Building Power Demand Research Articles

Electricity peak shaving for commercial buildings using machine learning and vehicle to building (V2B) system

Reducing electricity peak demand is essential to maintain the balance between supply and demand side in electricity power markets, as well as reduce utility costs and environmental impacts. With growth in adoption of Electric Vehicles (EVs), there is an emerging opportunity to balance electrical power demand of buildings by storing electricity in EVs during low demand periods and discharging electricity into buildings during peak demand periods. Due to uncertainty in time and magnitude of peak demand, decision makers are always faced with a challenging task to identify optimal schedules for charging and discharging EVs to minimize peak electricity demand. This paper presents the development of a novel system that is capable of predicting day-ahead building electricity demand profile and identifying optimum schedule of charging and discharging EVs to minimize electricity peak demand. The system is designed to comply with planned EV trip schedules and minimum state of charge (SOC). The system consists of (1) machine Learning (ML) model to predict electrical power demand, and (2) demand management optimization model to identify optimal schedule for charging and discharging EVs. Four methods are explored to develop the ML model, including histogram-based gradient boosting, random forest, deep artificial neural network (DNN), and long short-term memory (LSTM). A case study of multi-tenant commercial building is analyzed to evaluate the performance of the system and demonstrate its new capabilities. The results of the case study shows that LSTM has the best performance in terms of mean absolute error, root mean square error, and mean absolute percentage error with average values of 7.44, 17.78, and 20.08 %, respectively. Five scenarios for shaving peak electricity demand, including combinations of two electric vehicles, a stationary battery, and a PV system are investigated. Scenarios including the stationary battery and the PV system are considered to evaluate the full potential of peak demand reduction in the case study building. The results of the demand management optimization model show up to 36 % reduction in peak demand using two EVs, one stationary battery, and PV system of 40 kW capacity. The key contributions that this study adds to existing knowledge are: (1) developing machine learning models to predict day-ahead electricity demand in 15-minute intervals, (2) integrating machine learning and optimization algorithms in identifying EV charging and discharging schedules to minimize utility cost by shaving peak energy demand, and (3) considering planned EV trips and minimum SOC requirements in identifying optimal charging and discharging of EVS to shave peak energy demand in buildings. The implementation of this system provides practical solutions for managing electricity demand in commercial buildings using EVs. By reducing energy consumption and promoting the innovative use of EVs, this system offers a sustainable approach to managing electricity demand in commercial buildings.

Building Power Demand Estimation on Unusual Days Considering Facility Operation Status and Number of Occupants

To optimally conduct capital investment and energy management for buildings with storage batteries and photovoltaic (PV) generation, it is essential to accurately predict the electricity demand not only on weekdays but also on holidays and event days. However, predicting the building power demand on holidays and event days is difficult because facility operations and the number of occupants in a building are different from weekdays. In this paper, we propose a method to estimate unusual power demand by considering the power demand structure of existing buildings. The proposed method is an application of disaggregation techniques based on linear regression. First, a power demand structure model is developed by calculating the facility status demand contribution and number of occupants from weekday operational data. Second, the daily power demand is estimated with high accuracy by applying hypothetical operational scenarios, such as holidays, to the model. Experimental data from real offices confirm that the proposed method could estimate the power demand during weekends and vacation periods when the attendance rate is reduced by half.

A review on integration of surging plug-in electric vehicles charging in energy-flexible buildings: Impacts analysis, collaborative management technologies, and future perspective

Today’s plug-in electric vehicle (PEV) technology is one of the important ways to address the dependence of fossil energy and greenhouse gas emissions. With the explosive growth of the number of PEVs worldwide, surging PEVs charging has offered new opportunities and challenges for large amounts of existing buildings where more and more charging infrastructures have been added. By means of microgrid technology, energy-flexible buildings are nowadays integrated with multiple distributed energy sources, energy storages and various types of consumers, allowing the possibility of providing measures to mitigate the growing impact of surging PEVs charging demand and accommodate more PEVs in buildings. Collaborative management between PEVs and energy-flexible buildings is attracting more attention of energy professionals and provides incentives to make more efforts in developing more extensive and optimal control methods for the integrated building energy systems. However, there is little review focusing on the impact analysis and co-management technologies on PEVs charging in buildings. This paper firstly provides an overview of the literatures related to the impact of PEV charging on the building power demand profile and the associated distribution transformer performance as well as the key impact factors. Additionally, a comprehensive review and analysis are conducted on the state-of-the-art researches with regard to the co-management of PEVs and energy-flexible buildings with various energy flexibility sources for the co-benefits of the integrated building energy system. The further efforts needed for developing more applicable co-management methods for facilitating surging penetration of PEVs in buildings are also addressed.

