Search results for: shape prediction

Authors: Faruk Ozdemir, Roy Kalawsky, Peter Hubbard

Abstract:

Accurate estimation of RUL provides a basis for effective predictive maintenance, reducing unexpected downtime for industrial equipment. However, while models such as the Random Forest have effective predictive capabilities, they are the so-called ‘black box’ models, where interpretability is at a threshold to make critical diagnostic decisions involved in industries related to aviation. The purpose of this work is to present a prognostic framework that embeds Explainable Artificial Intelligence (XAI) techniques in order to provide essential transparency in Machine Learning methods' decision-making mechanisms based on sensor data, with the objective of procuring actionable insights for the aviation industry. Sensor readings have been gathered from critical equipment such as turbofan jet engine and landing gear, and the prediction of the RUL is done by a Random Forest model. It involves steps such as data gathering, feature engineering, model training, and evaluation. These critical components’ datasets are independently trained and evaluated by the models. While suitable predictions are served, their performance metrics are reasonably good; such complex models, however obscure reasoning for the predictions made by them and may even undermine the confidence of the decision-maker or the maintenance teams. This is followed by global explanations using SHAP and local explanations using LIME in the second phase to bridge the gap in reliability within industrial contexts. These tools analyze model decisions, highlighting feature importance and explaining how each input variable affects the output. This dual approach offers a general comprehension of the overall model behavior and detailed insight into specific predictions. The proposed framework, in its third component, incorporates the techniques of causal analysis in the form of Granger causality tests in order to move beyond correlation toward causation. This will not only allow the model to predict failures but also present reasons, from the key sensor features linked to possible failure mechanisms to relevant personnel. The causality between sensor behaviors and equipment failures creates much value for maintenance teams due to better root cause identification and effective preventive measures. This step contributes to the system being more explainable. Surrogate Several simple models, including Decision Trees and Linear Models, can be used in yet another stage to approximately represent the complex Random Forest model. These simpler models act as backups, replicating important jobs of the original model's behavior. If the feature explanations obtained from the surrogate model are cross-validated with the primary model, the insights derived would be more reliable and provide an intuitive sense of how the input variables affect the predictions. We then create an iterative explainable feedback loop, where the knowledge learned from the explainability methods feeds back into the training of the models. This feeds into a cycle of continuous improvement both in model accuracy and interpretability over time. By systematically integrating new findings, the model is expected to adapt to changed conditions and further develop its prognosis capability. These components are then presented to the decision-makers through the development of a fully transparent condition monitoring and alert system. The system provides a holistic tool for maintenance operations by leveraging RUL predictions, feature importance scores, persistent sensor threshold values, and autonomous alert mechanisms. Since the system will provide explanations for the predictions given, along with active alerts, the maintenance personnel can make informed decisions on their end regarding correct interventions to extend the life of the critical machinery.

Keywords: predictive maintenance, explainable artificial intelligence, prognostic, RUL, machine learning, turbofan engines, C-MAPSS dataset

Procedia PDF Downloads 7

Authors: Atish Bagchi, Siva Chandrasekaran

Abstract:

