algorithms

Article
Assessment of Solar Energy Generation Toward Net-Zero
Energy Buildings
Rayan Khalil 1 , Guilherme Vieira Hollweg 1 , Akhtar Hussain 2 , Wencong Su 1, * and Van-Hai Bui 1, *

1 Department of Electrical and Computer Engineering, University of Michigan—Dearborn,

Dearborn, MI 48128, USA; [email protected] (R.K.); [email protected] (G.V.H.)
2 Department of Electrical and Computer Engineering, Laval University, Quebec, QC G1V 0A6, Canada;
[email protected]
* Correspondence: [email protected] (W.S.); [email protected] (V.-H.B.)

Abstract: With the continuous rise in the energy consumption of buildings, the study and integration
of net-zero energy buildings (NZEBs) are essential for mitigating the harmful effects associated with
this trend. However, developing an energy management system for such buildings is challenging due
to uncertainties surrounding NZEBs. This paper introduces an optimization framework comprising
two major stages: (i) renewable energy prediction and (ii) multi-objective optimization. A prediction
model is developed to accurately forecast photovoltaic (PV) system output, while a multi-objective
optimization model is designed to identify the most efficient ways to produce cooling, heating, and
electricity at minimal operational costs. These two stages not only help mitigate uncertainties in
NZEBs but also reduce dependence on imported power from the utility grid. Finally, to facilitate
the deployment of the proposed framework, a graphical user interface (GUI) has been developed,
providing a user-friendly environment for building operators to determine optimal scheduling and
oversee the entire system.

Keywords: energy management system; forecasting model; GUI; multi-energy system; net-zero
energy buildings; optimization

Citation: Khalil, R.; Hollweg, G.V.;

1. Introduction
Hussain, A.; Su, W.; Bui, V.-H.
Assessment of Solar Energy The increasing demand for energy in buildings, driven by population growth and
Generation Toward Net-Zero Energy urbanization, has significantly raised energy consumption levels. Buildings are currently
Buildings. Algorithms 2024, 17, 528. responsible for consuming 74% of electricity in the U.S. [1], with projections indicating that
https://doi.org/10.3390/ building energy use will rise by 40% within the next two decades [2]. Such an escalation in
a17110528 energy consumption exacerbates environmental impacts, particularly through heightened
Academic Editor: Ming-Feng Ge
greenhouse gas (GHG) emissions. In response, the concept of net-zero energy buildings
(NZEBs) has been advocated as a vital solution designed to equalize the energy consumed
Received: 15 October 2024 with the energy produced on-site through renewable sources [3–5]. Beyond theoretical
Revised: 9 November 2024 frameworks, NZEBs offer a tangible approach to markedly reduce the carbon footprint of
Accepted: 14 November 2024 building infrastructures.
Published: 16 November 2024
Buildings can be classified into three major categories based on their primary use:
residential, commercial, and industrial. Numerous studies have been conducted on each
building type with the aim of achieving net-zero energy consumption. For example, it
Copyright: © 2024 by the authors.
is noted in ref. [6] that advancements in residential NZEBs focus on optimizing energy
Licensee MDPI, Basel, Switzerland. infrastructure connections, renewable energy sources, and energy-efficiency measures to
This article is an open access article reduce energy consumption and emissions. In ref. [7], a real-world residential building
distributed under the terms and is retrofitted and concluded that this can reduce greenhouse gas emissions by over 60%,
conditions of the Creative Commons with potential reductions of up to 96% through more advanced measures. Similarly, a
Attribution (CC BY) license (https:// net-zero energy management approach for commercial buildings is proposed in ref. [8]
creativecommons.org/licenses/by/ using renewable energy, pumped hydro storage, and hydrogen taxis, offering a framework
4.0/). for urban decarbonization by 2050. The key themes for commercial NZEB are discussed

Algorithms 2024, 17, 528. https://doi.org/10.3390/a17110528 https://www.mdpi.com/journal/algorithms