Evaluating performance of different generative adversarial networks for large-scale building power demand prediction

As an unsupervised-learning data-driven model, Generative Adversarial Networks (GANs) have recently attracted a lot of attention for various applications. There is potential to apply GANs for large-scale building power demand prediction, which is needed for power grid operation. However, there are many GAN variations and it is unclear which GAN is suitable for this application. To answer this question, this paper identifies five promising GANs (Original GAN, cGAN, SGAN, InfoGAN, and ACGAN) and evaluates their performance for predicting building power demand at a large scale. Physics-based building energy models are developed to generate training and reference data. A new evaluation indicator that combines accuracy and reproducibility is proposed to evaluate the performance of different GANs in predicting building power demand. The results show that SGAN and InfoGAN are not suitable because they cannot control the number of generated building samples for different building types. The prediction performance among the Original GAN, cGAN, and ACGAN can vary depending on training sample sizes and number of building types. If the training sample size is sufficiently large, Original GAN and cGAN can predict building power demand more accurately than ACGAN with the same number of samples. If training samples are limited, Original GAN provides better accuracy than cGAN and ACGAN. When the number of building types increase, the prediction accuracy increases for cGAN, decreases for ACGAN, and remains the same for Original GAN. As a result, cGAN and Original GAN are recommended for large-scale building power demand prediction.

Predictive analytics for building power demand: Day-ahead forecasting and anomaly prediction

The power demand forecast for buildings can generate useful day-ahead predictions in power system planning and operation. However, the information in the forecast needs to be interpreted by a person with domain expertise. Moreover, automated interpretation of upcoming abnormal behaviors needs ground-truth labeling, but labels are not always available from power meter data. In this paper, we propose a novel Predictive Power Demand Analytics Methodology (PPDAM), based on deep neural networks and symbolic aggregate approximation, to predict the pattern profile of power demand in a building and upcoming normal (motif) and anomalous (discord) behaviors. The experimental results indicate that a power forecast could be mapped as different foreseeable demand patterns, each with a specific probability of occurrence. The reliability of anomaly prediction is evaluated by a classification test of which the accuracy is 88% and the F1 score is 87.38%. The outcomes of this work could provide building operators with a solution to derive latent information in power consumption data. The derived information could be used to improve the working conditions of the building’s power system.

Impact of electric vehicle charging on the power demand of retail buildings

• Fast charging for electric vehicles will be more common at retail sites in the future. • Electric vehicle stations with high power levels can dwarf a building's power demand. • Electric vehicles may arrive at sites to charge when building AC loads are high. • Electricity bills increase most for cold climates and in high-demand-charge scenarios. As electric vehicle penetration increases, charging is expected to have a significant impact on the grid. Electric vehicle charging stations will greatly affect a building site's power demand, especially with the onset of fast charging with power levels as high as 350 kW per charger. Here, we assess how electric vehicle charging stations would impact a retail big box grocery store, exploring numerous station sizes, charging power levels, and utilization factors in various climate zones and seasons. We measure the effect of charging by assessing changes in monthly peak power demand, electricity usage, and annual electricity bill, computed using three distinct rate structures. We find that an electric vehicle station has the potential to dwarf a big box building's power demand if behind the same meter, increasing monthly peak power demand at the site by over 250%. Cold-climate areas paired with rate structures incorporating high demand charges are most susceptible for significant changes to the annual electricity bill, with increases as high as 88%.

A modeling framework to examine photovoltaic rooftop peak shaving with varying roof availability: A case of office building in Hong Kong

