
The ubiquitous presence of pharmaceuticals in the aquatic environment is a worldwide problem of today. Current analytical methods reveal ever more pharmaceuticals in different water bodies. The concentrations in waters that are detectable reach nano-and picogram ranges, i.e., ppb and ppt levels. Observed concentrations amount to as high as few micrograms. Among the major entry paths are wastewater treatment plants, which are often unable to eliminate the pharmaceuticals sufficiently relying on their three conventional purification stages. Hence, pharmaceuticals enter the aquatic environment without desirable deconstruction. Thus, advanced wastewater treatment processes are under development to retain or eliminate these trace contaminants. According to Decision 2018/840, a watchlist of 15 contaminants of significant interest has been established for the monitoring of surface waters in the European Union. The contaminants include biocides and pharmaceuticals, among them three estrogens, estrone (E1), 17-β-estradiol (E2) and 17-α-ethinylestradiol, (EE2), the antibiotics azithromycin, clarithromycin and erythromycin of macrolide type, ciprofloxacin and amoxicillin of fluoroquinolone and betalactame type. This review will provide an overview of the currently explored and researched methods for the realization of a fourth purification stage in wastewater treatment plants. To this purpose, biological, chemical and physical purification processes are reviewed and their characteristics and potential discussed. The degradation efficacy of the pharmaceuticals on the EU-Watch list will be compared and evaluated with respect to the most promising processes, which might be realized on large scale. Last but not least, recent and novel pilot plants will be presented and discussed.
Citation: Melanie Voigt, Alexander Wirtz, Kerstin Hoffmann-Jacobsen, Martin Jaeger. Prior art for the development of a fourth purification stage in wastewater treatment plant for the elimination of anthropogenic micropollutants-a short-review[J]. AIMS Environmental Science, 2020, 7(1): 69-98. doi: 10.3934/environsci.2020005
[1] | Biljana Ilić, Dragica Stojanovic, Gordana Djukic . Green economy: mobilization of international capital for financing projects of renewable energy sources. Green Finance, 2019, 1(2): 94-109. doi: 10.3934/GF.2019.2.94 |
[2] | Mukul Bhatnagar, Sanjay Taneja, Ercan Özen . A wave of green start-ups in India—The study of green finance as a support system for sustainable entrepreneurship. Green Finance, 2022, 4(2): 253-273. doi: 10.3934/GF.2022012 |
[3] | Nabil Haque, Sungida Rashid . Host country characteristics attracting climate projects through public-private partnerships. Green Finance, 2019, 1(4): 405-428. doi: 10.3934/GF.2019.4.405 |
[4] | Shahinur Rahman, Iqbal Hossain Moral, Mehedi Hassan, Gazi Shakhawat Hossain, Rumana Perveen . A systematic review of green finance in the banking industry: perspectives from a developing country. Green Finance, 2022, 4(3): 347-363. doi: 10.3934/GF.2022017 |
[5] | Laura Grumann, Mara Madaleno, Elisabete Vieira . The green finance dilemma: No impact without risk – a multiple case study on renewable energy investments. Green Finance, 2024, 6(3): 457-483. doi: 10.3934/GF.2024018 |
[6] | Goshu Desalegn . Insuring a greener future: How green insurance drives investment in sustainable projects in developing countries?. Green Finance, 2023, 5(2): 195-210. doi: 10.3934/GF.2023008 |
[7] | Wen-Tien Tsai . Green finance for mitigating greenhouse gases and promoting renewable energy development: Case study in Taiwan. Green Finance, 2024, 6(2): 249-264. doi: 10.3934/GF.2024010 |
[8] | María Cantero-Saiz, Sergio Sanfilippo-Azofra, Begoña Torre-Olmo, Violeta Bringas-Fernández . ESG and bank profitability: the moderating role of country sustainability in developing and developed economies. Green Finance, 2025, 7(2): 288-331. doi: 10.3934/GF.2025011 |
[9] | Yafei Wang, Jing Liu, Xiaoran Yang, Ming Shi, Rong Ran . The mechanism of green finance's impact on enterprises' sustainable green innovation. Green Finance, 2023, 5(3): 452-478. doi: 10.3934/GF.2023018 |
[10] | Raja Elyn Maryam Raja Ezuma, Nitanan Koshy Matthew . The perspectives of stakeholders on the effectiveness of green financing schemes in Malaysia. Green Finance, 2022, 4(4): 450-473. doi: 10.3934/GF.2022022 |
The ubiquitous presence of pharmaceuticals in the aquatic environment is a worldwide problem of today. Current analytical methods reveal ever more pharmaceuticals in different water bodies. The concentrations in waters that are detectable reach nano-and picogram ranges, i.e., ppb and ppt levels. Observed concentrations amount to as high as few micrograms. Among the major entry paths are wastewater treatment plants, which are often unable to eliminate the pharmaceuticals sufficiently relying on their three conventional purification stages. Hence, pharmaceuticals enter the aquatic environment without desirable deconstruction. Thus, advanced wastewater treatment processes are under development to retain or eliminate these trace contaminants. According to Decision 2018/840, a watchlist of 15 contaminants of significant interest has been established for the monitoring of surface waters in the European Union. The contaminants include biocides and pharmaceuticals, among them three estrogens, estrone (E1), 17-β-estradiol (E2) and 17-α-ethinylestradiol, (EE2), the antibiotics azithromycin, clarithromycin and erythromycin of macrolide type, ciprofloxacin and amoxicillin of fluoroquinolone and betalactame type. This review will provide an overview of the currently explored and researched methods for the realization of a fourth purification stage in wastewater treatment plants. To this purpose, biological, chemical and physical purification processes are reviewed and their characteristics and potential discussed. The degradation efficacy of the pharmaceuticals on the EU-Watch list will be compared and evaluated with respect to the most promising processes, which might be realized on large scale. Last but not least, recent and novel pilot plants will be presented and discussed.
Machinery has become an integral part of human life, especially considering the technological advancements related to Industry 4.0. Failures occurring on such crucial machinery lead to unplanned downtime, ultimately resulting in loss in economic aspects [1]. This is catastrophic in industry and public transport since these failures stop production, causing hassles to the public. Hence, machine diagnostics is of high importance in such instances. Fault prediction in an early stage dramatically improves the machine's lifetime, reducing costs and preventing downtimes.
Air and oil leaks are two of the predominant operational failures in public transport modalities, especially in metro trains, which are our prime objective. Air leaks are prone to occur in the dryer component, whereas oil leaks are prone to occur in the compressor component [2]. Various sensors, like pressure transducers, pneumatic sensors, motor current, etc., are used to analyze and diagnose the faults in metro trains [3,4,5]. An abnormal change can be observed in these sensors in the event of a fault in that component. Continuous monitoring of the vulnerable components with the sensors mentioned above can help identify the occurrence of a fault in that component [6].
Predictive maintenance has been an emerging technology in machine diagnostics, aiming to predict faults early and perform maintenance to prevent catastrophic events [7]. Also, anomaly detection based on sensor data on an early scale will reduce maintenance expenses and avoid downtime. Data curation, data pre-processing, diagnosis, and decision-making are the critical aspects of predictive maintenance [8]. This data-driven approach has been proven effective due to the vast availability of data and intelligent algorithms for automated analysis [9].
Artificial intelligence plays a promising role in fault prediction and predictive maintenance. Several machines and deep learning algorithms are trained on continuous data from sensors attached to the target machinery [10]. The proposed work uses machine and deep learning algorithms for anomaly detection (air and oil failure) on the air production unit of metro trains and real-time dashboard development, which is the first of this work as per our knowledge. The following section explains the existing works on machine usage and deep learning algorithms for predictive maintenance.
The following are the contributions of the proposed work:
1. To develop a deep learning algorithm for the simultaneous identification of the type and location of the fault, along with GPS quality monitoring from sensor data in metro trains.
2. To integrate an explainable AI technique into the model's prediction and highlight the key sensors contributing to the fault.
3. To develop a dashboard integrating the sensor data analytics as visual graphs, the deep learning model, and the explainable AI results for analysis by engineers.
This work will be of great aid to maintenance engineers for fault analysis in sensor values and assessment of GPS signals in real time. This predictive maintenance application can aid in reducing the downtime and service costs of machine parts, if found damaged.
The following is the outline of the research paper. Section 2 briefs about the existing work relevant to the field of interest. Section 3 explains the material description, the pre-processing techniques used, the different machine learning algorithms, the training parameters, and the methodologies for dashboard development. Section 4 represents the proposed methods' results, graphs, and supporting diagrams. Section 5 compares the proposed work with existing works and draws significant conclusions and future scope.
The detection of air leakage from the pneumatic door of a train is an attempt to reduce train downtime. Deep learning algorithms were applied for automatic feature extraction from extensive data obtained from continuous monitoring by sensors for the task of fault detection [11]. The OSR (open set recognition) concept was used for multi-task classification to predict the known class and detect unknown samples. A lightweight convolutional neural network (CNN) model streamlined with the OSR technique was trained to predict the air leakage. An 8-layer neural network consisting of 6 convolutions and two dense layers was used for air leakage. This model was trained using an SGD (stochastic gradient descent) optimizer with a learning rate of 0.001 for a batch size of 64.
The server air leakage in the breaking pipe results in breaking issues and decreases the train reliability [12]. Due to the visual constraints for air leak detection, the paper proposes a framework for the simultaneous prediction of the type and severity of air leakage using anomaly detection methods based on the on-and-off logs of the compressor. Around 632,683 data points were collected from May 2016 to October 2016 from 178 VIRM trains, of which 6957 are labeled as "Air Leakage" and 625,726 are labeled as "Normal". They have used a logistic classifier model for two different classes of compressor behavior for each separate train. One defines the boundary by separating two classes under everyday situations, and the other models the distribution of the compressor idle time and run time separately using logistic functions. It also further detects the context of compressor idle time erroneously classified as a compressor run time, aiding in anomaly detection. A density-based unsupervised clustering approach is adopted for anomaly detection before four weeks and can pre-filter anomalies to prevent false alarms.