Algorithms 2024, 17, 528 2 of 19

in ref. [9], identifying key themes such as energy efficiency and life cycle assessment
through bibliometric and qualitative analysis. In ref. [10], the authors discuss region-
specific transition plans and policy packages to achieve net-zero industrial emissions by
mid-century, such as material efficiency, carbon pricing, and decarbonized technologies.
The authors in ref. [11] outlined emission reduction challenges in the cement industry
and proposed solutions across its value chain to achieve net-zero emissions and minimize
environmental impacts.
In addition, the challenges associated with NZEBs can be broadly categorized into
two distinct phases: the planning phase and the operation phase. The planning phase
focuses on site selection, design, energy modeling, and sizing of energy equipment. Mean-
while, the operation phase emphasizes real-time energy monitoring, regular maintenance,
and optimal energy management. The benefits of multi-energy system co-planning in
nearly zero-energy districts are discussed in ref. [12] while identifying key research gaps
related to temporal and spatial representations. An investment planning approach for
NZEB in Canada is proposed in ref. [13] examining the influence of geographical factors on
energy retrofits. The significance of planning NZESs for large energy consumers, like data
centers, is discussed in ref. [14] to facilitate the integration of volatile renewable energy
sources. The operational and energy management aspects of NZEBs are discussed in
refs. [15–17] where energy management of a real building in China is accessed in ref. [15],
the role of HVAC in achieving net zero is discussed in ref. [16], and the role of efficiency
measures is discussed in ref. [17].
However, the adoption of NZEBs encounters considerable challenges. Primarily
among these are the unpredictability of renewable energy sources and the complexities
associated with efficient energy management [18]. Solar power’s intermittency introduces
uncertainties in the building energy system, affecting the consistency of power output from
photovoltaic (PV) systems. To mitigate these issues, various predictive models have been
developed. For instance, the authors in ref. [19] have developed a hybrid machine-learning
model integrating an extreme learning machine with a genetic algorithm and a similar day
analysis for precise hourly PV power output predictions. This model has been validated
through robust performance metrics such as the coefficient of determination, mean absolute
error, and normalized root mean square error, demonstrating high accuracy and stability in
day-ahead PV power predictions. Another study in ref. [20] has utilized machine learning
tools to develop a PV output prediction model that includes data quality checks, a machine
learning algorithm, weather clustering, and accuracy assessments, resulting in enhanced
prediction capabilities when linear regression coefficients are applied.
Despite advances in forecasting accuracy, precisely predicting the PV output under
constantly changing weather conditions remains impossible, leading to operational com-
plexities in NZEBs. Numerous studies have tackled the optimal operation of NZEBs under
such uncertainties [21]. For instance, the authors in ref. [22] have developed a framework
combining multi-objective optimization with robustness analysis to design integrated build-
ing energy systems, employing two-stage stochastic programming for balancing economic
and environmental goals while validating robustness through Monte Carlo simulations.
The authors in ref. [7] have developed a multi-objective optimization process aiming to
minimize the operating GHG emissions and the life-cycle cost. The process has been
applied to a typical multi-residential building and tested in the four Greek climate zones.
Although existing research studies [19,20,22,23] have significantly focused on enhanc-
ing the accuracy of prediction models or optimizing NZEB operations, more studies are
still required to develop comprehensive operational frameworks for building energy man-
agement systems with the necessary features. Additionally, developing a user-friendly
graphical user interface (GUI) that provides real-time visualization of system status and
integrates both predictive and optimization functionalities could greatly assist building
operators in managing and controlling building energy systems more effectively.
To address these multifaceted challenges, this study proposes an optimization frame-
work specifically designed for NZEBs. The framework comprises two primary phases:
Algorithms 2024, 17, 528 3 of 19

