Adapting Fan Motor, VFD Efficiency Correlations to Better Predict Energy Use
From eSociety, September 2019
Sometimes a building’s overall energy consumption differs from simulation. A wrong assessment of fan efficiencies at low-load conditions can be one of the reasons why.
Fans can contribute significantly to a building’s energy consumption, and the overall efficiency of fans is the combination of three factors: mechanical, motor and variable frequency drive efficiencies, according to “Adaptation of Fan Motor and VFD Efficiency Correlations Using Bayesian Inference” in Science and Technology for the Built Environment.
Manufacturers usually provide fan mechanical efficiency curves for a broad operating range, but motor and variable frequency drive (VFD) efficiencies are generally given at rating conditions only. To represent part-load conditions, regularly used correlations estimate motor and VFD efficiency variations to evaluate overall electricity consumption, according to the paper.
The study tests existing correlations for motor and VFD efficiency as a function of load and speed. This is done by comparing manufacturer data for a vendor that has shared its detailed test data. The study also seeks to improve such correlations using Bayesian inference to fit the data.
Lisa Rivalin with Lawrence Berkeley National Laboratory explains the research’s significance.
1. What is the significance of this research?
This research aims to update variable frequency drive and motor efficiency curves using strong expertise and a few data. These correlations are widely used in energy performance simulation. A building is equipped with a large number of fans, which overall efficiency is a combination of three factors: mechanical, motor and variable frequency drive efficiencies. Errors on these efficiencies accumulate and can lead to a significant error on the overall energy performance of the system or building.
Thanks to manufacturers’ data, we were able to test and update the existing correlations and reduce the efficiency assessment error between simulation and real data by 80%.
2. Why is it important to explore this topic now?
Today, reducing the overall energy consumption of buildings is a critical point for stakeholders. For that reason, energy performance contracting has emerged as an effective solution to increase the energy efficiency of new and existing buildings, reducing the financial risk for customers. The goal of this type of contract is to guarantee that customers’ buildings will achieve a specified energy performance (i.e., minimum energy savings). To do so requires prediction of future energy consumption of the building, at the design stage, before construction or major retrofit. To this end, building energy simulations taking into account all the major energy-using components are performed.
The correlations traditionally used to assess overall fan efficiencies were developed 20 years ago. Since then, motor and VFD technologies and regulations have evolved to more efficient ones, especially at part-load conditions. Modern efficient buildings use part-load operation more often, to save energy. Therefore, modeling fan performance using these obsolete correlations leads to overestimating energy consumption.
3. Were there any surprises or unforeseen challenges for you when preparing this research?
Our biggest surprise was to notice that VFD correlations from 1999 were still valid with today’s products, contrary to motors’ correlations that needed an update.
4. What lessons, facts and/or guidance can an engineer working in the field take away from this research?
The idea for conducting this research came from engineers who wanted to understand why the overall energy consumption of the building was so different from their simulations. The wrong assessment of fan efficiencies at low loads was one of the reasons. A lesson we can take away is that traditional correlation used in the industry may need to be updated to take into account modern technologies.
5. How can this research further the industry’s knowledge on this topic?
We have used Bayesian inference to update the correlations because we had strong expertise. This method has two advantages. The first one is it combines prior knowledge with data, stating that this process can be used to update the correlation each time new data points are measured. We can, for instance, imagine an update of the correlations each time a manufacturer released a new motor technology.
The second advantage is that it provides not only updated correlation coefficients but also the level of confidence associated with it. Thus, the confidence result may be used in further statistical analysis in larger building energy models to conduct risk analysis for energy performance contracting, for instance.