The challenges encountered by traditional manufacturing companies during their transition to intelligent factories, notably the scarcity of historical data for training machine learning models, were addressed by Mohan Rajashekarappa et al. [13]. A novel approach of artificially inducing anomalies for data labeling was introduced, and it underscored the importance of proactive readiness for potential future disruptions in newly installed systems. Through two experiments focused on air leakage detection, the proposed methodology demonstrates exceptional performance with RUS-Boosted bagged trees, yielding 98.73% accuracy, 99.40% precision, recall of 99.21%, and an F1 score of 99.30% on the test data.
The critical issue of energy efficiency and fault detection in air conditioning systems emphasizes their intricate nature and substantial energy consumption impact [14]. The study comprises two essential components: First, it investigates the ramifications of various faults within the air conditioning system on its coefficient of performance (COP), shedding light on the potential energy wastage associated with these faults. The research convincingly demonstrates that different faults lead to varying degradation levels in the COP. Second, the paper evaluates the effectiveness of three supervised learning classifier models in classifying these faults: deep learning, support vector machine (SVM), and multi-layer perceptron (MLP). The research assesses the performance of these classifiers across six distinct fault classes, revealing that different faults indeed exert varying negative impacts on the COP.
Predicting air failure of the air production unit (APU) in metro trains. The dataset used for this task was MetroPT, a 6-month analysis of metro trains in Portugal comprising analog, digital, and GPS sensors [15]. The GPS information was excluded from the dataset, and the timestamp was encoded using the label encoding technique. A random forest classifier algorithm was used for the multi-class classification of air failure prediction. The data was undersampled and then split into training and testing sets. A feature importance visualization technique was employed to identify the root cause of the air failure. The random forest classifier produced 85% and 97% accuracies for the binary and multi-class classification tasks, respectively.
A deep learning neural network for anomaly detection in metro trains was developed by Davari et al. [16]. The algorithms used for this task were the sparse autoencoder and variational autoencoder. This work is an unsupervised learning approach for anomaly detection of air failures in trains. The MetroPT dataset was used for this work with sensors placed in the air production unit. The two versions of the autoencoder were used for sensor data reconstruction, and a low-pass filter was used to perform anomaly detection and detect faults. The autoencoder algorithms using the digital data produced precision, recall, and F1 scores of 44%, 13%, and 32% better than that of the algorithms trained on the analog data.
An expert system for the multi-objective optimization of equipment was developed for highway optimization by Ali et al. [17]. The particle swarm optimization was used to simultaneously optimize the time, cost, and quality of the equipment for construction. This method reduced the time and cost by 35.4% and 39.1%, respectively. The application of predictive maintenance in concrete manufacturing was done by Alshboul et al. [18]. Seven different classification algorithms were used, out of which the cat boost classifier produced an F1-score of 0.985, an accuracy of 0.984, a recall of 0.983, and a ROC curve area of 0.984. A comparative analysis of machine learning algorithms for concrete strength estimation was performed by Alshboul et al. [19]. Three machine learning algorithms, namely, XGBoost, LighGBM, and genetic programming, were used, out of which the LightGBM and XGBoost algorithms surpassed other studied algorithms with a coefficient of determination of 95.74% and 93.27%, respectively.
The proposed work aims to develop a decision process for failure prediction and identification of the failure type and location using machine learning and multi-task models using deep learning algorithms.
The proposed work adopts the following workflow consisting of different blocks: data acquisition, data pre-processing, feature pre-processing, visualization, model development, validation, and deployment. Figure 1 represents the proposed workflow visually.
The dataset used for this work is named MetroPT [17,18], comprising of sensor information related to urban metro trains in Portugal collected during the year 2022. The dataset comprises different analogue, digital, and GPS sensors continuously capturing data from the metro trains for six months. The MetroPT dataset has been curated to develop AI algorithms for automated fault prediction and predictive maintenance of metro trains based on sensor data. Table 1 represents the different sensors, their description, and the unit of measurement used for acquiring the MetroPT dataset.
Name | Description | Type of sensor | Unit |
TP2 | Compressor pressure | Analog | Bar |
TP3 | Pneumatic panel pressure | Analog | Bar |
H1 | Pressure of the valve that is activated when the pressure exceeds 10.2 bar | Analog | Bar |
DV_Pressure | Pressure drop due to water discharge by air dryers | Analog | Bar |
Reservoirs | Air tank pressure | Analog | Bar |
Oil temperature | Temperature of oil in compressor | Analog | Celsius |
Flowmeter | Airflow | Analog | m3/h |
Motor current | Current flowing in the motor | Analog | Ampere |
Comp | Electric signal of the compressor based on the air intake | Digital | - |
DV Electric | Electric signal of compressor outlet | Digital | - |
Towers | Specifies the two towers based on the action of air drying | Digital | - |
MPG | Trigger to start the compressor when the pressure is less than 8.2 bar | Digital | - |
LPS | Trigger when the pressure is less than 7 bar | Digital | - |
The pressure switch | Trigger when pressure is detected in the pilot control valve | Digital | - |
Oil level | Trigger when the oil level is less than the threshold | Digital | - |
Caudal Impulses | Trigger for the air flowmeter | Digital | - |
GPS Longitude | Longitude position | Analog | ° |
GPS Latitude | Latitude position | Analog | ° |
GPS Speed | Speed | Analog | Km/h |
The dataset comprises around 15 million sensor data records from Jan 1, 2022 to Jun 30, 2022. The dataset obtained from the source is not directly labeled. The dataset owners have provided information about the failure, like the start time, end time, type, and location of the failure. Based on the start and end times, the appropriate timestamps were found and the values between those timestamps were coded according to the fault type. Table 2 shows the statistical description of the dataset for the parameters, namely, TP2, TP3, H1, DV_pressure (DVP), Reservoirs, Oil_temperature (OT), Flowmeter, Motor_current (MoC), COMP, DV Electric (DVE), Towers, MPG, LPS, Pressure_switch (PrS), Oil_level (OL), Caudal_impulses (CaI), GPS Longitude (GPSLong), GPS Latitude (GPSLat), GPS Speed, GPS Quality, month, day, hour, minute, second. The statistical values of mean, standard deviation (std), minimum value, 25%, 50%, 75%, and maximum values for a dataset are evaluated here. Table 3 represents the label code for the different types of faults occurring in the metro train.
TP2 | TP3 | H1 | DVP | Reservoirs | OT | Flowmeter | MoC | |
mean | 0.947 | 8.989 | 8.038 | -0.019 | 1.63 | 65.843 | 20.128 | 2.040 |
std | 2.836 | 0.667 | 2.846 | 0.185 | 0.064 | 5.931 | 3.578 | 2.198 |
min | -0.03 | 0.937 | -0.033 | -0.036 | 1.349 | 18.575 | 18.8347 | -0.009 |
25% | -0.009 | 8.492 | 8.332 | -0.0279 | 1.608 | 61.825 | 18.9748 | 0.0024 |
50% | -0.007 | 8.996 | 8.876 | -0.025 | 1.635 | 66.475 | 19.03 | 0.007 |
75% | -0.006 | 9.506 | 9.438 | -0.025 | 1.667 | 70.575 | 19.040 | 3.837 |
max | 10.806 | 10.38 | 10.368 | 8.11 | 1.791 | 80.174 | 37.008 | 9.3375 |
COMP | DVE | Towers | MPG | LPS | PrS | OL | CaI | |
mean | 0.892 | 0.107 | 0.946 | 0.892 | 0.004 | 0.0 | 0.0 | 0.002 |
std | 0.309 | 0.309 | 0.225 | 0.309 | 0.068 | 0.0 | 0.0 | 0.047 |
min | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
25% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
50% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
75% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
max | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 1.0 |
GPSLong | GPSLat | GPSSpeed | GPSQuality | |||||
mean | -7.880 | 37.578 | 8.592 | 0.912 | ||||
std | 2.443 | 11.649 | 14.096 | 0.282 | ||||
min | -8.69 | 0.0 | 0.0 | 0.0 | ||||
25% | -8.66106 | 41.1696 | 0.0 | 1.0 | ||||
50% | -8.658 | 41.1858 | 0.0 | 1.0 | ||||
75% | -8.583 | 41.212 | 16.0 | 1.0 | ||||
max | 0.0 | 41.240 | 286.0 | 1.0 | ||||
month | day | hour | minute | second | ||||
mean | 1.0 | 16.046 | 13.139 | 29.507 | 29.499 | |||
std | 0.0 | 8.919 | 6.444 | 17.318 | 17.318 | |||
min | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | |||
25% | 1.0 | 8.0 | 9.0 | 15.0 | 14.0 | |||
50% | 1.0 | 16.0 | 14.0 | 30.0 | 29.0 | |||
75% | 1.0 | 24.0 | 19.0 | 45.0 | 44.0 | |||
max | 1.0 | 31.0 | 23.0 | 59.0 | 59.0 |
Label Code | Corresponding fault |
0 | Air leak in air dryer |
1 | Air leak in client chamber |
2 | Oil leak in compressor |
3 | No fault |
Timestamp data cannot be processed by machine learning and deep learning algorithms. Hence, it needs to be processed. This issue was tackled by extracting the month, week, day, hour, minute, and second information from the timestamp using the Pandas functionalities. Finally, the timestamp feature is removed from the dataset. Two new columns, one for the type of failure and another for the location of the failure, were created from the labels. These columns are created as these are the labels for the multi-task model. Table 4 represents the different failure types and location codes.
Label Code | Type of failure | Location of failure |
0 | Air leak | Air dryer |
1 | Oil leak | Client chamber |
2 | No failure | Compressor |
3 | Not reported | No location |
The final step in the data pre-processing was the operation of undersampling, especially in the dataset of stage 1. The stage-1 dataset comprises 2, 12,104 samples under the fault condition, about 1.63% of the entire data. Around 98.4% of the data belongs to the regular class, proving the dataset is highly imbalanced. Hence, the samples from the regular class were undersampled. Around 28, 00,000 randomly selected samples were taken from the entire dataset, which was used as the data for the first stage. The final dataset comprises about 25 input features and three target vectors. Table 5 represents the description of all the targets in the final dataset.
