Fraunhofer-Institut für Optronik, Systemtechnik und Bildauswertung IOSB

Filters

Reset filters

Settings
Now showing 1 - 10 of 14
  • Publication
    Probabilistic multi-step ahead streamflow forecast based on deep learning
    ( 2024)
    Karimanzira, Divas
    ;
    Richter, Lucas
    ;
    Hilbring, Desiree
    ;
    Lödige, Michaela
    ;
    Vogl, Jonathan
    The use of deep learning methods for fluvial flood forecasting is rapidly gaining traction, offering a promising solution to the challenges associated with accurate yet time-consuming numerical models. This paper presents two physics-inspired deep learning approaches specifically designed for fluvial flood forecasting, each embracing different learning principles: centralized and federated learning. The centralized model utilizes an Encoder-Decoder technique to handle input data of varying types and scales, while the federated model is based on a node-link graph with a seq2seq internal model. Both models are enhanced with a probabilistic forecasting head to account for the inherent uncertainty in streamflow forecasts. The objective of these approaches is to address the limitations of traditional numerical models while leveraging the potential of deep learning to improve the speed, accuracy, and scalability of flood forecasting. To validate their effectiveness, the models were tested across different use cases. The findings from the federated learning approach emphasize the importance of catchment clustering before model utilization and demonstrate the models’ ability to generalize effectively in catchments with similar properties. On the other hand, the results of the centralized method highlight the model’s reliance on the test set falling within the data range of the training set (Average NSE and KGE for the sixth hour ahead of 0.88 and 0.78, respectively). To address this limitation, the paper suggests the development of a method for the future, such as leveraging a numerical model or using Generative Adversarial Networks, to generate highly extreme events, particularly in the context of a changing climate. The models are implemented in a flexible operational framework based on open standards, ensuring their adaptability and usability in various settings.
  • Publication
    Application of machine learning and deep neural networks for spatial prediction of groundwater nitrate concentration to improve land use management practices
    ( 2023)
    Karimanzira, Divas
    ;
    Weis, Jonas
    ;
    Wunsch, Andreas
    ;
    Ritzau, Linda
    ;
    Liesch, Tanja
    ;
    Ohmer, Marc
    The prediction of groundwater nitrate concentration's response to geo-environmental and human-influenced factors is essential to better restore groundwater quality and improve land use management practices. In this paper, we regionalize groundwater nitrate concentration using different machine learning methods (Random forest (RF), unimodal 2D and 3D convolutional neural networks (CNN), and multi-stream early and late fusion 2D-CNNs) so that the nitrate situation in unobserved areas can be predicted. CNNs take into account not only the nitrate values of the grid cells of the observation wells but also the values around them. This has the added benefit of allowing them to learn directly about the influence of the surroundings. The predictive performance of the models was tested on a dataset from a pilot region in Germany, and the results show that, in general, all the machine learning models, after a Bayesian optimization hyperparameter search and training, achieve good spatial predictive performance compared to previous studies based on Kriging and numerical models. Based on the mean absolute error (MAE), the random forest model and the 2DCNN late fusion model performed best with an MAE (STD) of 9.55 (0.367) mg/l, R2 = 0.43 and 10.32 (0.27) mg/l, R2 = 0.27, respectively. The 3DCNN with an MAE (STD) of 11.66 (0.21) mg/l and largest resources consumption is the worst performing model. Feature importance learning from the models was used in conjunction with partial dependency analysis of the most important features to gain greater insight into the major factors explaining the nitrate spatial variability. Large uncertainties in nitrate prediction have been shown in previous studies. Therefore, the models were extended to quantify uncertainty using prediction intervals (PIs) derived from bootstrapping. Knowledge of uncertainty helps the water manager reduce risk and plan more reliably.
  • Publication
    Nutzung offener Standards für die Integration von Pegeldaten und KI-basierte Prognose von Pegel- und Abflussdaten zur Verbesserung der Frühwarnung bei Sturzfluten
    ( 2023)
    Vogl, Jonathan
    ;
    Hilbring, Desiree
    ;
    Karimanzira, Divas
    Im Kontext von Frühwarnsystemen, deren Bedeutung in Zeiten mit steigender Anzahl an Extremwetterereignissen wächst, spielt die Verfügbarkeit von Messdaten in Kombination mit einem schnellen Datenaustausch zwischen KI-Algorithmen zur Berechnung und Plattformen zur Visualisierung von Prognoseergebnissen eine entscheidende Rolle. Im Rahmen des Eigenforschungsprojekts PrognoSF des Fraunhofer IOSB werden Niederschlagsprognosedaten des DWD mit lokalen Sensor- daten aus Smart City Plattformen kombiniert und Pegelprognosen auf Basis von KI durchgeführt, die später auch Frühwarnungen, gerade bei Gefährdung von kritischer Infrastruktur, gewährleisten sollen. Für die flexible Integration von Daten und Modulen kommen offene Standards zum Einsatz. Für die Pegelprognose wird ein neuronales maschinelles Übersetzungsmodell trainiert, das aus tiefen rekurrenten Netzen besteht.