Background: According to World Bank national accounts data, the estimated global manufacturing value-added output in 2020 was 13.74 trillion USD. These manufacturing processes are monitored, modelled, and controlled by advanced, real-time, computer-based systems, e.g., Industrial IoT, PLC, SCADA, etc. These systems measure and manipulate a set of physical variables, e.g., temperature, pressure, etc. Despite the use of IoT, SCADA etc., in manufacturing, studies suggest that unplanned downtime leads to economic losses of approximately 864 billion USD each year. Therefore, real-time, accurate detection, classification and prediction of machine behaviour are needed to minimise financial losses. Although vast literature exists on time-series data processing using machine learning, the challenges faced by the industries that lead to unplanned downtimes are: The current algorithms do not efficiently handle the high-volume streaming data from industrial IoTsensors and were tested on static and simulated datasets. While the existing algorithms can detect significant 'point' outliers, most do not handle contextual outliers (e.g., values within normal range but happening at an unexpected time of day) or subtle changes in machine behaviour. Machines are revamped periodically as part of planned maintenance programmes, which change the assumptions on which original AI models were created and trained. Aim: This research study aims to deliver a Graph Neural Network(GNN)based hybrid forecasting model that interfaces with the real-time machine control systemand can detect, predict machine behaviour and behavioural changes (anomalies) in real-time. This research will help manufacturing industries and utilities, e.g., water, electricity etc., reduce unplanned downtimes and consequential financial losses. Method: The data stored within a process control system, e.g., Industrial-IoT, Data Historian, is generally sampled during data acquisition from the sensor (source) and whenpersistingin the Data Historian to optimise storage and query performance. The sampling may inadvertently discard values that might contain subtle aspects of behavioural changes in machines. This research proposed a hybrid forecasting and classification model which combines the expressive and extrapolation capability of GNN enhanced with the estimates of entropy and spectral changes in the sampled data and additional temporal contexts to reconstruct the likely temporal trajectory of machine behavioural changes. The proposed real-time model belongs to the Deep Learning category of machine learning and interfaces with the sensors directly or through 'Process Data Historian', SCADA etc., to perform forecasting and classification tasks. Results: The model was interfaced with a Data Historianholding time-series data from 4flow sensors within a water treatment plantfor45 days. The recorded sampling interval for a sensor varied from 10 sec to 30 min. Approximately 65% of the available data was used for training the model, 20% for validation, and the rest for testing. The model identified the anomalies within the water treatment plant and predicted the plant's performance. These results were compared with the data reported by the plant SCADA-Historian system and the official data reported by the plant authorities. The model's accuracy was much higher (20%) than that reported by the SCADA-Historian system and matched the validated results declared by the plant auditors. Conclusions: The research demonstrates that a hybrid GNN based approach enhanced with entropy calculation and spectral information can effectively detect and predict a machine's behavioural changes. The model can interface with a plant's 'process control system' in real-time to perform forecasting and classification tasks to aid the asset management engineers to operate their machines more efficiently and reduce unplanned downtimes. A series of trialsare planned for this model in the future in other manufacturing industries.

Keywords: GNN, Entropy, anomaly detection, industrial time-series, AI, IoT, Industry 4.0, Machine Learning

Procedia PDF Downloads 150

Authors: Essam Gooda, Chadi Abdallah, Hamdi Seif, Safaa Baydoun, Rouya Hdeib, Hilal Obeid

Abstract:

Floods are common natural disaster occurring in semi-arid regions in Lebanon. This results in damage to human life and deterioration of environment. Despite their destructive nature and their immense impact on the socio-economy of the region, flash floods have not received adequate attention from policy and decision makers. This is mainly because of poor understanding of the processes involved and measures needed to manage the problem. The current understanding of flash floods remains at the level of general concepts; most policy makers have yet to recognize that flash floods are distinctly different from normal riverine floods in term of causes, propagation, intensity, impacts, predictability, and management. Flash floods are generally not investigated as a separate class of event but are rather reported as part of the overall seasonal flood situation. As a result, Lebanon generally lacks policies, strategies, and plans relating specifically to flash floods. Main objective of this research is to improve flash flood prediction by providing new knowledge and better understanding of the hydrological processes governing flash floods in the East Catchments of El Assi River. This includes developing rainstorm time distribution curves that are unique for this type of study region; analyzing, investigating, and developing a relationship between arid watershed characteristics (including urbanization) and nearby villages flow flood frequency in Ras Baalbeck and Fekha. This paper discusses different levels of integration approach¬es between GIS and hydrological models (HEC-HMS & HEC-RAS) and presents a case study, in which all the tasks of creating model input, editing data, running the model, and displaying output results. The study area corresponds to the East Basin (Ras Baalbeck & Fakeha), comprising nearly 350 km2 and situated in the Bekaa Valley of Lebanon. The case study presented in this paper has a database which is derived from Lebanese Army topographic maps for this region. Using ArcMap to digitizing the contour lines, streams & other features from the topographic maps. The digital elevation model grid (DEM) is derived for the study area. The next steps in this research are to incorporate rainfall time series data from Arseal, Fekha and Deir El Ahmar stations to build a hydrologic data model within a GIS environment and to combine ArcGIS/ArcMap, HEC-HMS & HEC-RAS models, in order to produce a spatial-temporal model for floodplain analysis at a regional scale. In this study, HEC-HMS and SCS methods were chosen to build the hydrologic model of the watershed. The model then calibrated using flood event that occurred between 7th & 9th of May 2014 which considered exceptionally extreme because of the length of time the flows lasted (15 hours) and the fact that it covered both the watershed of Aarsal and Ras Baalbeck. The strongest reported flood in recent times lasted for only 7 hours covering only one watershed. The calibrated hydrologic model is then used to build the hydraulic model & assessing of flood hazards maps for the region. HEC-RAS Model is used in this issue & field trips were done for the catchments in order to calibrated both Hydrologic and Hydraulic models. The presented models are a kind of flexible procedures for an ungaged watershed. For some storm events it delivers good results, while for others, no parameter vectors can be found. In order to have a general methodology based on these ideas, further calibration and compromising of results on the dependence of many flood events parameters and catchment properties is required.

