Sensor Data Driven Proactive Management of Infrastructure Systems

Online prediction of dynamical system behavior based on a combination of simulation data and sensor measurement data has numerous applications. Examples include predicting safe flight configurations, forecasting storms and wildfire spread, estimating railway track and pipeline health conditions. In such applications, high-fidelity simulations may be used to accurately predict a system’s dynamical behavior offline (“non-real time”). However, due to the computational expense, these simulations have limited usage for online (“real-time”) prediction of a system’s behavior. To remedy this, one possible approach is to allocate a significant portion of the computational effort to obtain data through offline simulations. The obtained offline data can then be combined with online sensor measurements for online estimation of the system’s behavior with comparable accuracy as the off-line, high-fidelity simulation. The main contribution of this paper is in the construction of a fast data-driven spatiotemporal prediction framework that can be used to estimate general parametric dynamical system behavior. This is achieved through three steps. First, high-order singular value decomposition is applied to map high-dimensional offline simulation datasets into a subspace. Second, Gaussian processes are constructed to approximate model parameters in the subspace. Finally, reduced-order particle filtering is used to assimilate sparsely located sensor data to further improve the prediction. The effectiveness of the proposed approach is demonstrated through a case study. In this case study, aeroelastic response data obtained for an aircraft through simulations is integrated with measurement data obtained from a few sparsely located sensors. Through this case study, the authors show that along with dynamic enhancement of the state estimates, one can also realize a reduction in uncertainty of the estimates.

Download Full-text

A Hybrid Approach: Dynamic Diagnostic Rules for Sensor Systems in Industry 4.0 Generated by Online Hyperparameter Tuned Random Forest

10.20944/preprints202007.0548.v1 ◽

2020 ◽

Author(s):

Ahlam Mallak ◽

Madjid Fathi

Keyword(s):

Random Forest ◽

Random Search ◽

Hybrid Approach ◽

Fault Detection And Diagnosis ◽

Data Driven ◽

Sensor Data ◽

Sensor Systems ◽

Hydraulic Test ◽

Model Based ◽

Detection And Diagnosis

In this work, A hybrid component Fault Detection and Diagnosis (FDD) approach for industrial sensor systems is established and analyzed, to provide a hybrid schema that combines the advantages and eliminates the drawbacks of both model-based and data-driven methods of diagnosis. Moreover, spotting the light on a new utilization of Random Forest (RF) together with model-based diagnosis, beyond its ordinary data-driven application. RF is trained and hyperparameter tuned using 3-fold cross-validation over a random grid of parameters using random search, to finally generate diagnostic graphs as the dynamic, data-driven part of this system. Followed by translating those graphs into model-based rules in the form of if-else statements, SQL queries or semantic queries such as SPARQL, in order to feed the dynamic rules into a structured model essential for further diagnosis. The RF hyperparameters are consistently updated online using the newly generated sensor data, in order to maintain the dynamicity and accuracy of the generated graphs and rules thereafter. The architecture of the proposed method is demonstrated in a comprehensive manner, as well as the dynamic rules extraction phase is applied using a case study on condition monitoring of a hydraulic test rig using time series multivariate sensor readings.

Download Full-text

Integrating Rule-Based Systems and Data Analytics Tools Using Open Standard PMML

Volume 1A: 35th Computers and Information in Engineering Conference ◽

10.1115/detc2015-46412 ◽

2015 ◽

Cited By ~ 2

Author(s):

Yunpeng Li ◽

Utpal Roy ◽

Y. Tina Lee ◽

Sudarsan Rachuri

Keyword(s):

Expert System ◽

Expert Systems ◽

Data Analytics ◽

Domain Knowledge ◽

Prediction Models ◽

Data Driven ◽

Sensor Data ◽

Formal Representation ◽

Rule Based ◽

Open Standard

Rule-based expert systems such as CLIPS (C Language Integrated Production System) are 1) based on inductive (if-then) rules to elicit domain knowledge and 2) designed to reason new knowledge based on existing knowledge and given inputs. Recently, data mining techniques have been advocated for discovering knowledge from massive historical or real-time sensor data. Combining top-down expert-driven rule models with bottom-up data-driven prediction models facilitates enrichment and improvement of the predefined knowledge in an expert system with data-driven insights. However, combining is possible only if there is a common and formal representation of these models so that they are capable of being exchanged, reused, and orchestrated among different authoring tools. This paper investigates the open standard PMML (Predictive Model Mockup Language) in integrating rule-based expert systems with data analytics tools, so that a decision maker would have access to powerful tools in dealing with both reasoning-intensive tasks and data-intensive tasks. We present a process planning use case in the manufacturing domain, which is originally implemented as a CLIPS-based expert system. Different paradigms in interpreting expert system facts and rules as PMML models (and vice versa), as well as challenges in representing and composing these models, have been explored. They will be discussed in detail.