accurately predicting renewable energy outputs, especially from PV systems, and strate-
gically optimizing energy production and consumption. The initial phase utilizes neural
network (NN)-based prediction models to forecast the output power of the PV system
accurately, taking into account a range of environmental and technical factors. The subse-
quent phase focuses on optimizing building systems for cooling, heating, and electricity
generation using mixed integer linear programming (MILP), aiming to minimize oper-
ational costs and maximize efficiency. This approach not only enhances the reliability
and performance of NZEBs but also reduces their dependence on external power sources,
thereby bolstering the sustainability of building operations. The integration of this model
promotes PV as an energy source for buildings, further providing a sustainable approach
to powering buildings. The increased use of renewable energy sources like PV will help
reduce harmful environmental impacts. This approach provides a feasible environmental
solution because it shows building operators that their buildings can efficiently operate
without the need for nonrenewable energy as the core energy provider. Additionally, the
implementation of a GUI facilitates intuitive and user-friendly interaction with the system,
enabling building operators to manage and optimize their energy resources efficiently. The
major contributions of this study are listed as follows:
• An operation strategy is proposed for NZEBs, including deep neural network-based renew-
able energy output predictions and a MILP-based optimization for multi-energy balance.
• Reliability and operational efficiency are significantly enhanced by accurately forecast-
ing PV output and optimizing building systems for cooling, heating, and electricity,
thus reducing dependency on external systems.
• A user-friendly GUI is developed for building operators, enabling effective energy
resource management and optimizing energy use in real time.
• The proposed operation offers a scalable and user-friendly solution that highlights the
feasibility of achieving NZEB objectives, further supporting the shift to sustainable
building operations.

2. System Models
2.1. System Configuration
Figure 1 shows the typical configuration and energy flow of a building energy system
(BES), which integrates various sources of energy to fulfill electric, heat, and cooling energy
demands. By combining the energy generated from these sources, the BES helps ensure
that the energy supply throughout the building is balanced and efficiently distributed.
This integration, which allows the system to harness energy from multiple sources, will
increase the reliability of the BES while reducing dependency on a single energy provider.
Not only will the system become more reliable, but there is also higher risk mitigation
associated with energy supply disruptions (e.g., power outages). The system presents a
way to seamlessly obtain energy from other energy sources in the building.
In the test system, the PV systems, distributed generators (DGs), and energy storage
systems (ESSs) are used to supply electricity to the building energy system, fulfilling all
electrical loads as well as powering other heating and cooling sources, namely the electric
heat pump, boiler, and electric chiller. The system is also connected to the utility grid,
which can help supply the entire system during periods of power shortage when the output
from local resources is insufficient. The utility grid and ESS both support bi-directional
power flow, allowing the system to efficiently adapt to fluctuating supply conditions and
energy demands. Surplus energy can be given back to these components, or the system can
draw energy from them when necessary. The heat generated by the boiler and electric heat
pump is used to meet the heat load and the heat energy consumed by the absorption chiller
(AC). The cooling load is fulfilled by the electric chiller (EC) and the AC. The operation of
the entire building’s energy is managed by a building energy management system.
 electric heat pump is used to meet the heat load and the heat energy consumed by the
absorption chiller (AC). The cooling load is fulfilled by the electric chiller (EC) and the
Algorithms 2024, 17, 528 AC. The operation of the entire building’s energy is managed by a building energy 4man-of 19
agement system.

Utility grid

DG1 DG2 … DGn ESS

PV
1

PV
2 Building Energy System
… Electric Electric
PV heat pump load
n

Boiler Heat load

Absorption
chiller
Electric Cooling
chiller load

Figure 1.
Figure 1. Energy
Energy building
building systems.
systems.