Keywords: flood risk management, flash flood, semi arid region, El Assi River, hazard maps

Procedia PDF Downloads 478

Authors: Gertraud Koch, Teresa Stumpf, Alejandra Tijerina García

Abstract:

Discourse analysis research approaches belong to the central research strategies applied throughout the humanities; they focus on the countless forms and ways digital texts and images shape present-day notions of the world. Despite the constantly growing number of relevant digital, multimodal discourse resources, digital humanities (DH) methods are thus far not systematically developed and accessible for discourse analysis approaches. Specifically, the significance of multimodality and meaning plurality modelling are yet to be sufficiently addressed. In order to address this research gap, the D-WISE project aims to develop a prototypical working environment as digital support for the sociology of knowledge approach to discourse analysis and new IT-analysis approaches for the use of context-oriented embedding representations. Playing an essential role throughout our research endeavor is the constant optimization of hermeneutical methodology in the use of (semi)automated processes and their corresponding epistemological reflection. Among the discourse analyses, the sociology of knowledge approach to discourse analysis is characterised by the reconstructive and accompanying research into the formation of knowledge systems in social negotiation processes. The approach analyses how dominant understandings of a phenomenon develop, i.e., the way they are expressed and consolidated by various actors in specific arenas of discourse until a specific understanding of the phenomenon and its socially accepted structure are established. This article presents insights and initial findings from D-WISE, a joint research project running since 2021 between the Institute of Anthropological Studies in Culture and History and the Language Technology Group of the Department of Informatics at the University of Hamburg. As an interdisciplinary team, we develop central innovations with regard to the availability of relevant DH applications by building up a uniform working environment, which supports the procedure of the sociology of knowledge approach to discourse analysis within open corpora and heterogeneous, multimodal data sources for researchers in the humanities. We are hereby expanding the existing range of DH methods by developing contextualized embeddings for improved modelling of the plurality of meaning and the integrated processing of multimodal data. The alignment of this methodological and technical innovation is based on the epistemological working methods according to grounded theory as a hermeneutic methodology. In order to systematically relate, compare, and reflect the approaches of structural-IT and hermeneutic-interpretative analysis, the discourse analysis is carried out both manually and digitally. Using the example of current discourses on digitization in the healthcare sector and the associated issues regarding data protection, we have manually built an initial data corpus of which the relevant actors and discourse positions are analysed in conventional qualitative discourse analysis. At the same time, we are building an extensive digital corpus on the same topic based on the use and further development of entity-centered research tools such as topic crawlers and automated newsreaders. In addition to the text material, this consists of multimodal sources such as images, video sequences, and apps. In a blended reading process, the data material is filtered, annotated, and finally coded with the help of NLP tools such as dependency parsing, named entity recognition, co-reference resolution, entity linking, sentiment analysis, and other project-specific tools that are being adapted and developed. The coding process is carried out (semi-)automated by programs that propose coding paradigms based on the calculated entities and their relationships. Simultaneously, these can be specifically trained by manual coding in a closed reading process and specified according to the content issues. Overall, this approach enables purely qualitative, fully automated, and semi-automated analyses to be compared and reflected upon.