Name of target vector | Description | Classes |
Type | Code for the type of failure | 0- Air failure |
1- Oil failure | ||
2- Normal | ||
Location | Code for the location of the failure | 0- Air dryer |
1- Client | ||
2- Compressor | ||
3- No location | ||
GPS quality | Quality of the GPS sensor | 0- Good |
1- Bad |
The dataset has no null or duplicate values since the dataset is obtained from sensors that continuously monitor the trains. The feature scaling technique, normalization, was adopted to bring all the features to the same scale (0-1). Appropriate data visualization techniques were used for the dataset's univariate, bivariate, and multivariate analysis. The dataset comprises continuous (analogue sensors) and categorical features (digital sensors); we have split them for data visualization. A histogram with a kernel density estimator function is used to visualize the continuous features in the dataset. A pie chart is used to visualize the categorical features. Finally, the information was visualized using a map that represents the train's route along with the train's speed. Figures 2, 4, 3, and 5 represent the visualization plots obtained from the continuous, categorical, GPS, and entire dataset, respectively. After visualization, a stratified split of ratio 80:20 was made for the training and testing sets, respectively.
Multi-task learning is the ability of a neural network to simultaneously obtain multiple outputs from a single input. In this work, based on the sensor information, the multi-task neural network is designed to simultaneously predict the type and location of the failure in the metro train. The multi-task neural network has common pre-processing layers, followed by branches, each corresponding to a particular task. Hence, the multi-task neural network uses parallel processing to simultaneously identify the type and location of the fault and assess the GPS quality.
The multi-task neural network comprises shared layers and task-specific layers. The input layer, a single hidden layer with four neurons, and regularization layers like dropout and batch normalization are common for both tasks. In contrast, there are individual output layers for each task. The output layer comprises three neurons for the task of failure identification, one neuron for the task of location identification, and one neuron for the task of GPS quality identification. Figure 6 represents the architecture diagram of the proposed multi-task model.
The abovementioned multi-task neural networks were trained using the Adam optimizer and a combination of categorical and binary cross-entropy loss functions. The batch size was set to 5000, and the models were trained for 20 epochs. A hybrid loss function was used to train the multi-task model since two tasks (fault type and location identification) were multi-class. In contrast, the task of GPS quality identification is binary. Equation 3.1 represents the hybrid loss function used to train the multi-task model.
LF=2∑i=11Nn∑JM∑k=1−ykjlog(p(ykj)i−1NN∑L2∑p=1ylplogPlp | (3.1) |
Evaluation is an essential component of the proposed workflow. Model evaluation is done to identify the model's performance on the testing set. Evaluation of the testing set is essential to identify if the trained model has overfit or underfit. The following performance metrics are derived from the confusion matrix.
The confusion matrix comprises four values: true positive, true negative, false positive, and false negative. The diagonal elements of the matrix indicate the correctly classified samples (true positive and true negative), and the non-diagonal elements indicate the misclassified samples (false positive and false negative). The following are the different metrics used to evaluate the trained models.
Accuracy is defined as the ratio of the correct classifications to that of the total classifications. Accuracy is considered the gold standard metric for the evaluation of classification algorithms. The formula for accuracy is mentioned in Equation 3.2.
Accuracy=TP+TNTP+TN+FP+FN | (3.2) |
Precision is defined as the ratio of true positives to that of the total positives. Precision is one of the metrics used to analyze a model's performance in class imbalance conditions. The formula for precision is mentioned in Equation 3.3.
Precision=TPTP+FP | (3.3) |
Recall is defined as the ratio of true positives to that of total samples. Recall is another metric that is used to analyze multi-class classification under the condition of class imbalance. Equation 3.4 represents the mathematical formula for recall.
Recall=TPTP+FN | (3.4) |
AUC-ROC is expanded as the area under the regional operating characteristic curve. The ROC curve is the plot between the false positive and true positive values. The area under that curve is termed an AUC score. An AUC value of less than 0.5 is considered a terrible score, a score of 0.5 is considered a random guess, and a score of more than 0.8 is considered a good score.
Model interpretability has been a focus and requirement for which the demand has risen in recent years. Many machine learning and deep learning algorithms are considered black boxes, providing output for input data without any logical interpretations. This is needed in areas like healthcare, where life-concerning critical decisions are made. In such instances, providing interpretability to the model by providing explanations of the predictions can greatly aid clinicians.
In this work, the local interpretable model agnostic explanations (LIME) [19] is used to derive the interpretations of the complex ensemble classifier. As the name suggests, LIME works locally, meaning it works on individual data samples. Also, this method is model agnostic, meaning it works for all models. LIME works by mapping a simple interpretable model (like linear regression) on a complex model [20]. The local region of the data space is considered, where synthetic samples are generated based on original samples. These synthetic samples are labeled based on the prediction of the complex model. Then, a simple interpretable linear regression is trained on the synthetic labeled data. The coefficients of the trained linear regression model represent the interpretations of the complex model on the local space. The LIME tabular function from the LIME library is used to get the interpretations for individual data samples.
The LIME function explains the type of fault, location of the fault, and quality of the GPS sensor, respectively. The tabular explainer from LIME was used to explain the predicted instances. A LIME explainer was used for each task: failure type, failure location, and GPS quality. The LIME explainer generates a figure that shows the input that contributed to that particular class (positive) and the features that contributed to the counter-class (negative). Hence, we can understand the features that positively and negatively contribute to a particular class from the plot.
The ultimate aim of the work is to develop a dashboard for real-time data analytics and predictions. A website was developed using a Python-based web development tool. The website will request the data recorded from the sensors as an Excel sheet of CSV (comma-separated values). The data visualization and prediction tasks are done simultaneously upon receiving the data. On the data visualization task, a stacked area chart of the continuous features, a stacked bar chart for the categorical features, and a map depicting the speed and route of the train are made. For the data prediction task, the latest values are processed and sent to the trained multi-task model for predictions on the failure type, failure location, and quality of the GPS, which are displayed on the website. Finally, the LIME explanations for the model on the provided sample are given for all three tasks: failure type identification, failure location identification, and GPS quality assessment.
The developed multi-task model was trained on the training set with the above-mentioned epochs and batch size. Table 6 represents the performance metrics of the model on the training and testing datasets for the fault type and location identification, respectively.
Set | Time consumption per epoch (ms) | Metric | Type identification | Location identification | GPS quality identification |
Training | 43 | Loss | 0.0031 | 0.0020 | 0.0029 |
Accuracy | 99.94 | 99.998 | 99.99 | ||
Precision | 99.99 | 100 | 100 | ||
Recall | 100 | 100 | 100 | ||
AUC | 1 | 1 | 1 | ||
Testing | 1 | Loss | 0.0035 | 0.0026 | 0.0033 |
Accuracy | 98.89 | 99.12 | 99.24 | ||
Precision | 99.56 | 99.67 | 99.84 | ||
Recall | 99.92 | 99.93 | 99.93 | ||
AUC | 1 | 1 | 1 |
The trained multi-task ANN model produced 98.89%, 99.12% and 99.24% accuracy for failure type identification, failure location identification and GPS quality assessment. Also, the trained model's precision, recall and AUC values are high, indicating that the model overcomes class imbalance issues.
The performance plots representing the values of the performance metrics for each epoch in the training phase are shown in Figure 7. The loss gradually decreases in each task's epoch for fault, location, and GPS quality, as shown in Figure 7(a). In contrast, the recall, AUC, and precision are increasing for each epoch, as shown in Figure 7(b), Figure 7(c), and Figure 7(d), respectively. This shows no fluctuations in the training phase and no signs of varying gradients.
The performance plots represent the values of the performance metrics for each epoch in the testing phase. The loss gradually decreases in each task's epoch, as shown in Figure 8(a). In contrast, the recall, AUC, and precision are increasing for each epoch for fault, location, and GPS quality as shown in Figure 8(b), Figure 8(c), and Figure 8(d). The precision, recall, and AUC values for GPS quality reached the maximum in the initial epochs and remained the same for the rest. Also, fluctuations in the precision values are observed for the tasks of fault type and fault location identification.
The confusion matrices on the testing set represent that the trained model has produced high true negatives and true positives. In contrast, it has produced very few false positives and false negatives. The confusion matrix for the fault type is shown in Figure 9(a), The fault location confusion matrix is shown in Figure 9(b) and the confusion matrix for GPS quality is shown in Figure 9(c).
The classification report for fault type classification indicates high precision, recall, and F1 scores for all three classes, showing no sign of class imbalance. The classification report for the trained multi-task model for fault type identification on the testing set is shown in Table 7. The classification report for the trained multi-task model for fault location identification on the testing set is shown in Table 8. The classification report for the trained multi-task model for GPS quality assessment on the testing set is shown in Table 9.
Classes | Precision | Recall | F1-Score |
0 | 0.98 | 1.00 | 0.99 |
1 | 1.00 | 1.00 | 1.00 |
2 | 1.00 | 1.00 | 1.00 |
Classes | Precision | Recall | F1-Score |
0 | 1 | 1 | 1 |
1 | 1 | 1 | 1 |
2 | 1 | 1 | 1 |
3 | 1 | 1 | 1 |
Classes | Precision | Recall | F1-Score |
0 | 1 | 1 | 1 |
1 | 1 | 1 | 1 |
A website was developed using Streamlit and hosted online using the Streamlit share, see [21] for the website URL. Figures 10 and 11 represent the snips of the developed website about data visualization and predictive modeling. Figures 12, 13, and 14 represent the LIME explanations for the tasks of failure type identification, failure location identification, and GPS quality assessment, respectively. Figure 10 represents the area chart for the continuous features and the bar chart for the categorical features. From this bar, anomalies in the sensor values can be visually identified. Figure 11 represents the map plot for the GPS data; the dots represent the map's route based on the latitude and longitude values, whereas the dot's intensity depicts the speed. This graph identifies the train's route and the crucial locations at which the train went fast/slow. Also, the multi-task neural network predictions are mentioned on the website for each dataset instance.
