Contributors:Simon Moeller, Amelie Bauer, Ines Weber, Franz Schröder, Hannes Harter
Variance Inflation factors for the OLS regression as presented in Table 5 and stepwise OLS regressions for October to April.
Contributors:Mohammad Royapoor, Sara Walker, Charalampos Patsios, Peter Davison, Mehdi Pajouhesh
Data in summary:
1- Building total B side: This is metered data from one of two mains busbars that supplies all none-emergency services and HVAC equipment
2- Building total A side: This is metered data from the second of two mains busbars that supplies all emergency services including fire safety, comm rooms, emergency lighting and public announcement. It also is connected to a PV array with peak electrical supply of around 33kWe.
3- Half hourly building demand and deferrable load breakdowns: This is processed data that includes building total and HH instances of deferrable loads for all sub-categories of loads considered in this work. It also includes HH instances of PV generation, and outside air temperature.
4- Early morning ramp rates following plant start-up: This is a file containing the difference between two instantaneous recordings of total building electricity consumption that shows the continuous fluctuation in total electricity demand in the building.
5- CO2-raw (Typical office): This files contains actual CO2 data in an office that represents typical space occupant density in the case study building.
6- CO2-raw (worst case): This files contains actual CO2 data in a teaching space that represents the highest observed space occupant density in the case study building.
7- Warming and cooling rates in the worst case zones: This file include actual data describing the operational temperature in the worst case zones most prone to overheating in summer and excessive heat loss in winter.
Contributors:Siu C.Y., Liao Z.
Converted Weather Data in EPW format for Toronto Pearson Airport from CWEEDs weather data in WYEC3 format.
Contributors:Ashkan Negahban, Avinash Pallikere, Parhum Delgoshaei, Robin Qiu
The following zip file contains the codes and research data for the implementation of the simulation optimization framework proposed in the associated paper. Detailed step-by-step instructions for running the code is provided in the "Instructions.docx" file. You will need both MATLAB and GAMS (with an MIP solver) installed on your computer to be able to use the code and replicate the results presented in the paper.
Contributors:Hannes Harter, Philipp Geyer, Werner Lang, Patricia Schneider-Marin, MANAV MAHAN SINGH
Life Cycle Inventory (LCI) data of input data for all Building Shapes and Building Development Levels (BDL 2-4) for the embedded energy calculation model, described in this study.
The Power_time_series.txt file contains the 34 electrical features related to power that were computed using current and voltage acquisitions made using our own acquisition system based on an Arduino MKR Zero microcontroller with a sampling frequency of 6.25 kHz (see reference "Design of an electricity consumption measurement system for Non Intrusive Load Monitoring", IEEE IREC 2019 from the same authors).
In the text file the power time series correspond to the active power P, its harmonic order Pk, where k is ranging between 1 and 15 and the sum of the harmonics PH. There are also the reactive power Q, its harmonic order Qk, where k is ranging between 1 and 15 and the sum of the harmonics QH.
All these 34 features are ordered by columns as follows:
P, P1, P2, P3, P4, P5, P6, P7, P8, P9, P10, P11, P12, P13, P14, P15, PH, Q, Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9, Q10, Q11, Q12, Q13, Q14, Q15, QH
These power time series correspond to the same scenario of 12 appliances (microwave, a DVD player, a fan, a screen, a vacuum, a waffle iron, a hair dryer, an iron, a flat iron, a mixer, a CFL and a LED lamp) that are randomly switched on/off every 3 seconds during almost one hour and half. This results in almots 1200 events of diferrent appliances.
The text file Events_time_stamping.txt contains the time stamps of each events that were hand-labelled after inspecting the signals. Timestamps of the labeled events were adjusted to match the transitions.
The text file Events_labelling corresponds to the labelling of each events (which appliance is responsable of the event).
Finally, the file MultivriateAbruptChangeDetectors.pdf is a repport containing the mathematicla derivation of each investigated detector in details.
Contributors:Mohamed Hamdy, Shady Attia, Sophie Schönfeldt Karlsen
Contributors:Giuseppina Buttitta, Donal Finn
The model is capable of creating stochastic multi-day occupancy profiles for building stock of different sizes and characterised by different shares of households belonging to the different occupancy categories identified in the UK. The model uses the Monte Carlo Markov Chain technique.
The occupancy categories are developed by the application of a data-mining clustering technique on data available from the UK Time Use Survey 2015. These categories are characterised by the following occupancy profiles:
1. Daily absence: unoccupied period from 09.00 to 04:00,
2. Working hours absence: unoccupied period from 08:20 to 18:10,
3. Lunchtime absence: unoccupied period from 11:10 to 16:10,
4. Constant presence 1,
5. Constant presence 2.
These occupancy categories are described in details in the associated paper and in a previous publication (https://doi.org/10.1016/j.enbuild.2019.05.056.).
In the associated paper the stochastic occupancy profiles are used as inputs in energy models of residential buildings, but the source code may be readily adapted for specific applications, with due acknowledgement to the authors.
Contributors:Alexander Stauch, Pascal Vuichard
Data used for the analysis presented in our article
Contributors:Trenbath K., Doherty B.
Raw_Data.csv is the time series data for the power consumption of a handful of device types in the Research Support Facility (RSF) at the National Renewable Energy Laboratory. The data is for the months of October, November, and December 2017 and was collected using Intellisockets smart plugs from Ibis Networks. The "Skyspark" column represents the plug load submeter for the B Wing East in the RSF.
Scaled_Supplemented_Skyspark_Ibis_RSF_B_Wing_East_OCt_Nov_Dec_2017_5min.csv presents raw data scaled by the estimated number of devices in the B Wing East. It also includes supplemental device estimates for devices that did not have smart plug data.All power values are in kW.
Single_Device_Avg_Power.csv is the same as the raw data devided by the number of devices monitored for each device type, meaning that the data is representing an average across all devices monitored of that type.
All power values are given in kW.