Keywords: entanglement of structural IT and hermeneutic-interpretative analysis, multimodality, plurality of meaning, sociology of knowledge approach to discourse analysis

Procedia PDF Downloads 226

Authors: Elhadi Nasr, Ibrahim Ramadan

Abstract:

Biostratigraphical analyses of well sections from the Maragh Low in the Eastern Sirt Basin has allowed high resolution correlations to be undertaken. Full integration of this data with available palaeoenvironmental, lithological, gravity, seismic, aeromagnetic, igneous, radiometric and wireline log information and a geochemical analysis of source rock quality and distribution has led to a more detailed understanding of the geological and the structural history of this area. Pre Sirt Unconformity two superimposed rifting cycles have been identified. The oldest is represented by the Amal Group of sediments and is of Late Carboniferous, Kasimovian / Gzelian to Middle Triassic, Anisian age. Unconformably overlying is a younger rift cycle which is represented the Sarir Group of sediments and is of Early Cretaceous, late Neocomian to Aptian in age. Overlying the Sirt Unconformity is the marine Late Cretaceous section. An assessment of pyrolysis results and a palynofacies analysis has allowed hydrocarbon source facies and quality to be determined. There are a number of hydrocarbon source rock horizons in the Maragh Low, these are sometimes vertically stacked and they are of fair to excellent quality. The oldest identified source rock is the Triassic Shale, this unit is unconformably overlain by sandstones belonging to the Sarir Group and conformably overlies a Triassic Siltstone unit. Palynological dating of the Triassic Shale unit indicates a Middle Triassic, Anisian age. The Triassic Shale is interpreted to have been deposited in a lacustrine palaeoenvironment. This particularly is evidenced by the dark, fine grained, organic rich nature of the sediment and is supported by palynofacies analysis and by the recovery of fish fossils. Geochemical analysis of the Triassic Shale indicates total organic carbon varying between 1.37 and 3.53. S2 pyrolysate yields vary between 2.15 mg/g and 6.61 mg/g and hydrogen indices vary between 156.91 and 278.91. The source quality of the Triassic Shale varies from being of fair to very good / rich. Linked to thermal maturity it is now a very good source for light oil and gas. It was once a very good to rich oil source. The Early Barremian Shale was also deposited in a lacustrine palaeoenvironment. Recovered palynomorphs indicate an Early Cretaceous, late Neocomian to early Barremian age. The Early Barremian Shale is conformably underlain and overlain by sandstone units belonging to the Sarir Group of sediments which are also of Early Cretaceous age. Geochemical analysis of the Early Barremian Shale indicates that it is a good oil source and was originally very good. Total organic carbon varies between 3.59% and 7%. S2 varies between 6.30 mg/g and 10.39 mg/g and the hydrogen indices vary between 148.4 and 175.5. A Late Barremian Shale unit of this age has also been identified in the central Maragh Low. Geochemical analyses indicate that total organic carbon varies between 1.05 and 2.38%, S2 pyrolysate between 1.6 and 5.34 mg/g and the hydrogen index between 152.4 and 224.4. It is a good oil source rock which is now mature. In addition to the non marine hydrocarbon source rocks pre Sirt Unconformity, three formations in the overlying Late Cretaceous section also provide hydrocarbon quality source rocks. Interbedded shales within the Rachmat Formation of Late Cretaceous, early Campanian age have total organic carbon ranging between, 0.7 and 1.47%, S2 pyrolysate varying between 1.37 and 4.00 mg/g and hydrogen indices varying between 195.7 and 272.1. The indication is that this unit would provide a fair gas source to a good oil source. Geochemical analyses of the overlying Tagrifet Limestone indicate that total organic carbon varies between 0.26% and 1.01%. S2 pyrolysate varies between 1.21 and 2.16 mg/g and hydrogen indices vary between 195.7 and 465.4. For the overlying Sirt Shale Formation of Late Cretaceous, late Campanian age, total organic carbon varies between 1.04% and 1.51%, S2 pyrolysate varies between 4.65 mg/g and 6.99 mg/g and the hydrogen indices vary between 151 and 462.9. The study has proven that both the Sirt Shale Formation and the Tagrifet Limestone are good to very good and rich sources for oil in the Maragh Low. High resolution biostratigraphical interpretations have been integrated and calibrated with thermal maturity determinations (Vitrinite Reflectance (%Ro), Spore Colour Index (SCI) and Tmax (ºC) and the determined present day geothermal gradient of 25ºC / Km for the Maragh Low. Interpretation of generated basin modelling profiles allows a detailed prediction of timing of maturation development of these source horizons and leads to a determination of amounts of missing section at major unconformities. From the results the top of the oil window (0.72% Ro) is picked as high as 10,700’ and the base of the oil window (1.35% Ro) assuming a linear trend and by projection is picked as low as 18,000’ in the Maragh Low. For the Triassic Shale the early phase of oil generation was in the Late Palaeocene / Early to Middle Eocene and the main phase of oil generation was in the Middle to Late Eocene. The Early Barremian Shale reached the main phase of oil generation in the Early Oligocene with late generation being reached in the Middle Miocene. For the Rakb Group section (Rachmat Formation, Tagrifet Limestone and Sirt Shale Formation) the early phase of oil generation started in the Late Eocene with the main phase of generation being between the Early Oligocene and the Early Miocene. From studying maturity profiles and from regional considerations it can be predicted that up to 500’ of sediment may have been deposited and eroded by the Sirt Unconformity in the central Maragh Low while up to 2000’ of sediment may have been deposited and then eroded to the south of the trough.