2.2.
2.2. Proposal
Proposal Optimization
Optimization Framework
Framework
The proposed optimization
The proposed optimization framework
framework includes two major
includes models:
two major a prediction
models: model
a prediction
of PV output
model power and
of PV output power a multi-objective optimization
and a multi-objective model model
optimization for the for
operation of entire
the operation of
building systems. These models are interdependent, as the output of the prediction
entire building systems. These models are interdependent, as the output of the prediction model
serves
model as inputasfor
serves configuring
input the building’s
for configuring operational
the building’s optimization.
operational The prediction
optimization. The pre-
model produces a forecast of expected output from a PV system
diction model produces a forecast of expected output from a PV system basedbased on environmen-
on envi-
tal factors. factors.
ronmental The detailed input/output
The detailed for both
input/output for prediction and optimization
both prediction and optimizationmodels are
models
summarized in Table 1.
are summarized in Table 1.
First, three prediction models were assessed and analyzed on their performance on
Table 1. Framework parameters.
PV generation data from Berlin, Germany, using the PVWatts API [23]. The prediction
models chosen were ARIMA, support vector
Model Type Inputsregression (SVR), and a feedforward
Outputs neural
network (FNN). To maintain consistency and accurate
City’s PV site generationcomparisons, the same input data
were used across all models. The input data used
Electric demands in the model were beam irradiance,
Predictions for 1, 3, and 7 days
Prediction
diffuse irradiance, ambient temperature,
Coolingwind speed, and date and
demands of PVtime. Theoutput
system output pa-
rameter was the output power of theHeating demands
PV system.
Operational cost of DGs Setpoint of DGs
Table 1. Framework parameters.
Optimization Operation boundary of DGs Energy flow
Load demand ESS
Model Type Inputs Outputs
City’s PV site generation
First, three prediction models were assessed and analyzed on their performance on
Electric demands Predictions for 1, 3, and 7
PV generationPrediction
data from Berlin, Germany, using the PVWatts API [23]. The prediction
Cooling demands days of PV system output
models chosen were ARIMA, support vector regression (SVR), and a feedforward neural
Heating demands
network (FNN). To maintain consistency and accurate comparisons, the same input data
were used across all models. The Operational
input datacost of in
used DGs Setpoint
the model were of DGs
beam irradiance,
diffuse irradiance, ambient temperature, wind speed, and date and time. flow
Optimization Operation boundary of DGs Energy The output
parameter was the output power of theLoad demand
PV system. ESS
After testing the three chosen models and gathering their actual and predicted output,
After testing
the following the three
testing error chosen
metrics,models and mean
including gathering their actual
absolute and predicted
error (MAE), out-
root mean
put, the error
squared following testing
(RMSE), anderror metrics,root
normalized including
mean mean
squared absolute error (MAE),
error (nRMSE), root mean
are calculated
for each prediction model. To better evaluate the developed models, these metrics are
also compared among one-day, three-day, and seven-day predictions. This study will
concentrate on the results for the one and three-day models, as the seven-day model was
included solely to demonstrate the scalability of the prediction and optimization model
over a longer time frame.
Algorithms 2024, 17, 528 5 of 19

Regarding the optimization model, we developed a MILP-based multi-objective opti-

mization model to find the most efficient way to operate the energy system to minimize
costs while at the same time ensuring that all energy demands from electricity, heating,
and cooling are met. Decision variables such as the generation output of the boiler, heat
pump, DGs, and chillers are the quantities that need to be optimized. To ensure that
optimal results are achieved by the model, power constraints were implemented. To find
the most efficient way to operate, the optimization model will explore many combinations
of decision variable values to end up with the smallest possible cost.
The project framework was developed to consider the variability present in envi-
ronmental parameters across different locations and climate zones, making it adaptable
for real-world applications. This adaptability allows the model to take in data specific
to an environment, retrain, and produce updated predictions suited to the building or
environment’s unique conditions. Despite the uncertainty in climate and renewable energy
generation, the option for continuous model retraining helps to increase the reliability of
its predictions for both power generation and operational costs. This energy management
framework would be particularly beneficial for commercial buildings, ideally the ones
equipped with rooftop solar/PV panels. Due to the model’s reasonable basis for scheduling
energy coverage, the commercial buildings would gain important insights into efficiently
meeting power demands.

2.3. Mathematical Models

In this section, a detailed MILP-based mathematical model is presented with a multi-
objective function aimed at minimizing building energy operation costs and reducing
dependency on the utility grid. Additionally, the operational constraints of DGs, ESSs,
and energy balances, including electrical, heating, and cooling systems, are also presented
in detail.
The following multi-objective Equation (1) is utilized in the optimization model for
optimal operation of energy building. The first term of the objective function presents the
operation cost of the entire building network, and the second term shows the dependence
of the building on the external utility grid.

