalexin using yeast surface-imprinted polymer prepared by atom transfer radical polymerization.

The first use of yeast as a support in the molecular imprinting field combined with atom transfer radical polymerization was described. Then, the as-prepared molecularly imprinted polymers were characterized by Fourier transmission infrared spectrometry, scanning electron microscope, thermogravimetric analysis, and elemental analysis. The obtained imprinted polymers demonstrated elliptical-shaped particles with the thickness of imprinting layer of 0.63 µm. The batch mode experiments were adopted to investigate the adsorption equilibrium, kinetics, and selectivity. The kinetic properties of imprinted polymers were well described by the pseudo-second-order kinetic equation, indicating the chemical process was the rate-limiting step for the adsorption of cefalexin (CFX). The equilibrium data were well fitted by the Freundlich isotherm, and the multimolecular layers adsorption capacity of imprinted polymers was 34.07 mg g(-1) at 298 K. The selectivity analysis suggested that the imprinted polymers exhibited excellent selective recognition for CFX in the presence of other compounds with related structure. Finally, the analytical method based on the imprinted polymers extraction coupled with high-performance liquid chromatograph was successfully used for CFX analysis in spiked pork and water samples.

Preparation of thermal-responsive magnetic molecularly imprinted polymers for selective removal of antibiotics from aqueous solution.

A novel thermal-responsive magnetic molecularly imprinted polymers (TMMIPs), maghemite/silica/poly (N-isopropylacrylamide-co-acrylamide-co-ethylene glycol dimethacrylate) (γ-Fe(2)O(3)/SiO(2)/P (NIPAm-co-AAm-co-EGDMA)), were developed as a potential effective adsorbent for selectively remove sulfamethazine (SMZ) exist in aquatic environments, which has been recognized as a warranting considerable issue. Free radical polymerization of NIPAm, AAm and EGDMA was performed in dimethyl sulfoxide/water (DMSO/H(2)O) (v/v=9/1) with 2,2'-azobisisobutyronitrile (AIBN) as initiator to coat γ-Fe(2)O(3)/SiO(2)/3-(methacryloxyl) propyl trimethoxysilane (MPS) microspheres through the capture of oligomers with the aid of vinyl groups on their surfaces. The unique aspect of TMMIPs was that they combined molecular recognition, magnetic separation and thermo-responsiveness. The got material was characterized by SEM, TEM, FT-IR and VSM. Batch mode adsorption studies were carried out to investigate the specific adsorption equilibrium, kinetics, and selective recognition ability of TMMIPs. Reversible recognition and release of template molecule were realized by changing environmental temperatures. Several other antibiotics were selected as model analytes to evaluate the selective recognition performance of TMMIPs. The TMMIPs have good temperature response, selectivity and reusability, making them possible in applying for antibiotics separation and controlled release.

Preparation of molecularly imprinted nanoparticles with superparamagnetic susceptibility through atom transfer radical emulsion polymerization for the selective recognition of tetracycline from aqueous medium.

In the work, we reported an effective method for the preparation of molecularly imprinted nanoparticles with superparamagnetic susceptibility through atom transfer radical emulsion polymerization (ATREP), and then as-prepared magnetic molecularly imprinted nanoparticles (MMINs) were evaluated as adsorbents for selective recognition of tetracycline (TC) molecules from aqueous medium. The resulting nanoparticles were characterized by FT-IR, TGA, VSM, SEM and TEM. The results demonstrated MMINs with a narrow diameter distribution were cross-linked with modified Fe3O4 particles, composed of imprinted layer and exhibited good magnetic sensitivity, magnetic and thermal stability. Batch rebinding studies were carried out to determine the specific adsorption equilibrium, kinetics, and selective recognition. The estimated adsorption capacity of MMINs towards TC by the Langmuir isotherm model was 12.10 mgg(-1) at 298 K, which was 6.33 times higher than that of magnetic non-molecularly imprinted nanoparticles (MNINs). The kinetic property of MMINs was well-described by the pseudo-second-order rate equation. The results of selective recognition experiments demonstrated outstanding affinity and selectivity towards TC over competitive antibiotics. The reusability of MMINs showed no obviously deterioration at least five repeated cycles in performance. In addition, the MMINs prepared were successfully applied to the extraction of TC from the spiked pork sample.

Selective recognition of 4-nitrophenol from aqueous solution by molecularly imprinted polymers with functionalized tetratitanate whisker composites as support.

Three kinds of molecularly imprinted polymers (MIPs) were obtained with surface molecular imprinting technique on functionalized potassium tetratitanate whisker (F-PTW). The results of adsorption experiments indicated that MIP prepared using PTW modified with N-(2-aminoethyl)-3-(trimethoxysilyl)propylamine (AAPTS) (F-PTW A) as support [MIP(1)] was superior to the other two polymers, then MIP(1) was selected to analyze the 4-nitrophenol (4-NP) adsorption process from aqueous solution in this study. AAPTS offered hydrophilic exterior that allowed to self-assemble with the template 4-NP through intermolecular interaction rather than based on the interactions between the functional monomers and template. Equilibrium adsorption data were analyzed by the Langmuir and Freundlich isotherm models at various temperatures. Kinetic properties were successfully investigated by pseudo-first-order model, pseudo-second-order model, intraparticle diffusion equation, initial adsorption rate, half-adsorption time. A diffusion-controlled process as the essential adsorption rate-controlling step was also proposed. The performance of such imprinted polymer was further demonstrated by high-performance liquid chromatography, and the results showed that the selectivity of MIP(1) exhibited higher affinity for template 4-NP over competitive phenolic compounds than that of non-imprinted polymer NIP(1). MIP(1) could be reused four times without significant loss in the adsorption capacity.