Author: jeantan

XMYs for HVAC: statistical analysis

eXtreme Meteorological Year (XMY) data for HVAC represents conditions that produce an extremely high or low building energy consumption across an entire year, and are important for building energy simulations that can help us gain understanding on how a building performs in an “extreme” year or a year of extreme seasons (like a hot, humid summer and a cold wet winter).

In our development of XMY data for HVAC, we ran historical weather data for Canberra and Brisbane through EnergyPlus simulations to obtain heating and cooling energy data for our three archetype buildings: Supermarket, 3-storey and 10-storey office building. Preliminary analysis of the energy data indicate a reasonable correlation between the three building types, and the results fit a rough Gaussian distribution (bell curve), reassuring us that standard statistical techniques can be applied.

When performing statistical analyses, probability values (P-values) provide an insight into the statistical likelihood of a dataset or event. In our study, P-values of annual HVAC energy consumption were calculated by applying each calendar year of historical weather to estimate the average energy consumption across the three building archetypes in Canberra and Brisbane, then calculating the mean and standard deviation of the result.

Using the empirical rule, in the knowledge that the data (approximately) fits a Gaussian distribution, we estimate that 68% is within one standard deviation of the mean, 95% is within two standard deviations of the mean and 99.7% of the data is within three standard deviations of the mean. The P01, P10, P90 and P99 data are those years that would result in energy consumption that is expected to exceed 1%, 10%, 90% and 99% of the cases (respectively) in a temporal sample.

In the figure below we have ranked the average annual energy use across the three archetypes from lowest to highest, and inserted the P-values that arise from this distribution (green, yellow, orange and red bars).

One option for creating a representative P-value climate data set is to take the historical year that most closely matches the target. So, for example, we know that a P90 year results in slightly more than 20 kWh/m2 (averaged across the three archetypes). The closest historical year is 1992, which resulted in marginally less consumption than the P90 target. In fact, for this distribution, the probability of exceeding the energy consumption of 1992 is 91.9 per cent.

The closest historical data to the targeted P-values are listed below:

  • P99 is 0.988 (1996)
  • P90 is 0.919 (1992)
  • P10 is 0.127 (2015)
  • P01 is 0.014 (2017)

We think we can improve on this.

Our next step in developing XMY data for HVAC will be to devise a technique to concatenate a series of months to create an artificial year of 12 months which more closely results in the target consumption at the P1, P10, P90 and P99 level. In the process, we will remain alert for lessons indicating how best to synthesize years which may not be realistic but which will allow simulators to evaluate high heating months/seasons with high cooling months/seasons in the same 12 month simulation.

This data will be used in simulations to test the robustness of building designs. Potential applications include risk assessments for developers, owners and regulators, as well as Green Star certification, NABERS Energy commitments and other areas of Energy Efficiency (for example NCC Section J compliance or Net Zero Emissions declarations).

Readers interested to engage with us in our development of XMY data are invited to make contact soon via exemplary.energy@exemplary.com.au

Developing XMYs for HVAC: Is one “extreme” valid for all?

In a previous article we outlined the concept of eXtreme Meteorological Year (XMY) as a hypothetical data set representing an extreme year of weather. An XMY for HVAC represents conditions that produce extremely high or low energy consumption across the entire year (note the focus on energy, as distinct from the extreme design conditions used for HVAC sizing which evaluate peak power demand).

As discussed previously, XMY data is important in building energy simulations to give us an insight into building energy performance as climatic conditions vary in the near future. If Representative Meteorological Year (RMY) data can tell us about the expected energy demand, the XMY data indicates the uncertainty due to climate variability.

In our work to develop XMYs for HVAC, one of the first questions to arise was whether an “extreme” year of climate should be considered as extreme for all building types.

To answer this, we ran historical weather data from 1990 to 2017 for Canberra and Brisbane through a series of EnergyPlus simulations to calculate heating and cooling energy data for our three archetype buildings: Supermarket, 3-storey and 10-storey office buildings. The analysis of the data were conducted for heating and cooling separately as well as combined, for individual building types and as a collective of buildings, over yearly and monthly time periods. Our results indicate a reasonable correlation between the three building types.

Correlation of HVAC energy use between the three building archetypes in Canberra

Pairwise comparison of the annual cooling energy use between the 3-storey and 10-storey office buildings in Canberra

We have also found that the energy results fit a rough Gaussian distribution (bell curve). This is important as it means that standard statistical techniques can be applied to the next stages of analysis which we will discuss further in a future post.