T n  o n o
Min ∑ α. C1 ∗ P1 (t) + C2 ∗ P2 (t) + pGrid (t) ∗ PGrid (t) − PGrid
buy sell buy
(t) + β. PGrid (t) (1)
t =1

where : α + β = 1
This multi-objective function not only aims to minimize the operation cost of the entire
building energy system but also to reduce the building’s dependency on external energy
systems. The cost of power produced by DGs and the cost/profit of trading power with
the utility grid are calculated at every time step. When the objective function is minimized,
the optimal values for the decision variables are determined. This ensures efficient and
cost-effective usage of our building energy system, along with encouraging the use of
sustainable energy.
There are various constraints that must be fulfilled during the operation of buildings.
The given Equations (2)–(6) show the operational constraints and energy balance of a
system involving DGs, ESS, and energy flows between various components. The operation
boundary of DGs is given in (2), ensuring that their output stays within the minimum
and maximum limits. Equations (3) and (4) describe the charging and discharging power
constraints of the ESS. Equation (3) ensures that the charging power does not exceed
the maximum allowable charging rate considering the current SOC and the charging
loss. Similarly, Equation (4) shows the discharging power based on the current SOC and
discharging loss. SOC update is given in Equation (5), where the SOC at the next step
depends on the current SOC and the actual amount of charging and/or discharging power.
Finally, the power balance within the system is ensured by Equation (6), where the total
power generated from PVs, DGs, ESS discharging, and grid purchases must meet the total
Algorithms 2024, 17, 528 6 of 19

load, including the building load, electric consumption by the heat pump, boiler, and
electric chiller, along with the amount of power exported to the utility grid.

Pimin ≤ Pi (t) ≤ Pimax (2)

char max (1 − SOC (t))

0 ≤ PESS (t) ≤ PESS . (3)
ηc
dis max
0 ≤ PESS (t) ≤ PESS .SOC (t).ηd (4)
char dis
SOC (t + 1) = SOC (t) + PESS (t) − PESS (t) (5)

dis ( t ) − Pchar ( t ) + P buy sell

PPV (t) + P1 (t) + P2 (t) + PESS ESS Grid ( t ) − PGrid
(6)
= PL (t) + PHP (t) + PBl (t) + PEC (t)
Similarly, the heat generated by the HP and boiler should balance the heat load and
the heat supplied to the AC, as shown in Equations (7)–(9). The electric-to-heat ratio of the
HP = 2 and η Bl = 3, respectively.
HP and boiler are ηe2h e2h

HHP (t) + HBl (t) = HL (t) + H AC (t) (7)

HP
HHP (t) = ηe2h × PHP (t) (8)
Bl
HBl (t) = ηe2h × PBl (t) (9)
Finally, the cooling load should always be fulfilled by the cooling generated by the AC
and the EC, as expressed in Equations (10)–(12). The heat-to-cooling and electric-to-cooling
AC = 2.5 and η EC = 2.5, respectively.
ratios are selected as ηh2c e2c

C AC (t) + CEC (t) = CL (t) (10)

AC
C AC (t) = ηh2c × H AC (t) (11)
EC
CEC (t) = ηe2c × PEC (t) (12)

3. Numerical Results
In this section, a detailed analysis is provided for dataset correlation calculation, the
prediction model, and the optimal operation of the energy building. The BES, previously
showcased in Figure 1, is utilized for the test system in this study. The prediction and
optimization model will be applied to a medium-sized office building located in Berlin,
Germany. PV generation data obtained for this building is a significant factor in the energy
system and will be incorporated into the BES for testing the models. First, we select features
that exhibit a high correlation with PV output to improve the accuracy of the PV output
prediction model. Next, the predicted output is integrated into the optimization model
to schedule the operation of the entire building energy system, aiming to minimize both
operational costs and the building’s dependency on external energy systems.
Both Tables 2 and 3 present the variables defined for the key components used in
the optimization model. Another critical component in the model is the battery, with a
maximum capacity of 1000, initial charge of 0.2, and a minimum and maximum charging
state of 0.2 and 0.9, respectively. The parameters associated with the battery, distributed
generators, and heating and cooling equipment will be integrated into the optimization
model to determine the operational cost of powering the building.