  • Publication
    Advanced Spatio-Temporal Event Detection System for Groundwater Quality Based on Deep Learning
    ( 2023)
    Karimanzira, Divas
    ;
    Ritzau, Linda
    ;
    Martin, Tobias
    ;
    Fischer, Thilo
    It is very important in sensor networks for monitoring, e.g. groundwater quality, to detect sensor failures and anomalous spatio-temporal events such as spills and identify affected areas. Most of the method for anomaly detection do not truly utilize spatial and temporal information. In this paper a novel method based on deep learning (DL) is proposed which truly utilize multivariate spatio-temporal information in anomalous events detection. Anomalous events are quite rare, which makes it very challenging to obtain labeled anomaly datasets. It is therefore purposeful to use an unsupervised approach for sensor anomaly and event detection with labels only being used to set thresholds on prediction errors. Two method for an unsupervised anomaly detection in multivariate spatio-temporal data using deep learning are proposed in this paper. The first framework is composed of a Long Short Term Memory (LSTM) Encoder followed by an LSTM decoder and a LSTM predictor for temporal anomaly detection. In a further step, a Deep Neural Network (DNN) based classifier is used to classify the encoded and trained latent representation to their spatial corresponding to form a temporal and spatial anomaly detector. The second framework is based on a CNN encoder and a LSTM decoder to capture both spatial and temporal features. The encoder component can use either 3D convolutions or Multichannel CNN to capture complex spatial dependencies in each spatial neighborhood. The 3D tensor input for the encoder is formed by stacking the data from the nearest spatial neighbors of each data point. Both methods produce similar results for event detection, detecting different types of anomalies (point, context, etc.). After the training phase to learn the normal system behavior, both methods are capable of detecting anomalies that have never been seen before with a very good accuracy (values ranging between 88% and 96%). To validate the accuracy and efficiency of the DL-based methods, they were compared to a modified ST-DBSCAN algorithm. The results show the superiority of the DL-based methods.
  • Publication
    Running a Reverse Osmosis Plant at Maximum Renewable Energy Use
    ( 2023)
    Karimanzira, Divas
    ;
    Rauschenbach, Thomas
    Decentralized water and energy systems are increasingly seen as the most effective solutions for remote areas. This paper presents the development of a solar-powered reverse osmosis (RO) plant to provide water to a remote village. The objective is to optimize the use of renewable energy and maximize water production by implementing a control mechanism that can quickly adapt to changing operating conditions. To achieve this, the RO plant operates in batch mode and is controlled by a nonlinear model-based predictive controller. It also serves as an active load to effectively utilize solar energy. Fortunately, in arid and water-scarce regions, the water demand aligns well with solar power availability. Although a diesel generator is used as a backup, energy is stored as pressure in a pressure chamber and in batteries for direct electrical energy consumption by the plant. The energy-intensive tasks, such as driving the high-pressure pump and membrane cleaning, are scheduled during periods of abundant solar energy. Additionally, a module for predicting membrane fouling is developed to enable the controller to appropriately schedule membrane cleaning and pump operation, while considering the various limitations imposed by the RO plant components. The system has successfully met the water demand while maximizing solar power utilization.
  • Publication
    Event Detection in Groundwater Sensor Networks using Artificial Intelligence
    ( 2023)
    Karimanzira, Divas
    ;
    Ritzau, Linda
    In this paper, a method for detecting spatial and temporal anomalous events in groundwater sensor networks (high-dimensional time series data), such as system faults and attacks will be developed. Unlike recently developed deep learning frameworks for anomaly detection which do not consider the dependences between the variables and apply the existing relationships to predict the expected behavior of the sensors, the method in this paper extracts the relationships between the sensors spatially and temporally and learn to detect and simultaneously explain deviations from these relationships. This challenge is solved by using graph attention neural networks and structured learning. Attention neural networks can give useful interpretability in context of the anomalies detected and allows to identify their causes. To improve robustness the method considers aleatoric and parametric uncertainties by using ensemble specific value prediction and prediction intervals without assuming any data distribution. Furthermore, the model was connected to a fully connected classifier to classify typical groundwater network anomalies. The method was applied to a study area and it could be shown that the method could capture in 92% of the cases the complex correlations between the high dimensional variables, and enabled analysts to identify the causes of the anomalies.