Keywords: Geochemical analysis of the source rocks from wells in Eastern Sirt Basin.

Procedia PDF Downloads 409

Authors: Manavjit Singh Dhindsa, Ursula Das, Kshirasagar Naik, Marzia Zaman, Richard Purcell, Srinivas Sampalli, Abdul Mutakabbir, Chung-Horng Lung, Thambirajah Ravichandran

Abstract:

These days, wildland fires, also known as forest fires, are more prevalent than ever. Wildfires have major repercussions that affect ecosystems, communities, and the environment in several ways. Wildfires lead to habitat destruction and biodiversity loss, affecting ecosystems and causing soil erosion. They also contribute to poor air quality by releasing smoke and pollutants that pose health risks, especially for individuals with respiratory conditions. Wildfires can damage infrastructure, disrupt communities, and cause economic losses. The economic impact of firefighting efforts, combined with their direct effects on forestry and agriculture, causes significant financial difficulties for the areas impacted. This research explores different forest fire spread models and presents a comprehensive review of various techniques and methodologies used in the field. A forest fire spread model is a computational or mathematical representation that is used to simulate and predict the behavior of a forest fire. By applying scientific concepts and data from empirical studies, these models attempt to capture the intricate dynamics of how a fire spreads, taking into consideration a variety of factors like weather patterns, topography, fuel types, and environmental conditions. These models assist authorities in understanding and forecasting the potential trajectory and intensity of a wildfire. Emphasizing the need for a comprehensive understanding of wildfire dynamics, this research explores the approaches, assumptions, and findings derived from various models. By using a comparison approach, a critical analysis is provided by identifying patterns, strengths, and weaknesses among these models. The purpose of the survey is to further wildfire research and management techniques. Decision-makers, researchers, and practitioners can benefit from the useful insights that are provided by synthesizing established information. Fire spread models provide insights into potential fire behavior, facilitating authorities to make informed decisions about evacuation activities, allocating resources for fire-fighting efforts, and planning for preventive actions. Wildfire spread models are also useful in post-wildfire mitigation strategies as they help in assessing the fire's severity, determining high-risk regions for post-fire dangers, and forecasting soil erosion trends. The analysis highlights the importance of customized modeling approaches for various circumstances and promotes our understanding of the way forest fires spread. Some of the known models in this field are Rothermel’s wildland fuel model, FARSITE, WRF-SFIRE, FIRETEC, FlamMap, FSPro, cellular automata model, and others. The key characteristics that these models consider include weather (includes factors such as wind speed and direction), topography (includes factors like landscape elevation), and fuel availability (includes factors like types of vegetation) among other factors. The models discussed are physics-based, data-driven, or hybrid models, also utilizing ML techniques like attention-based neural networks to enhance the performance of the model. In order to lessen the destructive effects of forest fires, this initiative aims to promote the development of more precise prediction tools and effective management techniques. The survey expands its scope to address the practical needs of numerous stakeholders. Access to enhanced early warning systems enables decision-makers to take prompt action. Emergency responders benefit from improved resource allocation strategies, strengthening the efficacy of firefighting efforts.

Keywords: artificial intelligence, deep learning, forest fire management, fire risk assessment, fire simulation, machine learning, remote sensing, wildfire modeling

Procedia PDF Downloads 81