Soft Sensor Application in Identification of the Activated Sludge Bulking Considering the Technological and Economical Aspects of Smart Systems Functioning (original) (raw)
Related papers
Improving activated sludge classification based on imbalanced data
Journal of Hydroinformatics, 2014
A fast and accurate classification method for sewage sludge biological activity classification is of great significance for wastewater treatment. However, the data are often imbalanced and the accuracy of traditional classification algorithms applied to imbalanced small classes of data is very low. Such small classes are crucial application data. Therefore, based on the analysis of eight microorganisms, a novel method is proposed in this paper for the classification of activated sludge known as balanced support-vector-based back-propagation (SV-BP) neural network. It first splits the multiclass classification problem into a plurality of pairwise classification problems and uses a support vector machine (SVM) to achieve equalization. Second, the new dataset is produced, following which back-propagation neural network (BPNN) is used for training and classification. To examine the efficiency of the model, 1731 real data points are collected from a wastewater treatment factory and divide the data into four classes with the help of wastewater experts. Based on the new model, data redundancy and noise are greatly reduced. With area under the curve (AUC) measurements, we find that the AUC of SV-BP is 6.9% higher than classical BPNN. In addition, the small-class recognition rate of SV-BP is far better than that by classical BPNN and SVM algorithms.
E3S Web of Conferences, 2017
In the study, models developed using data mining methods are proposed for predicting wastewater quality indicators: biochemical and chemical oxygen demand, total suspended solids, total nitrogen and total phosphorus at the inflow to wastewater treatment plant (WWTP). The models are based on values measured in previous time steps and daily wastewater inflows. Also, independent prediction systems that can be used in case of monitoring devices malfunction are provided. Models of wastewater quality indicators were developed using MARS (multivariate adaptive regression spline) method, artificial neural networks (ANN) of the multilayer perceptron type combined with the classification model (SOM) and cascade neural networks (CNN). The lowest values of absolute and relative errors were obtained using ANN+SOM, whereas the MARS method produced the highest error values. It was shown that for the analysed WWTP it is possible to obtain continuous prediction of selected wastewater quality indicators using the two developed independent prediction systems. Such models can ensure reliable WWTP work when wastewater quality monitoring systems become inoperable, or are under maintenance.
1995
This study introduces the application of a neural network method to estimate chemical oxygen demand (COD) in a wastewater treatment process. It presents the back propagation algorithm with a generalized sigmoid function in detail. Several simulations were investigated in order to select suitable learning rates and relative coefficients in the network for accelerating the speed of convergence in learning. The results of simulation to estimate real sampling data of COD for a specific real world wastewater treatment process are explained.
Cost Effective Soft Sensing for Wastewater Treatment Facilities
IEEE Access
Wastewater treatment plants are complex, non-linear, engineered systems of physical, biological and chemical processes operating at different timescales. Sensor systems are used to monitor wastewater treatment plants in order to ensure public safety and for efficient management of the plants. However, parameters of interest for wastewater can require expensive or inaccurate sensors or may require off-site laboratory analysis. For example, ammonium is important as a prime indicator of treatment efficiency and is highly regulated in discharge water. But ammonium sensors are also expensive at over $10,000 (AUD) per sensor. Soft sensors are computational models that accurately estimate process variables using the measurements from few physical sensors and can offer a cost-effective substitute for expensive wastewater sensors such as ammonium. In this paper, we propose a hybrid neural network architecture for learning soft sensors for complex phenomena. Our network architecture fuses sequential modelling with Gated Recurrent Neural Network units (GRUs) to capture global trends, with Convolution Neural Network (CNN) kernels to facilitate learning of local behaviours. We demonstrate the effectiveness of our technique using real-world data from a wastewater treatment plant with two-stage high-rate anaerobic and high-rate algal treatments. Secondly, we propose a novel data preparation algorithm that enables the deep learning techniques to learn from a limited data and facilitates fair evaluation. We develop and learn a soft sensor to predict ammonium and study its generalization. Our results demonstrate fit for purpose accuracy and that the soft sensor model is able to capture complex temporal patterns of the ground truth sensor time series. Finally, we publicly release an annotated data set of a secondary wastewater treatment plant to accelerate the research in the development of soft sensors. INDEX TERMS Wastewater treatment, high rate algal ponds, ammonium, soft sensors, hybrid model, recurrent neural network, deep learning.
Intelligent System for the Predictive Analysis of an Industrial Wastewater Treatment Process
Sustainability
Considering the exponential growth of today’s industry and the wastewater results of its processes, it needs to have an optimal treatment system for such effluent waters to mitigate the environmental impact generated by its discharges and comply with the environmental regulatory standards that are progressively increasing their demand. This leads to the need to innovate in the control and management information systems of the systems responsible to treat these residual waters in search of improvement. This paper proposes the development of an intelligent system that uses the data from the process and makes a prediction of its behavior to provide support in decision making related to the operation of the wastewater treatment plant (WWTP). To carry out the development of this system, a multilayer perceptron neural network with 2 hidden layers and 22 neurons each is implemented, together with process variable analysis, time-series decomposition, correlation and autocorrelation techniqu...