Table 10 compares the performance of the proposed multi-task neural network to that of the existing works related to the application of predictive maintenance in train fault analysis using sensor data.
Author | Algorithm | Result |
Rajashekarappa et al. [13] | RUS Boosted Classifier | 98.73% accuracy |
Najjar et al. [15] | Random Forest Classifier | 97% accuracy |
Davari et al. [16] | Autoencoder | 44% improvement in precision compared to baseline |
Proposed work | Multi-task Artificial Neural Network | 99% average accuracy |
Based on our knowledge, only two works [15,16] use artificial intelligence for the predictive maintenance of urban metro trains. Najjar et al. [15] worked on predicting air failure of the air production unit (APU) in metro trains. The dataset used for this task was MetroPT, a 6-month analysis of metro trains in Portugal comprising analog, digital, and GPS sensors. The GPS information was excluded from the dataset, and the timestamp was encoded using the label encoding technique. A random forest classifier algorithm was used for the multi-class classification of air failure prediction. The data was undersampled and then split into training and testing sets. A feature importance visualization technique was employed to identify the root cause of the air failure. The proposed work produced testing accuracies of 84% and 87% on the binary and multi-class classification tasks and F1 scores in the ranges of 0.83–0.5 and 0.73–0.97 for the binary and multi-class classification tasks. The proposed model produced better results than the work and has also included oil failures in addition to air failures. Davari et al. [16] developed a deep learning neural network for anomaly detection in metro trains. The algorithms used for this task were the sparse autoencoder and variational autoencoder. This work is an unsupervised learning approach for anomaly detection of air failures in trains. The sparse autoencoder trained on the digital data produced 42% more than those trained on analog data. Also, the variational autoencoder performed better than the sparse autoencoder by 37%. The proposed work considered both analog, digital, and GPS sensors worked on both air and oil failures, and produced state-of-the-art results.
The multi-task model has produced excellent results on the training and testing datasets. The performance plots prove that the model has been trained perfectly and shows no signs of overfitting or underfitting. Also, the confusion matrices and classification report suggest that the trained model is generalized and does not exhibit any signs of class imbalance.
The proposed work has some advantages in comparison to the existing works. The proposed work addresses all issues faced in the metro trains (air and oil failures). Another advantage is the excellent results obtained by the trained models. The third advantage is agility; the proposed multi-task model has taken less time to train and predict batch data. Finally, the explainable AI technique, namely LIME, has been implemented to provide interpretations to the outputs given by the multi-task model. This provides belief in the model prediction and can also be useful for engineers to deeply analyze the issue.
Figure 12 represents the LIME explanation for a local instance related to the task of fault type prediction. As observed from the plot, the predicted instance is air failure with 100% confidence, and the features positively and negatively contribute to the prediction.
Figure 13 represents the LIME explanation for a local instance related to the task of fault location prediction. As observed from the plot, the predicted instance is a client with 100% confidence, and the features positively and negatively contribute to the prediction. Motor current, DV pressure, and the day positively contributed to the prediction, followed by GPS speed and reservoir. The GPS longitude, minute, and flowmeter have negatively contributed to the prediction.
Figure 14 represents the LIME explanation for a local instance related to the task of GPS quality prediction. As observed from the plot, the predicted instance is air failure with 100% confidence, and the features positively and negatively contribute to the prediction. Oil temperature has majorly contributed positively, whereas H1, day, and minute have majorly contributed negatively. For all plots, the range or condition of the input features is given, which might be of great use for the fault analysis.
One reason for achieving good results is the split of the output labels into fault type and location. This reduces the number of interdependent classes in each stage, which might have improved the performance of the algorithms. Also, we observed a performance rise of 6% when the features were standardized. Considering the deep learning aspects, we developed a multi-task model which splits the classes into individual tasks, allowing for more attention, resulting in better results and quicker periods.
However, the study has some limitations. First, the dataset was undersampled to 30, 00,000 data points, roughly 20% of the entire dataset. The second one was the expansion of the target vectors into new columns, which introduced more computations and the need to train more models. While this approach was considered to improve the holistic performance of the models, this resulted in the creation of multiple datasets and, ultimately, multiple ML models for training, leading to computational costs. The third one was the limited selection of machine learning algorithms. Many good machine learning algorithms like support vector machine and ensemble learning techniques were not implemented due to the computational constraints and long training durations (the SVM algorithm did not train even after 30 minutes!).
Hence in this work, a multi-task model was developed for the identification of failures simultaneously. The proposed method has produced 98.89%, 99.12%, and 99.24% accuracies in the testing set for failure type, failure location, and GPS quality predictions, respectively, exceeding the state-of-the-art methods. The model produced 99.56%, 99.67%, and 99.84% precision in the testing set for failure type, failure location, and GPS quality predictions, respectively. The high accuracy and precision values indicate the good performance of the model and no signs of class imbalance. The deep learning model took 43 seconds for training and 1 second for inferencing for test data, showing fast predictions, needed for predictive maintenance applications. Moreover, a real-time interactive dashboard was developed, performing dynamic data visualization and predictions. Finally, using the LIME explainable AI technique provides explanations for the predictions, adding belief and better analysis for engineers. The developed system would be advantageous for engineers to perform fault analysis and predictive maintenance effectively. Future work will use a database to store the streaming data and deploy the system in real time. Also, we will develop deep learning algorithms on the entire dataset and employ online learning strategies to update the learned model in real time.
Pratik Vinayak Jadhav: Data Curation and Analysis, Research Design and Methodology, writing draft. Sairam V. A: Data Curation and Analysis, Research Design and Methodology, writing draft. Siddharth Sonkavade: Data Curation and Analysis, Research Design and Methodology, writing draft. Shivali Amit Wagle : Conceptualization, Supervision, Project Development, Writing, Review, and Editing. Preksha Pareek: Conceptualization, Supervision, Project Development, Writing, and Review. Ketan Kotecha: Writing, Review, and Editing, Funding and Resources. Tanupriya Choudhury: Data Analysis, Writing, Review, and Editing.
The authors declare no conflict of interest.
[1] | Tijani JO, Fatoba OO, Petrik LF (2013) A review of pharmaceuticals and endocrine-disrupting compounds: Sources, effects, removal, and detections. Water Air Soil Pollut 224. |
[2] |
Gerbersdorf SU, Cimatoribus C, Class H, et al. (2015) Anthropogenic Trace Compounds (ATCs) in aquatic habitats-Research needs on sources, fate, detection and toxicity to ensure timely elimination strategies and risk management. Environ Int 79: 85-105. doi: 10.1016/j.envint.2015.03.011
![]() |
[3] |
Burke V, Richter D, Greskowiak J, et al. (2016) Occurrence of Antibiotics in Surface and Groundwater of a Drinking Water Catchment Area in Germany. Water Environ Res 88: 652-659. doi: 10.2175/106143016X14609975746604
![]() |
[4] |
Schwarzenbach RP (2006) The Challenge of Micropollutants in Aquatic Systems. Science 313: 1072-1077. doi: 10.1126/science.1127291
![]() |
[5] |
Luo Y, Guo W, Ngo HH, et al. (2014) A review on the occurrence of micropollutants in the aquatic environment and their fate and removal during wastewater treatment. Sci Total Environ 473-474: 619-641. doi: 10.1016/j.scitotenv.2013.12.065