The United Nations Sustainable Development Goal 7 (UNSDG 7) “Affordable and Clean Energy” strives to increase renewable energy share in energy consumption. Building-integrated photovoltaics generate renewable energy and offset fossil fuels carbon emissions. As roof space availability varies among different buildings, the potential to utilize rooftops as photovoltaic systems is uncertain. This study's aim is justified by the need to develop a systematic modeling framework to examine photovoltaic rooftops with varying roof availability and to propose peak shaving strategies for reducing peak load and carbon emissions. The framework incorporates computational simulations to model the building power demand and evaluate different peak shaving strategies. By identifying building peak loads, peak shaving offsets part of the loads with photovoltaic electricity. This framework can be applied globally given suitable data, and is demonstrated on a 10-story reference office building with photovoltaic installations occupying 10%, 30%, and 50% of the roof area. In this demonstration, 9 peak shaving strategy options are considered for covering a varying number of peaks that last for different durations on weekdays. Peak load reduction and carbon emission saving in each option are assessed, optimal peak shaving options are identified according to seasonal changes and available area for photovoltaics. For example, when 50% of the roof's area is available, a full-office-hour strategy is proposed in summer, saving approximately 595 kg CO2 per weekday. In winter, solar generation is reduced. A proposed strategy covering full-office-hour except lunch hours yielded an estimated 271 kg CO2 saving per weekday. Based on these findings, policies are recommended to encourage utilization of existing rooftops, construct photovoltaic-ready roofs on new buildings, and guide urban planning to avoid excessive shading.

Prediction of office building electricity demand using artificial neural network by splitting the time horizon for different occupancy rates

Due to the impact of occupants’ activities in buildings, the relationship between electricity demand and ambient temperature will show different trends in the long-term and short-term, which show seasonal variation and hourly variation, respectively. This makes it difficult for conventional data fitting methods to accurately predict the long-term and short-term power demand of buildings at the same time. In order to solve this problem, this paper proposes two approaches for fitting and predicting the electricity demand of office buildings. The first proposed approach splits the electricity demand data into fixed time periods, containing working hours and non-working hours, to reduce the impact of occupants’ activities. After finding the most sensitive weather variable to non-working hour electricity demand, the building baseload and occupant activities can be predicted separately. The second proposed approach uses the artificial neural network (ANN) and fuzzy logic techniques to fit the building baseload, peak load, and occupancy rate with multi-variables of weather variables. In this approach, the power demand data is split into a narrower time range as no-occupancy hours, full-occupancy hours, and fuzzy hours between them, in which the occupancy rate is varying depending on the time and weather variables. The proposed approaches are verified by the real data from the University of Glasgow as a case study. The simulation results show that, compared with the traditional ANN method, both proposed approaches have less root-mean-square-error (RMSE) in predicting electricity demand. In addition, the proposed working and non-working hour based regression approach reduces the average RMSE by 35%, while the ANN with fuzzy hours based approach reduces the average RMSE by 42%, comparing with the traditional power demand prediction method. In addition, the second proposed approach can provide more information for building energy management, including the predicted baseload, peak load, and occupancy rate, without requiring additional building parameters.

The Effect of Deep Energy Retrofit on The Hourly Power Demand of Finnish Detached Houses

This study examines how the energy renovation of old detached houses affects the hourly power consumption of heating and electricity in Finland. As electrification of heating through heat pumps becomes more common, the effects on the grid need to be quantified. Increased fluctuation and peak power demand could increase the need for fossil-based peaking power plants or call for new investments to the distribution infrastructure. The novelty in this study is the focus on hourly power demand instead of just annual energy consumption. Identifying the influence of building energy retrofits on the instantaneous power demand can help guide policy and investments into building retrofits and related technology. The work was done through dynamic building simulation and utilized building configurations obtained through multi-objective optimization. Deep energy retrofits decreased both the total and peak heating power consumption. However, the use of air-source heat pumps increased the peak power demand of electricity in district heated and wood heated buildings by as much as 100%. On the other hand, peak power demand in buildings with direct electric heating was reduced by 30 to 40%. On the building stock level, the demand reduction in buildings with direct electric heating could compensate for the increase in the share of buildings with ground-source heat pumps, so that the national peak electricity demand would not increase. This prevents the increase of demand for high emission peaking power plants as heat pump penetration rises. However, a use is needed for the excess solar electricity generated by the optimally retrofitted buildings, because much of the solar electricity cannot be utilized in the single-family houses during summer.

Assessment of peak power demand reduction available via modulation of building ventilation systems.