Modeling Sludge volume index of activated sludge process in municipal WWTP is a difficult task to accomplish due to the high nonlinearity of the plant and the non-uniformity and variability of influent quantity, quality parameters, and operation condition. ANNs were developed for the prediction of the Sludge Volume Index using influent quality parameters and operating parameters of Batna Wastewater Treatment Plant from 2011 to 2014. The best model given by the neural network for the SVI prediction composed of one input layer with fifteen input variables, one hidden layer with thirteen nodes and one output layer with one output variable with R = 0.8784 and RMSE = 0.443. The results demonstrate the ability of the appropriate Neural Network models for the prediction of SVI. This provides a very useful tool that can be used by WWTP operators in their daily management to increase treatment process performances and WWTP reliability.
Modelling of Wastewater Treatment Plant With Decision and Regression Trees
2002
Wastewater treatment plants (WWTP) are dynamic and complex systems, where management can be improved by applying different approaches to modeling and predicting their operation. Machine learning tools (decision and regression trees) were used to build useful prediction models for wastewater treatment plant operation. Database used for building the models is composed of quantitative and qualitative data, including estimations and microbiological analysis on the WWTP. Attributes are given as daily averaged values of the measured data. Each record in the data set that we used in our study comprises information of a one-day situation of the plant. The situation is then identified (classified) according to the plant operation status, i.e., good operation, sludge problems, etc. Based on assumption that in reality problems with bulking sludge can not appear the same day as they are presented in the data set, we incorporated a one and two-day delay for sludge problems. Thus, new attributes were added, carrying the information of a day-before inflow. Two types of models are constructed. The first type is a decision tree model, predicting operational conditions of the plant. These models discover when a specific situation (i.e. bulking sludge) occurs on the WWTP. The second type is model for numeric prediction, which is carried out using model (regression) trees. These results are comparable with mathematical models results, since they predict (calculate) a specific outflow variable from the variables measured on the inflow point of the plant. The program package WEKA, which includes most of popular machine learning algorithms, was used for constructing the models.
Sludge Bulking Prediction Using Principle Component Regression and Artificial Neural Network
Mathematical Problems in Engineering, 2012
Sludge bulking is the most common solids settling problem in wastewater treatment plants, which is caused by the excessive growth of filamentous bacteria extending outside the flocs, resulting in decreasing the wastewater treatment efficiency and deteriorating the water quality in the effluent. Previous studies using molecular techniques have been widely used from the microbiological aspects, while the mechanisms have not yet been completely understood to form the deterministic cause-effect relationship. In this study, system identification techniques based on the analysis of the inputs and outputs of the activated sludge system are applied to the data-driven modeling. Principle component regression PCR and artificial neural network ANN were identified using the data from Chongqing wastewater treatment plant CQWWTP , including temperature, pH, biochemical oxygen demand BOD , chemical oxygen demand COD , suspended solids SSs , ammonia NH 4 , total nitrogen TN , total phosphorus TP , and mixed liquor suspended solids MLSSs . The models were subsequently used to predict the sludge volume index SVI , the indicator of the bulking occurrence. Comparison of the results obtained by both models is also presented. The results showed that ANN has better prediction power R 2 0.9 than PCR R 2 0.7 and thus provides a useful guide for practical sludge bulking control.
KSOM and MLP neural networks for on-line estimating the efficiency of an activated sludge process
Chemical Engineering Journal, 2006
This work is devoted to the study of the Saint Cyprien (south of France) activated sludge WWTP process and to the on-line estimation of chemical parameters (influent and effluent chemical oxygen demand, ammonia and suspended solids) not easily measurable on-line. Their knowledge makes it possible to estimate the process efficiency and to provide reliable information for the plant monitoring. A tool including Kohonen's self-organizing maps and a multi level perceptron is used. The Kohonen's self organizing maps neural network is applied to analyze the multidimensional Saint Cyprien process data and to diagnose the interrelationship of the process variables in an activated sludge WWTP. The multi level perceptron is used as estimation tool. The obtained results are satisfactory. The information provided by the estimation procedure is sufficiently reliable and precise to be exploitable by operators in charge of the plant monitoring and maintenance. It allows understanding how the system is evolving. The whole procedure (Kohonen's self-organizing maps and multi level perceptron) uses tools which proved to be efficient and complementary.
DESALINATION AND WATER TREATMENT
Sludge management account for high economic costs and energy consumption in wastewater treatment. Accurate forecasting of sewage sludge generation thus can be important for the planning, operation and optimization of processes at wastewater treatment plant (WWTP). In this study data from a municipal treatment plant with a capacity of 88 thousand cubic meters of sewage per day located in south part of Poland were used to find a good forecasting model for sludge amount prediction. Among models an autoregressive integrated moving average (ARIMA) is one of the popular linear models in time series forecasting. Since the ARIMA model cannot capture the non-linear structure of the data research activities in forecasting suggest using neural networks. Long-short term memory (LSTM) recurrent neural network proved its usability in time series forecasting. Looking at the curve representing data of sludge amount generated in the previous years at WWTP the linear and non-linear patterns could be distinguished. To address these issue a hybrid methodology that combines advantage of ARIMA and LSTM was proposed and used for forecasting purpose. Experimental results showed that the combined model can be an effective way to improve the forecasting accuracy of sludge amount generated at this WWTP.