![]() |
[6] |
Haddad T, Baginska E, Kümmerer K (2015) Transformation products of antibiotic and cytostatic drugs in the aquatic cycle that result from effluent treatment and abiotic/biotic reactions in the environment: An increasing challenge calling for higher emphasis on measures at the beginning of the pi. Water Res 72: 75-126. doi: 10.1016/j.watres.2014.12.042
![]() |
[7] |
Rivera-Utrilla J, Sánchez-Polo M, Ferro-García MÁ, et al. (2013) Pharmaceuticals as emerging contaminants and their removal from water. A review. Chemosphere 93: 1268-1287. doi: 10.1016/j.chemosphere.2013.07.059
![]() |
[8] |
Kümmerer K (2009) Antibiotics in the aquatic environment-a review-part I. Chemosphere 75: 417-434. doi: 10.1016/j.chemosphere.2008.11.086
![]() |
[9] |
Novo A, Andre S, Viana P, et al. (2013) Antibiotic resistance, antimicrobial residues and bacterial community composition in urban wastewater. Water Res 47: 1875-1887. doi: 10.1016/j.watres.2013.01.010
![]() |
[10] |
Margot J, Rossi L, Barry DA, et al. (2015) A review of the fate of micropollutants in wastewater treatment plants. Wiley Interdiscip Rev Water 2: 457-487. doi: 10.1002/wat2.1090
![]() |
[11] |
Sui Q, Cao X, Lu S, et al. (2015) Occurrence, sources and fate of pharmaceuticals and personal care products in the groundwater: A review. Emerg Contam 1: 14-24. doi: 10.1016/j.emcon.2015.07.001
![]() |
[12] |
Du B, Price AE, Scott WC, et al. (2014) Science of the Total Environment Comparison of contaminants of emerging concern removal, discharge, and water quality hazards among centralized and on-site wastewater treatment system ef fl uents receiving common wastewater in fl uent. Sci Total Environ 466-467: 976-984. doi: 10.1016/j.scitotenv.2013.07.126
![]() |
[13] | Decision 2018/840 (2018) Commission Implementing Decision (EU) 2018/840 of 5 June 2018 establishing a watch list of substances for Union-wide monitoring in the field of water policy pursuant to Directive 2008/105/EC of the European Parliament and of the Council and repealing Comm. |
[14] | Directive 2000/60/EC (2000) Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in the field of water policy. |
[15] | Directive 2008/105/EC Environmental quality standards applicable to surface water. |
[16] | Directive 2013/39/EU Directive 2013/39/EU of the European Parliament and of the Council of 12 August 2013 amending Directives 2000/60/EC and 2008/105/EC as regards priority substances in the field of water policy Text with EEA relevance. |
[17] | Decision 2015/495/EU Commission Implementing Decision (EU) 2015/495 of 20 March 2015 establishing a watch list of substances for Union-wide monitoring in the field of water policy pursuant to Directive 2008/105/EC of the European Parliament and of the Council (notified under. |
[18] | aus der Beek T, Weber FA, Bergmann A, et al. (2016) Pharmaceuticals in the environment: Global occurrence and potential cooperative action under the Strategic Approach to International Chemicals Management (SAICM). UBA Texte 67/2016. |
[19] |
Gogoi A, Mazumder P, Kumar V, et al. (2018) Groundwater for Sustainable Development Occurrence and fate of emerging contaminants in water environment : A review. Groundw Sustain Dev 6: 169-180. doi: 10.1016/j.gsd.2017.12.009
![]() |
[20] |
Liu J, Dan X, Lu G, et al. (2018) Investigation of pharmaceutically active compounds in an urban receiving water: Occurrence, fate and environmental risk assessment. Ecotoxicol Environ Saf 154: 214-220. doi: 10.1016/j.ecoenv.2018.02.052
![]() |
[21] |
Zhang R, Zhang R, Yu K, et al. (2018) Occurrence, sources and transport of antibiotics in the surface water of coral reef regions in the South China Sea: Potential risk to coral growth. Environ Pollut 232: 450-457. doi: 10.1016/j.envpol.2017.09.064
![]() |
[22] |
Jelic A, Gros M, Ginebreda A, et al. (2011) Occurrence, partition and removal of pharmaceuticals in sewage water and sludge during wastewater treatment. Water Res 45: 1165-1176. doi: 10.1016/j.watres.2010.11.010
![]() |
[23] |
Biel-Maeso M, Baena-Nogueras RM, Corada-Fernández C, et al. (2018) Occurrence, distribution and environmental risk of pharmaceutically active compounds (PhACs) in coastal and ocean waters from the Gulf of Cadiz (SW Spain). Sci Total Environ 612: 649-659. doi: 10.1016/j.scitotenv.2017.08.279
![]() |
[24] |
Mirzaei R, Yunesian M, Nasseri S, et al. (2018) Occurrence and fate of most prescribed antibiotics in different water environments of Tehran, Iran. Sci Total Environ 619-620: 446-459. doi: 10.1016/j.scitotenv.2017.07.272
![]() |
[25] | Paredes L, Omil F, Lema JM, et al. (2018) What happens with organic micropollutants during UV disinfection in WWTPs? A global perspective from laboratory to full-scale. J Hazard Mater 342: 670-678. |
[26] |
Kostich MS, Batt AL, Lazorchak JM (2014) Concentrations of prioritized pharmaceuticals in effluents from 50 large wastewater treatment plants in the US and implications for risk estimation. Environ Pollut 184: 354-359. doi: 10.1016/j.envpol.2013.09.013
![]() |
[27] |
Orias F, Perrodin Y (2013) Characterisation of the ecotoxicity of hospital effluents: A review. Sci Total Environ 454-455: 250-276. doi: 10.1016/j.scitotenv.2013.02.064
![]() |
[28] |
Botero-Coy AM, Martínez-Pachón D, Boix C, et al. (2018) 'An investigation into the occurrence and removal of pharmaceuticals in Colombian wastewater'. Sci Total Environ 642: 842-853. doi: 10.1016/j.scitotenv.2018.06.088
![]() |
[29] | Wang B, Dai G, Deng S, et al. (2017) Large-scale enzymatic membrane reactors for tetracycline degradation in WWTP effluents. Water Res 45: 1-11. |
[30] |
Tiedeken EJ, Tahar A, McHugh B, et al. (2017) Monitoring, sources, receptors, and control measures for three European Union watch list substances of emerging concern in receiving waters-A 20 year systematic review. Sci Total Environ 574: 1140-1163. doi: 10.1016/j.scitotenv.2016.09.084
![]() |
[31] |
Oulton RL, Kohn T, Cwiertny DM (2010) Pharmaceuticals and personal care products in effluent matrices: A survey of transformation and removal during wastewater treatment and implications for wastewater management. J Environ Monit 12: 1956. doi: 10.1039/c0em00068j
![]() |
[32] | Ahmed MB, Zhou JL, Ngo HH, et al. (2016) Progress in the biological and chemical treatment technologies for emerging contaminant removal from wastewater: A critical review. J Hazard Mater 323: 274-298. |
[33] |
Cecconet D, Molognoni D, Callegari A, et al. (2017) Biological combination processes for efficient removal of pharmaceutically active compounds from wastewater: A review and future perspectives. J Environ Chem Eng 5: 3590-3603. doi: 10.1016/j.jece.2017.07.020
![]() |
[34] | Bayona JM (2013) Removal of Pharmaceutical Compounds from Wastewater and Surface Water by Natural Treatments. 62. |
[35] |
Cole S (1998) The Emergence of Treatment Wetlands. Environ Sci Technol 32: 218A-223A. doi: 10.1021/es9834733
![]() |
[36] |
Verlicchi P, Zambello E (2014) How efficient are constructed wetlands in removing pharmaceuticals from untreated and treated urban wastewaters? A review. Sci Total Environ 470-471: 1281-1306. doi: 10.1016/j.scitotenv.2013.10.085
![]() |
[37] |
Fountoulakis MS, Terzakis S, Chatzinotas A, et al. (2009) Pilot-scale comparison of constructed wetlands operated under high hydraulic loading rates and attached biofilm reactors for domestic wastewater treatment. Sci Total Environ 407: 2996-3003. doi: 10.1016/j.scitotenv.2009.01.005
![]() |
[38] |
Hijosa-Valsero M, Fink G, Schlüsener MP, et al. (2011) Removal of antibiotics from urban wastewater by constructed wetland optimization. Chemosphere 83: 713-719. doi: 10.1016/j.chemosphere.2011.02.004
![]() |
[39] |
Hijosa-Valsero M, Matamoros V, Martín-Villacorta J, et al. (2010) Assessment of full-scale natural systems for the removal of PPCPs from wastewater in small communities. Water Res 44: 1429-1439. doi: 10.1016/j.watres.2009.10.032
![]() |
[40] | Dai Y, Dan A, Yang Y, et al. (2016) Factors Affecting Behavior of Phenolic Endocrine Disruptors, Estrone and Estradiol, in Constructed Wetlands for Domestic Sewage Treatment. EnvironSciTechnol 50: 11844-11852. |
[41] |
Song H, Nakano K, Taniguchi T, et al. (2009) Estrogen removal from treated municipal effluent in small-scale constructed wetland with different depth. Bioresour Technol 100: 2945-2951. doi: 10.1016/j.biortech.2009.01.045
![]() |
[42] |
Verlicchi P, Galletti A, Petrovic M, et al. (2013) Removal of selected pharmaceuticals from domestic wastewater in an activated sludge system followed by a horizontal subsurface flow bed-Analysis of their respective contributions. Sci Total Environ 454-455: 411-425. doi: 10.1016/j.scitotenv.2013.03.044
![]() |
[43] |