Peak power demand strains electrical grids and increases cost of electricity generation, transmission and distribution infrastructure. Many studies have examined ways of reducing peak power demand, including modification of room air temperature setpoints or reduction of lighting levels. However, very few or no studies have examined the peak power reduction resource offered by temporary curtailment of building ventilation systems. For this reason, we conducted a simulation campaign in which we examined the resource offered by temporary ventilation curtailment in commercial buildings of different use types across the United States and in residences in the state of California, with the essential constraint that any changes resulted in air quality acceptable to occupants through additional ventilation to compensate for the curtailment. To do this, we employed previously validated building models implemented in the airflow and contaminant transport tool CONTAM and building thermal and systems modeling tool EnergyPlus, in some cases co-simulated. Results show savings are highly dependent on building type and climate but range from 0–2 W/ft2 and up to 40% of total peak building power demand. Depending on building type, this power shed can be conducted for 1.5–8 h before acute exposure or odor concerns are expected, assuming a safety factor of 2. This reduction is of the same magnitude as that offered by thermal control strategies such as setpoint increase, or from lighting reduction strategies.

One‐day ahead predictive management of building hybrid power system improving energy cost and batteries lifetime

In recent times, the energy consumed by buildings facilities became considerable. Efficient local energy management is vital to deal with building power demand penalties. This operation becomes complex when a hybrid energy system is included in the power system. This study proposes new energy management between photovoltaic (PV) system, Battery Energy Storage System (BESS) and the power network in a building by controlling the PV/BESS inverter. The strategy is based on explicit model predictive control (MPC) to find an optimal power flow in the building for one-day ahead. The control algorithm is based on a simple power flow equation and weather forecast. Then, a cost function is formulated and optimised using genetic algorithms-based solver. The objective is reducing the imported energy from the grid preventing the saturation and emptiness of BESS. Including other targets to the control policy as energy price dynamic and BESS degradation, MPC can optimise dramatically the efficacy of the global building power system. The strategy is implemented and tested successfully using MATLAB/SimPowerSystems software, compared to classical hysteresis management, MPC has given 10% in energy cost economy and 25% improvement in BESS lifetime.

On the performance of LCC optimization software OPERA-MILP by comparison with building energy simulation software IDA ICE

From an economic point of view, it is crucial to minimize the life cycle costs (LCC) of buildings undergoing energy renovations, hence an optimization approach is needed. Building energy use and power demand as well as energy efficiency measures are important issues while performing an LCC optimization. Thus it is of great importance to accurately predict the building energy use and power demand before and after energy renovation.This paper aims to address the performance of an in-house LCC optimization software, OPERA-MILP, which has a rather fast optimization procedure. The aim is fulfilled through comparison with building energy simulation software IDA ICE before and after cost-optimal energy renovation. Three historic buildings with different layout and construction properties are used as a case study.The results show good agreement in the calculations of buildings' power demand and energy use between OPERA-MILP and IDA ICE. The percentage difference in calculated annual energy use and buildings' power demand with OPERA-MILP compared to IDA ICE is shown to be maximum 11% and 8% for the studied climate zones, respectively. Total impact on LCC is estimated to be equal to or less than 8%.

Building power demand forecasting using K-nearest neighbours model – practical application in Smart City Demo Aspern project

Following the ongoing transformation of the European power system, in the future, it will be necessary to locally balance the increasing share of decentralised renewable energy supply. Therefore, a reliable short-term load forecast at the level of single buildings is required. In this study, we use a forecaster, which is based on K-nearest neighbours approach and was introduced in an earlier publication, on three buildings of Smart City Demo Aspern project. The authors demonstrate how this forecaster can be applied on different buildings without any manual setup or parametrisation, showing that it is viable to replace load-profiling solutions for predicting electricity consumption at the level of single buildings.

A Comparative analysis of system efficiency for AC and DC residential power distribution paradigms

This paper presents a comparative efficiency study for the use of Direct Current (DC) and Alternating Current (AC) power in the residential distribution power system. The U.S. Residential Energy Splits data for year 2015 from ‘Buildings Energy Data Book’ has been used for modelling of residential buildings. Distribution system models have been made for both AC and DC powers. The system efficiency values have been tested for different cases. The AC system was found to be superior to DC with a minimum efficiency advantage of around 2% and a maximum of around 6%. Subsequently, the building air-conditioning loads were assumed to incorporate the modern concept of Variable Speed Drives (VSD) which use an intermediate DC power stage. If the building power supply is DC rather than AC, then the intermediate AC/DC conversion stage may be avoided. Taking the air-conditioning as a DC load roughly doubles the DC power demand of building. The DC system showed slightly (around 1%) higher efficiency for the VSD air-conditioning case. Back in history, it was apparently the efficiency of power transfer that allowed AC to oust DC, and now, once again efficiency may be a decisive factor for the reignited competition between AC and DC.

