R8tech in cooperation with Department of Civil Engineering & Architecture and Department of Computer Systems of Tallinn University of Technology (TalTech) have been granted a 165000€ R&D grant by Archimedes Foundation’s NUTIKAS program. The project started in 2019 and is finished in 2021.
Automatic Methods for Increasing the Efficiency of Ventilation Devices Based on Machine Learning and Rule-Based Error Detection
The real performance of air handling units (AHU) is often different than initially planned. It causes several problems, which are difficult to identify. If the AHU does not work effectively, the energy consumption can be higher than planned and the desired parameters of indoor climate may not be achieved. The main problems during the operation of AHUs are:
a) The efficiency the heat recovery of the AHU is lower than planned. That means that the energy consumption is significantly higher it should be;
b) The temperature of supply air is too low or too high which can cause discomfort to occupants;
c) Frost formation of heat-exchangers cause problems during the operation of AHUs;
d) The filters of AHUs are not replaced at the optimum time that causes problems to the indoor air quality or increases the operating costs of AHUs;
e) The control systems of the AHUs are not working in the optimal way which increases the energy consumption of the AHU and decreases the lifetime of its mechanical components.
The listed problems are usually caused by human error during the design, installation commissioning phases and/or equipment wear off during the usage. The detection of the root cause of the problem while minimizing systems downtime during unfavourable hours requires highly qualified expertise in this field making the time critical technical servicing time-consuming and expensive.
The main objective of this project is to develop solutions to automatically assess the mechanical condition of the ventilation machines heat recoveries and filters and to estimate the impact of suboptimal operation of the heat recovery and filters