Table 2. Distributed generator parameters for optimization model.

Components Min (kW/h) Max (kW/h) Cost (USD/kW/h)

DG1 0 400 0.2
DG2 0 500 0.3
 to determine the operational cost of powering the building.

Table 2. Distributed generator parameters for optimization model.

Components Min (kW/h) Max (kW/h) Cost (USD/kW/h)

DG1 0 400 0.2
Algorithms 2024, 17, 528 7 of 19
DG2 0 500 0.3

Table 3. Heating and cooling equipment parameters for optimization model.

Table 3. Heating and

Components cooling
Min (Wh) equipment parameters
Max (Wh) for optimization
Ratio model.
Heat Pump 0 6000 2
Components
Heat Boiler 0 Min (Wh) 6000 Max (Wh)3 Ratio
Cooling AC
Heat Pump 0 0 6000 6000 2 2
Cooling EC 0 6000 3
Heat Boiler 0 6000 3
3.1. Input Data
Cooling AC 0 6000 2
The one-year PV dataset is derived from the output of the building’s PV system,
Cooling EC 0 6000 3
sourced from PVWatts [23]. Observing the correlations between the variables in the da-
taset provides valuable insights into which parameters most significantly impact PV sys-
3.1.output.
tem Input This
Dataunderstanding can further assist in predictive modeling, PV system de-
sign, and the selection of installation sites. As shown in Figure 2, beam irradiance—the
The one-year PV dataset is derived from the output of the building’s PV system,
direct sunlight received by the panels—has the strongest correlation with PV system out-
sourced from
put (r = 0.89). PVWatts
This indicates[23]. Observing
that as the amountthe correlations
of direct betweenthe
sunlight increases, theenergy
variables in the dataset
provides
output of thevaluable insights into which parameters most significantly impact PV system
system rises.
The second
output. This strongest correlationcan
understanding is between
furtherdiffuse
assistirradiance and PV system
in predictive modeling,outputPV system design,
(r = 0.77). Diffuse irradiance refers to sunlight that has been scattered by the atmosphere
and the selection of installation sites. As shown in Figure 2, beam irradiance—the direct
rather than direct sunlight. Despite this distinction, it still contributes positively to system
sunlight
output. Bothreceived
irradianceby the panels—has
measures demonstrate the strongest
a positive correlation
correlation withasPV system output
with output,
(r = 0.89).
shown This3, indicates
in Figure thatPVasenergy
indicating that the amount
productionofincreases
direct sunlight
with total increases,
sunlight ex- the energy output
posure,
of the whether
systemdirect
rises.or indirect.

Figure
Figure2. 2.
Correlation matrix
Correlation illustrating
matrix the relationships
illustrating between beam
the relationships irradiance,
between beamdiffuse irradi- diffuse irradiance,
irradiance,
ance, ambient temperature, wind speed, and PV system output.
ambient temperature, wind speed, and PV system output.

The second strongest correlation is between diffuse irradiance and PV system output
(r = 0.77). Diffuse irradiance refers to sunlight that has been scattered by the atmosphere
rather than direct sunlight. Despite this distinction, it still contributes positively to system
output. Both irradiance measures demonstrate a positive correlation with output, as shown
Algorithms 2024, 17, x FOR PEER REVIEW 8 of 19
in Figure 3, indicating that PV energy production increases with total sunlight exposure,
whether direct or indirect.

(a) (b)
Figure 3. Correlation plots between beam irradiance, diffuse irradiance, and PV system output. (a)
Figure 3. Correlation plots between beam irradiance, diffuse irradiance, and PV system output.
PV output and beam irradiance; (b) PV output and diffuse irradiance.
(a) PV output and beam irradiance; (b) PV output and diffuse irradiance.
In contrast, ambient temperature has a weaker correlation with PV system output,
with an r value of 0.44 (Figure 2), suggesting that while temperature may have some in-
fluence, it is not a key factor in determining system performance. The weakest correlation
is found between wind speed and PV system output, with an r value of 0.058 (Figure 2),
reflecting the minimal impact of wind speed on the system. Figure 4 further illustrates the
 (a) (b)
Algorithms 2024, 17, 528 8 of(a)
Figure 3. Correlation plots between beam irradiance, diffuse irradiance, and PV system output. 19
PV output and beam irradiance; (b) PV output and diffuse irradiance.

