Anomaly detection in streaming environmental sensor data: A data-driven modeling approach
Environmental Modelling & Software
Increasing availability of industrial systems through data stream mining
Computers and Industrial Engineering
| Hi-index | 0.00 |
Development both of sensor and data communication technologies made a serious impact in development of SCADA systems in various domains. Possibility of access to sensor data in real time mode changed methods of monitoring and control. Data quality control is essential for next generation monitoring and control systems which acts on issues in data and performs model predictive control of equipment connected to data acquisition devices. Authors offer a concept of web-based real-time data quality evaluation and improvement system for energy domain. The system evaluates data in real-time based on a set of rules applied by the operator (identifying gaps, outliers, anomalies etc) and classifies cases into categories: errors, warnings and statistics. All cases are registered and tracked by the system. The system creates reports, improves data automatically (performs data imputation using curtain interpolation algorithm) or reacts on manual data modification. domain specific query language allows to create user-defined rules. It contains advanced data checks based on statistics and machine learning methods. The system is designed to be domain independent and to be able to process data from different data sources. The concept was designed and implemented in program code. The functionality of it can be extended by plug-ins written according to special requirements.