Download Full-text

Multi-agent approach to distributed processing big sensor data based on fog computing model for the monitoring of the urban infrastructure systems

2016 International Conference System Modeling & Advancement in Research Trends (SMART) ◽

10.1109/sysmart.2016.7894540 ◽

2016 ◽

Cited By ~ 7

Author(s):

Danila Parygin ◽

Nikita Nikitsky ◽

Valery Kamaev ◽

Anna Matokhina ◽

Alexey Finogeev ◽

...

Keyword(s):

Fog Computing ◽

Distributed Processing ◽

Urban Infrastructure ◽

Sensor Data ◽

Infrastructure Systems ◽

Computing Model ◽

Multi Agent

Download Full-text

Sensor Data-Driven UAV Anomaly Detection using Deep Learning Approach

10.1109/milcom52596.2021.9653036 ◽

2021 ◽

Author(s):

Julio Galvan ◽

Ashok Raja ◽

Yanyan Li ◽

Jiawei Yuan

Keyword(s):

Deep Learning ◽

Anomaly Detection ◽

Data Driven ◽

Sensor Data ◽

Learning Approach

Download Full-text

A critical review of the data pipeline: how wastewater system operation flows from data to intelligence

Water Science & Technology ◽

10.2166/wst.2020.393 ◽

2020 ◽

Vol 82 (12) ◽

pp. 2613-2634

Author(s):

Jean-David Therrien ◽

Niels Nicolaï ◽

Peter A. Vanrolleghem

Keyword(s):

Current Practice ◽

Review Paper ◽

State Of The Art ◽

Data Driven ◽

Sensor Data ◽

Measurement Technology ◽

Efficient Operation ◽

Smart Water ◽

Data Pipeline ◽

Systems Models

Abstract Faced with an unprecedented amount of data coming from evermore ubiquitous sensors, the wastewater treatment community has been hard at work to develop new monitoring systems, models and controllers to bridge the gap between current practice and data-driven, smart water systems. For additional sensor data and models to have an appreciable impact, however, they must be relevant enough to be looked at by busy water professionals; be clear enough to be understood; be reliable enough to be believed and be convincing enough to be acted upon. Failure to attain any one of those aspects can be a fatal blow to the adoption of even the most promising new measurement technology. This review paper examines the state-of-the-art in the transformation of raw data into actionable insight, specifically for water resource recovery facility (WRRF) operation. Sources of difficulties found along the way are pinpointed, while also exploring possible paths towards improving the value of collected data for all stakeholders, i.e., all personnel that have a stake in the good and efficient operation of a WRRF.

Download Full-text

Reaction Time Optimization Based on Sensor Data-Driven Simulation for Snow Removal Projects

Construction Research Congress 2018 ◽

10.1061/9780784481288.047 ◽

2018 ◽

Author(s):

Parinaz Jafari ◽

Emad Mohamed ◽

Mostafa Ali ◽

Ming-Fung Francis Siu ◽

Simaan Abourizk ◽

...

Keyword(s):

Reaction Time ◽

Data Driven ◽

Sensor Data ◽

Snow Removal ◽

Time Optimization

Download Full-text

Rapid Multi-Dimensional Impact Assessment of Floods

Sustainability ◽

10.3390/su12104246 ◽

2020 ◽

Vol 12 (10) ◽

pp. 4246 ◽

Cited By ~ 1

Author(s):

David Pastor-Escuredo ◽

Yolanda Torres ◽

María Martínez-Torres ◽

Pedro J. Zufiria

Keyword(s):

Social Media ◽

Impact Assessment ◽

Rural Areas ◽

Disaster Response ◽

Heterogeneous Data ◽

Data Driven ◽

Sensor Data ◽

Mobile Phone Data ◽

Aggregated Data ◽

The Impact

Natural disasters affect hundreds of millions of people worldwide every year. The impact assessment of a disaster is key to improve the response and mitigate how a natural hazard turns into a social disaster. An actionable quantification of impact must be integratively multi-dimensional. We propose a rapid impact assessment framework that comprises detailed geographical and temporal landmarks as well as the potential socio-economic magnitude of the disaster based on heterogeneous data sources: Environment sensor data, social media, remote sensing, digital topography, and mobile phone data. As dynamics of floods greatly vary depending on their causes, the framework may support different phases of decision-making during the disaster management cycle. To evaluate its usability and scope, we explored four flooding cases with variable conditions. The results show that social media proxies provide a robust identification with daily granularity even when rainfall detectors fail. The detection also provides information of the magnitude of the flood, which is potentially useful for planning. Network analysis was applied to the social media to extract patterns of social effects after the flood. This analysis showed significant variability in the obtained proxies, which encourages the scaling of schemes to comparatively characterize patterns across many floods with different contexts and cultural factors. This framework is presented as a module of a larger data-driven system designed to be the basis for responsive and more resilient systems in urban and rural areas. The impact-driven approach presented may facilitate public–private collaboration and data sharing by providing real-time evidence with aggregated data to support the requests of private data with higher granularity, which is the current most important limitation in implementing fully data-driven systems for disaster response from both local and international actors.

Download Full-text