Song H-L, Yang X-L, Nakano K, et al. (2011) Elimination of estrogens and estrogenic activity from sewage treatment works effluents in subsurface and surface flow constructed wetlands. Int J Environ Anal Chem 91: 600-614. doi: 10.1080/03067319.2010.496046
![]() |
[44] |
Hu X, Bao Y, Hu J, et al. (2017) Occurrence of 25 pharmaceuticals in Taihu Lake and their removal from two urban drinking water treatment plants and a constructed wetland. Env Sci Pollut Res 24: 14889-14902. doi: 10.1007/s11356-017-8830-y
![]() |
[45] |
Kimura K, Hara H, Watanabe Y (2007) Elimination of selected acidic pharmaceuticals from municipal wastewater by an activated sludge system and membrane bioreactors. Environ Sci Technol 41: 3708-3714. doi: 10.1021/es061684z
![]() |
[46] |
Fazal S, Zhang B, Zhong Z, et al. (2015) Industrial Wastewater Treatment by Using MBR (Membrane Bioreactor) Review Study. J Environ Prot (Irvine, Calif) 06: 584-598. doi: 10.4236/jep.2015.66053
![]() |
[47] |
Carmosini N, Lee LS (2009) Ciprofloxacin sorption by dissolved organic carbon from reference and bio-waste materials. Chemosphere 77: 813-820. doi: 10.1016/j.chemosphere.2009.08.003
![]() |
[48] |
Zuehlke S, Duennbier U, Lesjean B, et al. (2006) Long-Term Comparison of Trace Organics Removal Performances Between Conventional and Membrane Activated Sludge Processes. Water Environ Res 78: 2480-2486. doi: 10.2175/106143006X111826
![]() |
[49] | Baumgarten S (2007) Membranbioreaktoren zur industriellen Abwasserreinigung. |
[50] |
Janssens R, Mandal MK, Dubey KK, et al. (2017) Slurry photocatalytic membrane reactor technology for removal of pharmaceutical compounds from wastewater: Towards cytostatic drug elimination. Sci Total Environ 599-600: 612-626. doi: 10.1016/j.scitotenv.2017.03.253
![]() |
[51] |
Abass OK, Wu X, Guo Y, et al. (2015) Membrane Bioreactor in China: A Critical Review. Int J Membr Sci Technol 2: 29-47. doi: 10.15379/2410-1869.2015.02.02.04
![]() |
[52] |
Besha AT, Gebreyohannes AY, Tufa RA, et al. (2017) Removal of emerging micropollutants by activated sludge process and membrane bioreactors and the effects of micropollutants on membrane fouling: A review. J Environ Chem Eng 5: 2395-2414. doi: 10.1016/j.jece.2017.04.027
![]() |
[53] |
Fan H, Li J, Zhang L, et al. (2014) Contribution of sludge adsorption and biodegradation to the removal of five pharmaceuticals in a submerged membrane bioreactor. Biochem Eng J 88: 101-107. doi: 10.1016/j.bej.2014.04.008
![]() |
[54] |
Kovalova L, Siegrist H, Singer H, et al. (2012) Hospital wastewater treatment by membrane bioreactor: Performance and efficiency for organic micropollutant elimination. Environ Sci Technol 46: 1536-1545. doi: 10.1021/es203495d
![]() |
[55] | Le T, Ng C, Tran NH, et al. (2018) Removal of antibiotic residues, antibiotic resistant bacteria and antibiotic resistance genes in municipal wastewater by membrane bioreactor systems. Water Res 498-508. |
[56] |
Zhang W, Ding L, Luo J, et al. (2016) Membrane fouling in photocatalytic membrane reactors (PMRs) for water and wastewater treatment: A critical review. Chem Eng J 302: 446-458. doi: 10.1016/j.cej.2016.05.071
![]() |
[57] |
Huang BC, Guan YF, Chen W, et al. (2017) Membrane fouling characteristics and mitigation in a coagulation-assisted microfiltration process for municipal wastewater pretreatment. Water Res 123: 216-223. doi: 10.1016/j.watres.2017.06.080
![]() |
[58] |
Liébana R, Arregui L, Belda I, et al. (2015) Membrane bioreactor wastewater treatment plants reveal diverse yeast and protist communities of potential significance in biofouling. Biofouling 31: 71-82. doi: 10.1080/08927014.2014.998206
![]() |
[59] | Oppenländer T (2003) Photochemical Purification of Water and Air: Advanced Oxidation Processes (AOPs): Principles, Reaction Mechanisms, Reactor Concepts (Chemistry), Weinheim, WILEY-VCH Verlag. |
[60] | Parsons S (2004) Advanced Oxidation Processes for Water and Wastewater Treatment, London, IWA Publishing. |
[61] |
Andreozzi R, Caprio V, Insola A, et al. (1999) Advanced oxidation processes (AOP) for water purification and recovery. Catal Today 53: 51-59. doi: 10.1016/S0920-5861(99)00102-9
![]() |
[62] |
Voigt M, Bartels I, Nickisch-hartfiel A, et al. (2018) Elimination of macrolides in water bodies using photochemical oxidation. AIMS Environ Sci 5: 372-388. doi: 10.3934/environsci.2018.5.372
![]() |
[63] |
Homem V, Santos L (2011) Degradation and removal methods of antibiotics from aqueous matrices--a review. J Environ Manage 92: 2304-2347. doi: 10.1016/j.jenvman.2011.05.023
![]() |
[64] |
Monteagudo JM, Durán A, San Martín I (2014) Mineralization of wastewater from the pharmaceutical industry containing chloride ions by UV photolysis of H2O2/Fe(II) and ultrasonic irradiation. J Environ Manage 141: 61-69. doi: 10.1016/j.jenvman.2014.03.020
![]() |
[65] |
Shahidi D, Roy R, Azzouz A (2015) Advances in catalytic oxidation of organic pollutants-Prospects for thorough mineralization by natural clay catalysts. Appl Catal B Environ 174-175: 277-292. doi: 10.1016/j.apcatb.2015.02.042
![]() |
[66] | Brillas E (2014) A review on the degradation of organic pollutants in waters by UV photoelectro-fenton and solar photoelectro-fenton. J Braz Chem Soc 25: 393-417. |
[67] |
Voigt M, Bartels I, Nickisch-Hartfiel A, et al. (2017) Photoinduced degradation of sulfonamides, kinetic, and structural characterization of transformation products and assessment of environmental toxicity. Toxicol Environ Chem 99: 1304-1327. doi: 10.1080/02772248.2017.1373777
![]() |
[68] |
Jelic a., Michael I, Achilleos a., et al. (2013) Transformation products and reaction pathways of carbamazepine during photocatalytic and sonophotocatalytic treatment. J Hazard Mater 263: 177-186. doi: 10.1016/j.jhazmat.2013.07.068
![]() |
[69] |
Vasconcelos TG, Henriques DM, König A, et al. (2009) Photo-degradation of the antimicrobial ciprofloxacin at high pH: Identification and biodegradability assessment of the primary by-products. Chemosphere 76: 487-493. doi: 10.1016/j.chemosphere.2009.03.022
![]() |
[70] | Suslick S, Fang M (1999) Acoustic cavitation and its chemical consequences. Phil Trans R Soc Lond A 335-353. |
[71] | Torres-Palma RA, Serna-Galvis EA (2018) Sonolysis, Advanced Oxidation Processes for Waste Water Treatment, Elsevier, 177-213. |
[72] |
Serna-Galvis EA, Botero-Coy AM, Martínez-Pachón D, et al. (2019) Degradation of seventeen contaminants of emerging concern in municipal wastewater effluents by sonochemical advanced oxidation processes. Water Res 154: 349-360. doi: 10.1016/j.watres.2019.01.045
![]() |
[73] |
Zoschke K, Börnick H, Worch E (2014) Vacuum-UV radiation at 185 nm in water treatment-a review. Water Res 52: 131-145. doi: 10.1016/j.watres.2013.12.034
![]() |
[74] |
Crapulli F, Santoro D, Sasges MR, et al. (2014) Mechanistic modeling of vacuum UV advanced oxidation process in an annular photoreactor. Water Res 64: 209-225. doi: 10.1016/j.watres.2014.06.048
![]() |
[75] | Heit G, Neuner A, Saugy P, et al. (1998) Vacuum-UV (172 nm) Actinometry. The Quantum Yield of the Photolysis of Water. J Phys Chem A 5639: 5551-5561. |
[76] | Ratpukdi T (2014) Degradation of Paracetamol and Norfloxacin in Aqueous Solution Using Vacuum Ultraviolet (VUV) Process. J Clean Energy Technol 2: 168-170. |
[77] |
Szabó RK, Megyeri C, Illés E, et al. (2011) Phototransformation of ibuprofen and ketoprofen in aqueous solutions. Chemosphere 84: 1658-1663. doi: 10.1016/j.chemosphere.2011.05.012
![]() |
[78] |
Voigt M, Jaeger M (2017) On the photodegradation of azithromycin, erythromycin and tylosin and their transformation products-A kinetic study. Sustain Chem Pharm 5: 131-140. doi: 10.1016/j.scp.2016.12.001
![]() |
[79] | Gottschalk C, Libra JA, Saupe A (2010) Ozonation of Water and Waste Water-A Practical Guide to Understanding Ozone and its Applications, Weinheim, WILEY-VCH Verlag GmbH & Co. KGaA. |
[80] |
Antoniou MG, Hey G, Rodríguez Vega S, et al. (2013) Required ozone doses for removing pharmaceuticals from wastewater effluents. Sci Total Environ 456-457: 42-49. doi: 10.1016/j.scitotenv.2013.03.072
![]() |
[81] |
Ribeiro AR, Nunes OC, Pereira MFR, et al. (2015) An overview on the advanced oxidation processes applied for the treatment of water pollutants defined in the recently launched Directive 2013/39/EU. Environ Int 75: 33-51. doi: 10.1016/j.envint.2014.10.027
![]() |
[82] |
Matilainen A, Sillanpää M (2010) Removal of natural organic matter from drinking water by advanced oxidation processes. Chemosphere 80: 351-365. doi: 10.1016/j.chemosphere.2010.04.067
![]() |
[83] |
Tong L, Eichhorn P, Pérez S, et al. (2011) Photodegradation of azithromycin in various aqueous systems under simulated and natural solar radiation: kinetics and identification of photoproducts. Chemosphere 83: 340-348. doi: 10.1016/j.chemosphere.2010.12.025
![]() |
[84] |
Frontistis Z, Kouramanos M, Moraitis S, et al. (2015) UV and simulated solar photodegradation of 17α-ethynylestradiol in secondary-treated wastewater by hydrogen peroxide or iron addition. Catal Today 252: 84-92. doi: 10.1016/j.cattod.2014.10.012