  • Publication
    Simultaneous Pipe Leak Detection and Localization Using Attention-Based Deep Learning Autoencoder
    ( 2023)
    Karimanzira, Divas
    Water distribution networks are often susceptible to pipeline leaks caused by mechanical damages, natural hazards, corrosion, and other factors. This paper focuses on the detection of leaks in water distribution networks (WDN) using a data-driven approach based on machine learning. A hybrid autoencoder neural network (AE) is developed, which utilizes unsupervised learning to address the issue of unbalanced data (as anomalies are rare events). The AE consists of a 3DCNN encoder, a ConvLSTM decoder, and a ConvLSTM future predictor, making the anomaly detection robust. Additionally, spatial and temporal attention mechanisms are employed to enhance leak localization. The AE first learns the expected behavior and subsequently detects leaks by identifying deviations from this expected behavior. To evaluate the performance of the proposed method, the Water Network Tool for Resilience (WNTR) simulator is utilized to generate water pressure and flow rate data in a water supply network. Various conditions, such as fluctuating water demands, data noise, and the presence of leaks, are considered using the pressure-driven demand (PDD) method. Datasets with and without pipe leaks are obtained, where the AE is trained using the dataset without leaks and tested using the dataset with simulated pipe leaks. The results, based on a benchmark WDN and a confusion matrix analysis, demonstrate that the proposed method successfully identifies leaks in 96% of cases and a false positive rate of 4% compared to two baselines: a multichannel CNN encoder with LSTM decoder (MC-CNN-LSTM) and a random forest and model based on supervised learning with a false positive rate of 8% and 15%, respectively. Furthermore, a real case study demonstrates the applicability of the developed model for leak detection in the operational conditions of water supply networks using inline sensor data.
  • Publication
    Optimal Control of Fish Growth Under Uncertainty Using Chance Constraint Model Predictive Controller
    ( 2022)
    Karimanzira, Divas
    ;
    Rauschenbach, Thomas
    To achieve sustainability goals (social, economic and environmental), recirculation aquaculture systems (RAS) should be operated at optimal conditions. Recirculation aquaculture systems have various additive and unpredictable disturbances such as irregular temperature in big tanks, uneven distribution of feed and uncertain rate constants for nutrients utilization by fish. With such disturbance the performance of RAS cannot be met with probability one. Thus, the performance level should be attained under a desired probability (reliability) level. Therefore, in this paper we apply the chance-constraint model predictive control (CC-MPC) approach with state estimation using an Unscented Kalman Filter (UKF) to meet this requirement. The cost function is based on the feed conversion ratio metric, profit maximization with desired growth reference trajectory tracking. A bio-energetic and econometric model of fish growth in a RAS is utilized to illustrate the application of the formulation through simulations. Healthy fish growth is enabled by applying a health monitoring estimator based on artificial Intelligence (AI) in the feedback loop. The CC-MPC is compared to a deterministic MPC, with a focus on constraint breaching, computation time, and operational behavior. The simulations show similar performance for the fish growth for both types of MPC, while the computation time increases slightly for the CC-MPC, together with operational behaviors getting limited. In the case study, a final average fish weight of 433g is reached at a reliability level of 95% compared to 429g of the deterministic.
  • Publication
    Moving Horizon vs. Unscented Kalman Filter for State Estimation in Streamflow Prediction
    ( 2022)
    Karimanzira, Divas
    ;
    Rauschenbach, Thomas
    In this paper, a Moving Horizon Estimator (MHE) and an Unscented Kalman Filter (UKF) are applied and compared for state estimation in flood forecasting. The investigations are based on a conceptual rainfall-runoff model proposed by Lorent/Gevers for streamflow forecasting. Data for the investigations was collected from the region Trusetal in Germany. Streamflow prediction, especially for watersheds with fast response to intense rain, require the knowledge of the current state of the system (e.g., soil moisture content). Firstly, a Moving Horizon Estimator (MHE) was applied for the state estimation, due to our good experience with it in other applications, its ability to deal with non-Gaussian disturbances and the fact that the hydrologic model is nonlinear, and its states satisfy equality and inequality constraints. Due to computational intensity of the MHE, an UKF was also implemented for comparison. Even though theory and most literature conclude the superiority of MHE to UKF, in this application example the results show that the UKF and the MHE produce almost similar results with UKF slightly better, which might be due to several reasons such as problems with the initialization of the hessian matrix, choice of prediction horizon and existence of local optima in MHE. Therefore, comprehensive investigations were performed in this respect.