Readers interested to engage with us in our development of XMY data are invited to make contact soon via exemplary.energy@exemplary.com.au

CSIRO timing offset error in several weather elements

A critical review of the CSIRO Weather and Climate Data (Part 4) 

The CSIRO developed a set of Representative Meteorological Year (RMY) weather and climate data sets as the baseline for the organisation’s work in creating so-called “predictive” weather files that can be used to investigate the impact of climate change on building energy consumption (further information on this important work is available at https://acds.csiro.au/future-climate-predictive-weather). The RMY data are presented in the EnergyPlus Weather (.epw) format, transcribed from the Australian Climate Data Bank (ACDB) which is the basis for climate information in the Nationwide House Energy Rating Scheme (NatHERS) software tools and provides data for 70 geographic climate zones across Australia.

These climate data sets have been made freely available by the CSIRO since August 2021, and have become the de facto standard for building energy modellers seeking to demonstrate compliance with the energy efficiency requirements of the National Construction Code (NCC) along with a variety of other applications. Thus the accuracy of the RMY data sets has significant implications for the energy efficiency of Australia’s future building stock, and Exemplary Energy have undertaken a timely review the CSIRO weather and climate data sets ahead of the 2022 publication of the NCC.

Our critique has already highlighted several major shortcomings with the data sets, and discussions of the first three issues can be accessed by clicking on the following links:

1. Reliance on weather data ending in 2015 for the characterisation of a warming climate ;

2. A 30-minute error in solar data in the .epw format; and

3. A lack of coincident precipitation data despite the .epw format expressly inviting it.

The differences between the .epw and ACDB formats mean that the transcription is non-trivial and is grossly flawed regardless of the method. For example, solar radiation data in the ACDB format is timestamped at the centre of the time period (each hourly data point representing 30-minutes either side of the timestamp), whereas in the .epw format it is timestamped at the end of the period. These data should always be integrated from the original high frequency observations.

On the other hand, the transcription of instantaneous1 elements such as dry bulb temperature, dew point and wind speed should be straightforward. However, the CSIRO method appears to introduce a 60-minute offset error in several weather elements including dry bulb, dew point, atmospheric pressure and wind.

The issues arising from these errors need to be considered by policymakers and modellers alike. In mid-November 2021, we advised our colleagues at CSIRO and the Australian Department of Industry Science, Energy and Resources (DISER, responsible for the NCC) of these findings but they have yet to even add a warning to the distribution website. We will continue to work with them to avoid further propagation of the errors and offer our support to improve the data going forward. We urge users and policymakers to be mindful of these issues as modelling inaccuracies arising now are embedded in building operations for many years to come.

In the interests of full disclosure, we note that Exemplary Energy offers high quality climate and weather data, including ersatz future climate data, that avoid the issues of the CSIRO datasets. These are available for modellers demonstrating NCC compliance through the JV3 pathway (simulating a compliant reference building as well as the actual building being proposed), along with non-regulatory applications in design and optimisation and resilience testing of buildings and energy systems.

Notes:
1 Most observations (i.e. those other than solar radiation and precipitation) are actually averaged from a series of high frequency measurements taken over a period on the order of a few seconds, or in the case of wind observations a period of ten minutes. For our purposes, these are taken is representing the instantaneous conditions at the time of the timestamp.

Supporting Student Talent through the APSRC

Exemplary Energy had the great pleasure of supporting student talent at the recent Asia-Pacific Solar Research Conference (APSRC) by supporting three awards to the finest student researchers based on the quality of the extended abstract and the oral presentation or poster:

The Wal Read Memorial Award for Best Student Poster was awarded to Ryan Hall, a PhD candidate at UNSW Sydney, for his research titled “Top Cell Materials Via High-Throughput Materials Screening And Density Functional Theory Calculations”.

Exemplary Energy’s executive director Trevor Lee (fourth from the left) with the recipients of the APSRC 2021 student prizes. Ryan Hall, winner of the APSRC 2021 Wal Read Memorial Award, is second from the right. Milad Mohsenzadeh, recipient of the John Ballinger Award, is on the far left.

The John Ballinger Award for Best Buildings-related Student Presentation was awarded to Milad Mohsenzadeh, a PhD candidate for the Renewable Energy and Energy Efficiency Group at the University of Melbourne for his paper on “An Innovative Cost-effective Floating Solar Still with Integrated Condensation Coils”. Mr Mohsenzadeh’s PhD research project focusses on developing an innovative solar-driven desalination technique for producing drinking water in remote areas. He has published seven papers, and also holds an award-winning patent on innovative concentrated photovoltaic thermal solar collector.

The Monica Oliphant Prize for Best Female Student Poster went to Shukla Poddar, a PhD student in the School of Photovoltaics and Renewable Energy Engineering at UNSW Sydney, for her work on “Characterization of Solar Ramps across Australia”. Ms Poddar’s PhD research involves estimating future solar PV productivity over Australia at all timescales and understanding the impacts of climate change on future solar PV productivity using high resolution climate models.

Shukla Poddar, the recipient of the APSRC 2021 Monica Oliphant Prize

Exemplary Energy congratulates all the recipients of the APSRC 2021 awards, and wishes them all the very best in their research endeavours.