![]() |
[85] |
Ma X, Zhang C, Deng J, et al. (2015) Simultaneous degradation of estrone, 17β-estradiol and 17α-ethinyl estradiol in an aqueous UV/H2o2 system. Int J Environ Res Public Health 12: 12016-12029. doi: 10.3390/ijerph121012016
![]() |
[86] |
Voigt M, Savelsberg C, Jaeger M (2017) Photodegradation of the antibiotic spiramycin studied by high-performance liquid quadrupole time-of-flight mass spectrometry. Toxicol Environ Chem 99: 624-640. doi: 10.1080/02772248.2017.1280039
![]() |
[87] | Hernández F, Bakker J, Bijlsma L, et al. (2019) The role of analytical chemistry in exposure science: Focus on the aquatic environment. Chemosphere 564-583. |
[88] | Voigt M, Savelsberg C, Jaeger M (2018) Identification of Pharmaceuticals in The Aquatic Environment Using HPLC-ESI-Q-TOF-MS and Elimination of Erythromycin Through Photo-Induced Degradation. J Vis Exp. |
[89] | Bobu M, Yediler A, Siminiceanu I, et al. (2013) Comparison of different advanced oxidation processes for the degradation of two fluoroquinolone antibiotics in aqueous solutions. J Environ Sci Heal Part A 48: 251-262. |
[90] |
Yu F, Li Y, Han S, et al. (2016) Adsorptive removal of antibiotics from aqueous solution using carbon materials. Chemosphere 153: 365-385. doi: 10.1016/j.chemosphere.2016.03.083
![]() |
[91] |
Sharif F, Westerhoff P, Herckes P (2013) Sorption of trace organics and engineered nanomaterials onto wetland plant material. Environ Sci Process Impacts 15: 267-274. doi: 10.1039/C2EM30613A
![]() |
[92] |
Gao Y, Li Y, Zhang L, et al. (2012) Adsorption and removal of tetracycline antibiotics from aqueous solution by graphene oxide. J Colloid Interface Sci 368: 540-546. doi: 10.1016/j.jcis.2011.11.015
![]() |
[93] |
Mailler R, Gasperi J, Coquet Y, et al. (2016) Removal of emerging micropollutants from wastewater by activated carbon adsorption: Experimental study of different activated carbons and factors influencing the adsorption of micropollutants in wastewater. J Environ Chem Eng 4: 1102-1109. doi: 10.1016/j.jece.2016.01.018
![]() |
[94] | Kaub M, Biebersdorf N (2014) Kläranlage Höxter 4. Reinigungsstufe zur Elimination von Mikroschadstoffen, Bochum. |
[95] |
Grover DP, Zhou JL, Frickers PE, et al. (2011) Improved removal of estrogenic and pharmaceutical compounds in sewage effluent by full scale granular activated carbon : Impact on receiving river water. J Hazard Mater 185: 1005-1011. doi: 10.1016/j.jhazmat.2010.10.005
![]() |
[96] |
Kovalova L, Siegrist H, Von Gunten U, et al. (2013) Elimination of micropollutants during post-treatment of hospital wastewater with powdered activated carbon, ozone, and UV. Environ Sci Technol 47: 7899-7908. doi: 10.1021/es400708w
![]() |
[97] |
Mailler R, Gasperi J, Coquet Y, et al. (2016) Removal of a wide range of emerging pollutants from wastewater treatment plant discharges by micro-grain activated carbon in fluidized bed as tertiary treatment at large pilot scale. Sci Total Environ 542: 983-996. doi: 10.1016/j.scitotenv.2015.10.153
![]() |
[98] | Lima DRS, Baêta BEL, Aquino SF, et al. (2014) Removal of Pharmaceuticals and Endocrine Disruptor Compounds from Natural Waters by Clarification Associated with Powdered Activated Carbon. Water Air Soil Pollut 225. |
[99] |
Rubirola A, Llorca M, Rodriguez-Mozaz S, et al. (2014) Characterization of metoprolol biodegradation and its transformation products generated in activated sludge batch experiments and in full scale WWTPs. Water Res 63: 21-32. doi: 10.1016/j.watres.2014.05.031
![]() |
[100] |
Kårelid V, Larsson G, Björlenius B (2017) Pilot-scale removal of pharmaceuticals in municipal wastewater: Comparison of granular and powdered activated carbon treatment at three wastewater treatment plants. J Environ Manage 193: 491-502. doi: 10.1016/j.jenvman.2017.02.042
![]() |
[101] |
Li L, Quinlivan PA, Knappe DRU (2002) Effects of activated carbon surface chemistry and pore structure on the adsorption of organic contaminants from aqueous solution. Carbon N Y 40: 2085-2100. doi: 10.1016/S0008-6223(02)00069-6
![]() |
[102] |
Benstoem F, Nahrstedt A, Boehler M, et al. (2017) Performance of granular activated carbon to remove micropollutants from municipal wastewater-A meta-analysis of pilot- and large-scale studies. Chemosphere 185: 105-118. doi: 10.1016/j.chemosphere.2017.06.118
![]() |
[103] | Margot J, Magnet A (2011) Elimination des micropolluants dans les eaux usées-Essais pilotes à la station d épuration de Lausanne. gwa 7: 487-493. |
[104] |
Margot J, Kienle C, Magnet A, et al. (2013) Treatment of micropollutants in municipal wastewater: Ozone or powdered activated carbon? Sci Total Environ 461-462: 480-498. doi: 10.1016/j.scitotenv.2013.05.034
![]() |
[105] |
Baresel C, Malmborg J, Ek M, et al. (2016) Removal of pharmaceutical residues using ozonation as intermediate process step at Linköping WWTP, Sweden. Water Sci Technol 73: 2017-2024. doi: 10.2166/wst.2016.045
![]() |
[106] |
Mamo J, Insa S, Monclús H, et al. (2016) Fate of NDMA precursors through an MBR-NF pilot plant for urban wastewater reclamation and the effect of changing aeration conditions. Water Res 102: 383-393. doi: 10.1016/j.watres.2016.06.057
![]() |
[107] |
Hofman-Caris CHM, Siegers WG, van de Merlen K, et al. (2017) Removal of pharmaceuticals from WWTP effluent: Removal of EfOM followed by advanced oxidation. Chem Eng J 327: 514-521. doi: 10.1016/j.cej.2017.06.154
![]() |
[108] | Cédat B, de Brauer C, Métivier H, et al. (2016) Are UV photolysis and UV/H 2 O 2 process efficient to treat estrogens in waters? Chemical and biological assessment at pilot scale. Water Res 100: 357-366. |
[109] |
El-taliawy H, Ekblad M, Nilsson F, et al. (2017) Ozonation efficiency in removing organic micro pollutants from wastewater with respect to hydraulic loading rates and different wastewaters. Chem Eng J 325: 310-321. doi: 10.1016/j.cej.2017.05.019
![]() |
[110] |
Ibáñez M, Borova V, Boix C, et al. (2017) UHPLC-QTOF MS screening of pharmaceuticals and their metabolites in treated wastewater samples from Athens. J Hazard Mater 323: 26-35. doi: 10.1016/j.jhazmat.2016.03.078
![]() |
1. | Piotr F. Borowski, Nexus between water, energy, food and climate change as challenges facing the modern global, European and Polish economy, 2020, 6, 2471-2132, 397, 10.3934/geosci.2020022 | |
2. | Christopher Schulz, Udisha Saklani, The future of hydropower development in Nepal: Views from the private sector, 2021, 179, 09601481, 1578, 10.1016/j.renene.2021.07.138 | |
3. | Fanqi Zou, Tinghui Li, Feite Zhou, Does the Level of Financial Cognition Affect the Income of Rural Households? Based on the Moderating Effect of the Digital Financial Inclusion Index, 2021, 11, 2073-4395, 1813, 10.3390/agronomy11091813 | |
4. | Divya Narain, Hoong Chen Teo, Alex Mark Lechner, James E.M. Watson, Martine Maron, Biodiversity risks and safeguards of China’s hydropower financing in Belt and Road Initiative (BRI) countries, 2022, 5, 25903322, 1019, 10.1016/j.oneear.2022.08.012 | |
5. | Edoardo Borgomeo, Bill Kingdom, Judith Plummer-Braeckman, Winston Yu, Water infrastructure in Asia: financing and policy options, 2022, 0790-0627, 1, 10.1080/07900627.2022.2062707 | |
6. | Judith Plummer Braeckman, Sanna Markkanen, Perceptions of Risk in Relation to Large Hydropower Projects: A Finance Perspective, 2021, 1556-5068, 10.2139/ssrn.4011266 | |
7. | Emmanuel Yamoah Tenkorang, Francis Enu-Kwesi, Franklin Bendu, Pon Souvannaseng, Evolving Lending Regimes and the Political Economy of Dam Financing in Ghana, 2022, 1556-5068, 10.2139/ssrn.4013584 | |
8. | Yan Yang, Bo Yang, Zijun Xin, Green finance development, environmental attention and investment in hydroelectric power: From the perspective of environmental protection law, 2024, 69, 15446123, 106167, 10.1016/j.frl.2024.106167 | |
9. | Pon Souvannaseng, Fast Finance and the Political Economy of Catastrophic Dam Collapse in Lao PDR: The Case of Xe Pian-Xe Namnoy, 2024, 97, 0030-851X, 261, 10.5509/2024972-art7 |
Name | Description | Type of sensor | Unit |
TP2 | Compressor pressure | Analog | Bar |
TP3 | Pneumatic panel pressure | Analog | Bar |
H1 | Pressure of the valve that is activated when the pressure exceeds 10.2 bar | Analog | Bar |
DV_Pressure | Pressure drop due to water discharge by air dryers | Analog | Bar |
Reservoirs | Air tank pressure | Analog | Bar |
Oil temperature | Temperature of oil in compressor | Analog | Celsius |
Flowmeter | Airflow | Analog | m3/h |
Motor current | Current flowing in the motor | Analog | Ampere |
Comp | Electric signal of the compressor based on the air intake | Digital | - |
DV Electric | Electric signal of compressor outlet | Digital | - |
Towers | Specifies the two towers based on the action of air drying | Digital | - |