A method of determining the thermal power demand of buildings connected to the district heating system with usage of heat accumulation

The paper presents a new method of determining the thermal power demand of buildings connected to the district heating system, which included the actual heat demand and the possibility of balancing the thermal power using the thermal storage capacity of district heating network and internal heat capacity of buildings. Moreover, the analysis of the effect of incidence of external air temperature and duration of episodes with the lowest outdoor temperatures on the thermal power demand of district heating system was conducted.

Experimental determination of parameters in models of indoor air temperature response to reduction in supplied energy

To diminish peaks of energy power demand in buildings, we need to determine parameters in transmittance models that describe indoor temperature response to changes in heating. This article discusses first- and second-order models as well as an integration model. Calculations deal with indoor temperature drop in the range to −2 K following the heating power reduction for a period of 4–6 h. The second-order model was found to be the most accurate in showing the decrease in temperature. The calculated values of parameters vary within wide ranges, with significant differences occurring even with values obtained in consecutive measurement series. It was shown that the use of averaged values produce large result errors in the predictions of indoor temperature changes. This renders an offline method useless and promotes an online method. The average value of the time constant in the first-order model correlates well with the value of the time constant calculated as a ratio of the heat storage mass to the room heat loss coefficient. This is true if we take into account exclusively this part of the building partition layer that contributes to the heat storage.

Building power demand forecasting

Abstract Buildings acting as flexible loads have been often proposed to mitigate the volatility of renewable energy sources. Thereby, an accurate short-term demand forecast is indispensable for effective demand side management. At the same time, standardized load profiles, commonly used in the distribution grid, are inadequate for load forecasting within building domain. For this PhD, project a novel short-term forecasting model is proposed for that domain. It considers not only residual load, but also scheduled demand response as well as the PV-generation of the building. Moreover, it is not building specific and is, therefore, suitable for area-wide application within building domain.

Preemptive Demand Response Management for Buildings

A building energy management system (BEMS), which forms an integral part of a smart grid, enables building operators to monitor, manage, and control the energy utilized in their buildings, thus reducing the demand and consumption of energy. Building operators are responsible for the day-to-day maintenance and operation of their buildings’ heating, cooling, mechanical, and electrical systems. This paper proposes an intelligent preemptive demand response management (DRM) using the BEMS to ensure contracted capacity or demand limit (CC/DL) is not exceeded and at the same time reduce energy consumption in buildings. In this paper, dynamic electric vehicle (EV) charge scheduling, speed control of air conditioning (AC) system’s variable speed drive (VSD), and priority-based load shedding are considered in the DRM program. The performance of the proposed DRM program to keep the building power demand within the CC/DL and reduce the energy consumption is tested and analyzed using the BEMS.

Building power demand response methods toward smart grid

Smart grid has been drawing attention particularly when renewable generations are integrated. In order to ensure high power reliability and energy efficiency in an electrical grid, research and application has been conducted at power supply side to solve the grid critical issues: peak load and power imbalance. However, as the major end-users at power demand side, buildings can also play a significant and cost-effective role by making use of their power demand responses. Different demand response programs (e.g., time- and incentive-based) have been developed and applied for encouraging the end-users to change their energy usage behaviors expected by the grid. Generally, buildings are able to limit and/or shift the power demands according to their own considerations under the specific incentives. A comprehensive review on the building power demand response methods is still missing, although research and application has been investigated and conducted on power demand aspects concerning the building system configuration and the control strategies of power demand optimization. This article, therefore, presents a comprehensive review on the strategies, impacts, and benefits of building power demand response in a grid to systematically evaluate and make better use of their demand response potentials. The possibility of developing proper building power demand response strategies for offline and online applications of the smart grid is also discussed.

A Novel Air-conditioning System for Proactive Power Demand Response to Smart Grid

This paper presents a novel air-conditioning system with proactive demand response to smart grid. The system consists of a chilled water storage system (CWS) and a temperature and humidity independent control (THIC) air-conditioning system. Using this system, building power demand can be flexibly controlled as desired by implementing two demand response strategies: demand side bidding (DSB) strategy and demand as frequency controlled reserve (DFR) strategy, in respond to the day-ahead and hour-ahead power balance requirements of the grid, respectively. Considerable benefits can be achieved for both power companies and end-users under incentive pricing mechanisms. A case study concerning on the demand response performance of the proposed system is also conducted in an office building.