RTU Research Information System
Latviešu English

Innovation application: Weather Prediction Algorithm Based on Historical Data Using Kalman Filter

Title Weather Prediction Algorithm Based on Historical Data Using Kalman Filter
Abstract

We offer a new method of collection and processing of meteorological data of meteorological service. This operation is based on observations and correction of numerical weather forecast errors by using a new algorithm. This algorithm considerably increases the accuracy of the short-term forecast of external air temperature. The algorithm ensures correction of predicted air temperature for the next three hours. Processing of temperature data using Kalman Filter allows to decrease errors of the predicted temperature.

Keywords Kalman filter, weather prediction, Python
Authors Nikolajs Bogdanovs
Romualds Beļinskis
Aleksandrs Ipatovs
Department (13100) Telekomunikāciju institūts
Statistical Classification of Economic Activities, NACE 2 Manufacture of computer, electronic and optical products
Computer programming, consultancy and related activities
Description of the technology

The predictive methodology is based on parameter refinement procedures and it is able to adjust the heating curve in order to obtain required supply temperature consistent to the actual demand for heating in room. Adjustment coefficients for modulating the heating curve have been refined by the test and error procedure. Overall results show the potential of energy savings increase by ~ 6-8% at significant decrease in room temperature deviation from settings. Based on this work, recommendations will be made for the building owners about the possibilities for exploiting heating engineering systems. The accuracy improving algorithm has been implemented using programming language Python.

Fig. 1. Application of the weather prediction algorithm.

Applications The predictive methodology is based on parameter refinement procedures and it is able to adjust the heating curve in order to obtain required supply temperature consistent to the actual demand for heating in room. Adjustment coefficients for modulating the heating curve have been refined by the test and error procedure.
Advantages

Building owners will receive recommendations for more efficient use of heating systems.

Technology Readiness Level Experimental proof of concept
Partnership offer • Service contracts.
Publications
ID 129
Contact information Linda Šufriča, e-mail: inovacijas@rtu.lv; phone.: 28442736