ANN-Based Soft Sensor to Predict Effluent Violations in Wastewater Treatment Plants
Ivan Pisa1, Ignacio Santín2, Jose Lopez Vicario3
1Department of Telecommunications and Systems Engineering, Escola d'Enginyeria, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain. ivan.pisa@uab.cat.
Sensors (Basel, Switzerland)
|March 16, 2019
Summary
This study introduces an artificial neural network (ANN) soft sensor using Long-Short Term Memory (LSTM) to predict nitrogen pollutants in wastewater. The system accurately forecasts violations, aiding in environmental protection and regulatory compliance.
Area of Science:
- Environmental Engineering
- Artificial Intelligence
- Water Quality Management
Background:
- Wastewater treatment plants (WWTPs) face challenges in meeting strict effluent regulations for pollutants like nitrogen.
- Existing control strategies sometimes fail to prevent violations, leading to environmental damage such as eutrophication and aquatic toxicity.
- Accurate prediction of nitrogen-derived components is crucial for proactive control and compliance.
Purpose of the Study:
- To develop and evaluate an artificial neural network (ANN)-based soft sensor for predicting ammonium (SNH) and total nitrogen (SNtot) in wastewater effluent.
- To integrate the soft sensor with existing control strategies to enhance their effectiveness in preventing regulatory violations.
- To assess the prediction accuracy and violation detection capabilities of the proposed soft sensor.
Main Methods:
- Implementation of a Long-Short Term Memory (LSTM) network as a soft sensor within the Benchmark Simulation Model N.2 (BSM2).
- Testing the soft sensor's performance with three control strategies of varying complexity.
- Evaluating the accuracy of nitrogen-derived component predictions and the probability of detecting effluent limit violations.
Main Results:
- The LSTM-based soft sensor demonstrated high accuracy in predicting nitrogen-derived products.
- The system achieved an 86%–94% probability of detecting violations of BSM2 effluent limits.
- Calibration of the soft sensor can further improve prediction accuracy, potentially achieving perfect violation prediction at the cost of increased false positives.
Conclusions:
- The proposed ANN-based soft sensor effectively predicts key nitrogen parameters in wastewater.
- The predictive capabilities enhance the performance of WWTP control strategies, reducing the likelihood of effluent violations.
- This technology offers a valuable tool for improving water quality management and ensuring environmental compliance.
Related Concept Videos
02:27Predicting Molecular Geometry
45.7K
VSEPR Theory for Determination of Electron Pair Geometries
45.7K
02:01Plant Cells and Tissues
65.6K
Plant tissues are collections of similar cells performing related functions. Different plant tissues will have their own specialized roles and can be combined with other tissues to form organs such as flowers, fruit, stem, and leaves. Two major types of plant tissue include meristematic and permanent tissue.
65.6K
01:56Plant Hormones
27.5K
Plant hormones—or phytohormones—are chemical molecules that modulate one or more physiological processes of a plant. In animals, hormones are often produced in specific glands and circulated via the circulatory system. However, plants lack hormone-producing glands.
27.5K
00:53Tonicity in Plants
59.7K
Tonicity describes the capacity of a cell to lose or gain water. It depends on the quantity of solute that does not penetrate the membrane. Tonicity delimits the magnitude and direction of osmosis and results in three possible scenarios that alter the volume of a cell: hypertonicity, hypotonicity, and isotonicity. Due to differences in structure and physiology, tonicity of plant cells is different from that of animal cells in some scenarios.
59.7K
02:43Plant Cell Wall
60.2K
The plant cell wall gives plant cells shape, support, and protection. As a cell matures, its cell wall specializes according to the cell type. For example, the parenchyma cells of leaves possess only a thin, primary cell wall.
60.2K
01:03Prediction Intervals
3.4K
The interval estimate of any variable is known as the prediction interval. It helps decide if a point estimate is dependable.
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
3.4K