MPG | Trigger to start the compressor when the pressure is less than 8.2 bar | Digital | - |
LPS | Trigger when the pressure is less than 7 bar | Digital | - |
The pressure switch | Trigger when pressure is detected in the pilot control valve | Digital | - |
Oil level | Trigger when the oil level is less than the threshold | Digital | - |
Caudal Impulses | Trigger for the air flowmeter | Digital | - |
GPS Longitude | Longitude position | Analog | ° |
GPS Latitude | Latitude position | Analog | ° |
GPS Speed | Speed | Analog | Km/h |
TP2 | TP3 | H1 | DVP | Reservoirs | OT | Flowmeter | MoC | |
mean | 0.947 | 8.989 | 8.038 | -0.019 | 1.63 | 65.843 | 20.128 | 2.040 |
std | 2.836 | 0.667 | 2.846 | 0.185 | 0.064 | 5.931 | 3.578 | 2.198 |
min | -0.03 | 0.937 | -0.033 | -0.036 | 1.349 | 18.575 | 18.8347 | -0.009 |
25% | -0.009 | 8.492 | 8.332 | -0.0279 | 1.608 | 61.825 | 18.9748 | 0.0024 |
50% | -0.007 | 8.996 | 8.876 | -0.025 | 1.635 | 66.475 | 19.03 | 0.007 |
75% | -0.006 | 9.506 | 9.438 | -0.025 | 1.667 | 70.575 | 19.040 | 3.837 |
max | 10.806 | 10.38 | 10.368 | 8.11 | 1.791 | 80.174 | 37.008 | 9.3375 |
COMP | DVE | Towers | MPG | LPS | PrS | OL | CaI | |
mean | 0.892 | 0.107 | 0.946 | 0.892 | 0.004 | 0.0 | 0.0 | 0.002 |
std | 0.309 | 0.309 | 0.225 | 0.309 | 0.068 | 0.0 | 0.0 | 0.047 |
min | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
25% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
50% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
75% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
max | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 1.0 |
GPSLong | GPSLat | GPSSpeed | GPSQuality | |||||
mean | -7.880 | 37.578 | 8.592 | 0.912 | ||||
std | 2.443 | 11.649 | 14.096 | 0.282 | ||||
min | -8.69 | 0.0 | 0.0 | 0.0 | ||||
25% | -8.66106 | 41.1696 | 0.0 | 1.0 | ||||
50% | -8.658 | 41.1858 | 0.0 | 1.0 | ||||
75% | -8.583 | 41.212 | 16.0 | 1.0 | ||||
max | 0.0 | 41.240 | 286.0 | 1.0 | ||||
month | day | hour | minute | second | ||||
mean | 1.0 | 16.046 | 13.139 | 29.507 | 29.499 | |||
std | 0.0 | 8.919 | 6.444 | 17.318 | 17.318 | |||
min | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | |||
25% | 1.0 | 8.0 | 9.0 | 15.0 | 14.0 | |||
50% | 1.0 | 16.0 | 14.0 | 30.0 | 29.0 | |||
75% | 1.0 | 24.0 | 19.0 | 45.0 | 44.0 | |||
max | 1.0 | 31.0 | 23.0 | 59.0 | 59.0 |
Label Code | Corresponding fault |
0 | Air leak in air dryer |
1 | Air leak in client chamber |
2 | Oil leak in compressor |
3 | No fault |
Label Code | Type of failure | Location of failure |
0 | Air leak | Air dryer |
1 | Oil leak | Client chamber |
2 | No failure | Compressor |
3 | Not reported | No location |
Name of target vector | Description | Classes |
Type | Code for the type of failure | 0- Air failure |
1- Oil failure | ||
2- Normal | ||
Location | Code for the location of the failure | 0- Air dryer |
1- Client | ||
2- Compressor | ||
3- No location | ||
GPS quality | Quality of the GPS sensor | 0- Good |
1- Bad |
Set | Time consumption per epoch (ms) | Metric | Type identification | Location identification | GPS quality identification |
Training | 43 | Loss | 0.0031 | 0.0020 | 0.0029 |
Accuracy | 99.94 | 99.998 | 99.99 | ||
Precision | 99.99 | 100 | 100 | ||
Recall | 100 | 100 | 100 | ||
AUC | 1 | 1 | 1 | ||
Testing | 1 | Loss | 0.0035 | 0.0026 | 0.0033 |
Accuracy | 98.89 | 99.12 | 99.24 | ||
Precision | 99.56 | 99.67 | 99.84 | ||
Recall | 99.92 | 99.93 | 99.93 | ||
AUC | 1 | 1 | 1 |
Classes | Precision | Recall | F1-Score |
0 | 0.98 | 1.00 | 0.99 |
1 | 1.00 | 1.00 | 1.00 |
2 | 1.00 | 1.00 | 1.00 |
Classes | Precision | Recall | F1-Score |
0 | 1 | 1 | 1 |
1 | 1 | 1 | 1 |
2 | 1 | 1 | 1 |
3 | 1 | 1 | 1 |
Classes | Precision | Recall | F1-Score |
0 | 1 | 1 | 1 |
1 | 1 | 1 | 1 |
Author | Algorithm | Result |
Rajashekarappa et al. [13] | RUS Boosted Classifier | 98.73% accuracy |
Najjar et al. [15] | Random Forest Classifier | 97% accuracy |
Davari et al. [16] | Autoencoder | 44% improvement in precision compared to baseline |
Proposed work | Multi-task Artificial Neural Network | 99% average accuracy |
Name | Description | Type of sensor | Unit |
TP2 | Compressor pressure | Analog | Bar |
TP3 | Pneumatic panel pressure | Analog | Bar |
H1 | Pressure of the valve that is activated when the pressure exceeds 10.2 bar | Analog | Bar |
DV_Pressure | Pressure drop due to water discharge by air dryers | Analog | Bar |
Reservoirs | Air tank pressure | Analog | Bar |
Oil temperature | Temperature of oil in compressor | Analog | Celsius |
Flowmeter | Airflow | Analog | m3/h |
Motor current | Current flowing in the motor | Analog | Ampere |
Comp | Electric signal of the compressor based on the air intake | Digital | - |
DV Electric | Electric signal of compressor outlet | Digital | - |
Towers | Specifies the two towers based on the action of air drying | Digital | - |
MPG | Trigger to start the compressor when the pressure is less than 8.2 bar | Digital | - |
LPS | Trigger when the pressure is less than 7 bar | Digital | - |
The pressure switch | Trigger when pressure is detected in the pilot control valve | Digital | - |
Oil level | Trigger when the oil level is less than the threshold | Digital | - |
Caudal Impulses | Trigger for the air flowmeter | Digital | - |
GPS Longitude | Longitude position | Analog | ° |
GPS Latitude | Latitude position | Analog | ° |
GPS Speed | Speed | Analog | Km/h |
TP2 | TP3 | H1 | DVP | Reservoirs | OT | Flowmeter | MoC | |
mean | 0.947 | 8.989 | 8.038 | -0.019 | 1.63 | 65.843 | 20.128 | 2.040 |
std | 2.836 | 0.667 | 2.846 | 0.185 | 0.064 | 5.931 | 3.578 | 2.198 |
min | -0.03 | 0.937 | -0.033 | -0.036 | 1.349 | 18.575 | 18.8347 | -0.009 |
25% | -0.009 | 8.492 | 8.332 | -0.0279 | 1.608 | 61.825 | 18.9748 | 0.0024 |
50% | -0.007 | 8.996 | 8.876 | -0.025 | 1.635 | 66.475 | 19.03 | 0.007 |
75% | -0.006 | 9.506 | 9.438 | -0.025 | 1.667 | 70.575 | 19.040 | 3.837 |
max | 10.806 | 10.38 | 10.368 | 8.11 | 1.791 | 80.174 | 37.008 | 9.3375 |
COMP | DVE | Towers | MPG | LPS | PrS | OL | CaI | |
mean | 0.892 | 0.107 | 0.946 | 0.892 | 0.004 | 0.0 | 0.0 | 0.002 |
std | 0.309 | 0.309 | 0.225 | 0.309 | 0.068 | 0.0 | 0.0 | 0.047 |
min | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
25% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
50% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
75% | 1.0 | 0.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 |
max | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 1.0 |
GPSLong | GPSLat | GPSSpeed | GPSQuality | |||||
mean | -7.880 | 37.578 | 8.592 | 0.912 | ||||
std | 2.443 | 11.649 | 14.096 | 0.282 | ||||
min | -8.69 | 0.0 | 0.0 | 0.0 | ||||
25% | -8.66106 | 41.1696 | 0.0 | 1.0 | ||||
50% | -8.658 | 41.1858 | 0.0 | 1.0 | ||||
75% | -8.583 | 41.212 | 16.0 | 1.0 | ||||
max | 0.0 | 41.240 | 286.0 | 1.0 | ||||
month | day | hour | minute | second | ||||
mean | 1.0 | 16.046 | 13.139 | 29.507 | 29.499 | |||
std | 0.0 | 8.919 | 6.444 | 17.318 | 17.318 | |||
min | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | |||
25% | 1.0 | 8.0 | 9.0 | 15.0 | 14.0 | |||
50% | 1.0 | 16.0 | 14.0 | 30.0 | 29.0 | |||
75% | 1.0 | 24.0 | 19.0 | 45.0 | 44.0 | |||
max | 1.0 | 31.0 | 23.0 | 59.0 | 59.0 |
Label Code | Corresponding fault |
0 | Air leak in air dryer |
1 | Air leak in client chamber |
2 | Oil leak in compressor |
3 | No fault |
Label Code | Type of failure | Location of failure |
0 | Air leak | Air dryer |
1 | Oil leak | Client chamber |
2 | No failure | Compressor |
3 | Not reported | No location |
Name of target vector | Description | Classes |
Type | Code for the type of failure | 0- Air failure |
1- Oil failure | ||
2- Normal | ||
Location | Code for the location of the failure | 0- Air dryer |
1- Client | ||
2- Compressor | ||
3- No location | ||
GPS quality | Quality of the GPS sensor | 0- Good |
1- Bad |
Set | Time consumption per epoch (ms) | Metric | Type identification | Location identification | GPS quality identification |
Training | 43 | Loss | 0.0031 | 0.0020 | 0.0029 |
Accuracy | 99.94 | 99.998 | 99.99 | ||
Precision | 99.99 | 100 | 100 | ||
Recall | 100 | 100 | 100 | ||
AUC | 1 | 1 | 1 | ||
Testing | 1 | Loss | 0.0035 | 0.0026 | 0.0033 |
Accuracy | 98.89 | 99.12 | 99.24 | ||
Precision | 99.56 | 99.67 | 99.84 | ||
Recall | 99.92 | 99.93 | 99.93 | ||
AUC | 1 | 1 | 1 |
Classes | Precision | Recall | F1-Score |
0 | 0.98 | 1.00 | 0.99 |
1 | 1.00 | 1.00 | 1.00 |
2 | 1.00 | 1.00 | 1.00 |
Classes | Precision | Recall | F1-Score |
0 | 1 | 1 | 1 |
1 | 1 | 1 | 1 |
2 | 1 | 1 | 1 |
3 | 1 | 1 | 1 |
Classes | Precision | Recall | F1-Score |
0 | 1 | 1 | 1 |
1 | 1 | 1 | 1 |
Author | Algorithm | Result |
Rajashekarappa et al. [13] | RUS Boosted Classifier | 98.73% accuracy |
Najjar et al. [15] | Random Forest Classifier | 97% accuracy |
Davari et al. [16] | Autoencoder | 44% improvement in precision compared to baseline |
Proposed work | Multi-task Artificial Neural Network | 99% average accuracy |