Geodesic
Cleaning sensor data in intelligent heating control system
Vestnik Ûžno-Uralʹskogo gosudarstvennogo universiteta. Seriâ Vyčislitelʹnaâ matematika i informatika, Tome 10 (2021) no. 3, pp. 16-36

Voir la notice de l'article provenant de la source Math-Net.Ru

Sometimes, smart heating control applications are partially equipped with missing values and outliers in the sensor data due to software/hardware failures/human errors. To provide an effective analysis and decision-making, erroneous sensor data should be cleaned by imputation of missing values and smoothing outliers. In this paper, we present a case of the Smart Heating Control System (SHCS) installed in the South Ural State University, and describe the structure and development principles of Data Cleaning Module (DCM) of the system. We implement DCM through data mining and neural network technologies as a set of the following subsystems. The preprocessor extracts raw data from the system's data warehouse and prepares a training data for further processing. Predictor provides Recurrent Neural Network (RNN) to forecast the next value of a sensor based on its historical data. Reconstructor determines if the current value of a sensor is an outlier, and if so, imputes it by the synthetic value from Predictor. Finally, Anomaly Detector subsystem discovers anomalous sequences in the sensor data. In the experiments on the real sensor data, DCM showed relatively high and stable accuracy as well as adequate detection of anomalies.
Keywords: heating systems, control systems, cleaning, process control, temperature sensors, flowmeters, temperature measurement. 
@article{VYURV_2021_10_3_a1,
     author = {M. L. Zymbler and Ya. A. Kraeva and E. A. Latypova and E. V. Ivanova and D. A. Shnayder and A. A. Basalaev},
     title = {Cleaning sensor data in intelligent heating control system},
     journal = {Vestnik \^U\v{z}no-Uralʹskogo gosudarstvennogo universiteta. Seri\^a Vy\v{c}islitelʹna\^a matematika i informatika},
     pages = {16--36},
     publisher = {mathdoc},
     volume = {10},
     number = {3},
     year = {2021},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/VYURV_2021_10_3_a1/}
}
TY  - JOUR
AU  - M. L. Zymbler
AU  - Ya. A. Kraeva
AU  - E. A. Latypova
AU  - E. V. Ivanova
AU  - D. A. Shnayder
AU  - A. A. Basalaev
TI  - Cleaning sensor data in intelligent heating control system
JO  - Vestnik Ûžno-Uralʹskogo gosudarstvennogo universiteta. Seriâ Vyčislitelʹnaâ matematika i informatika
PY  - 2021
SP  - 16
EP  - 36
VL  - 10
IS  - 3
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/VYURV_2021_10_3_a1/
LA  - ru
ID  - VYURV_2021_10_3_a1
ER  - 
%0 Journal Article
%A M. L. Zymbler
%A Ya. A. Kraeva
%A E. A. Latypova
%A E. V. Ivanova
%A D. A. Shnayder
%A A. A. Basalaev
%T Cleaning sensor data in intelligent heating control system
%J Vestnik Ûžno-Uralʹskogo gosudarstvennogo universiteta. Seriâ Vyčislitelʹnaâ matematika i informatika
%D 2021
%P 16-36
%V 10
%N 3
%I mathdoc
%U http://geodesic.mathdoc.fr/item/VYURV_2021_10_3_a1/
%G ru
%F VYURV_2021_10_3_a1
M. L. Zymbler; Ya. A. Kraeva; E. A. Latypova; E. V. Ivanova; D. A. Shnayder; A. A. Basalaev. Cleaning sensor data in intelligent heating control system. Vestnik Ûžno-Uralʹskogo gosudarstvennogo universiteta. Seriâ Vyčislitelʹnaâ matematika i informatika, Tome 10 (2021) no. 3, pp. 16-36. http://geodesic.mathdoc.fr/item/VYURV_2021_10_3_a1/