In
In contrast,
contrast, ambient
ambient temperature
temperature has has aa weaker
weaker correlation
correlation with
with PV
PV system
system output,
output,
with
with anan rrvalue
valueofof0.44
0.44(Figure
(Figure2),2),suggesting
suggesting that while
that temperature
while temperature may have
may some
have somein-
fluence,
influence,it is not
it is a key
not factor
a key factorinindetermining
determiningsystem
systemperformance.
performance.The Theweakest
weakestcorrelation
correlation
is
is found
found between
between windwind speed
speed andand PVPV system
system output,
output, with
with an
an rr value
value of
of 0.058
0.058 (Figure
(Figure 2),
2),
reflecting
reflecting the minimal impact of wind speed on the system. system. Figure
Figure 44 further
further illustrates
illustrates the
the
lack of
lack of correlation
correlation between wind speed and output, highlighting its insignificant effect.n effect.n

(a) (b)
Figure
Figure 4. Correlation
Correlation plots
plotsbetween
betweenambient
ambienttemperature,
temperature,wind
wind speed,
speed, andand
PVPV system
system output.
output. (a) (a)
PV
PV output
output andand ambient
ambient temperature;
temperature; (b) PV
(b) PV output
output andand wind
wind speed.
speed.

The
The output
output power
powerofofPVsPVsisisobtained
obtainedbybyperforming
performing thethe
PVPV prediction
predictionmodel, as il-
model, as
lustrated
Algorithms 2024, 17, x FOR PEER REVIEW in
illustrated Figure
in Figure5a.
5a.The
Theelectric,
electric,heat, and
heat, andcooling
cooling loads
loads are
are taken
taken from
from a
a combined
combined9 of 19
cooling,
cooling, heat,
heat, and
and power
power system
system inin an
an office
office building
building [24],
[24], as
as depicted
depicted in in Figure
Figure 5b–d.
5b–d.
Similarly, the data for three-day-ahead optimization
Similarly, optimization are
are presented
presented in in Figure
Figure 6a–d.
6a–d.

(a) PV data (b) Electricity data

(c) Cooling data (d) Heating data

Figure 5.
5. Input
Input data
datafor
forone-day
one-daysimulation.
simulation.
Algorithms 2024, 17, 528 (c) Cooling data (d) Heating data 9 of 19
Figure 5. Input data for one-day simulation.

(a) PV data (b) Electricity data

(c) Cooling data (d) Heating data

Figure
Figure 6.
6. Input
Input data
data for
for three-day
three-day simulation.
simulation.

3.2. Prediction
3.2. Prediction Models
Models
The chosen
The chosen prediction
predictionmodel
modelutilizes
utilizesan
anNN
NNarchitecture.
architecture.This
This architecture
architecture refers
refers to
to the arrangement of neurons into layers, along with the connection patterns
the arrangement of neurons into layers, along with the connection patterns between lay- between
layers.
ers. It depicts
It depicts howhow a network
a network transforms
transforms the received
the received input input into output.
into output. The NN The NN
model
model integrated in this project uses Keras, which is a high-level neural network API. The
following characteristics of the model and its layers differ slightly among one-day and
three-day predictions to account for the different levels of complexity. The input layer
of the NN model represents the parameters used to help generate the output layer. The
specific input parameters utilized in the model are ambient temperature, beam irradiance,
diffuse irradiance, wind speed, and date and time. Both one- and three-day prediction
models have densely connected input layers with seven neurons.
The distinguishing factor between the models for one-day versus three-day simulation
is the size of the hidden layer. A hidden layer is an intermediary layer between the input
and the output. This is where the data received by the model are essentially studied for
intricate patterns to produce accurate predictions. The hidden layer size for the one- and
the three-day simulations is 7 and 14 neurons, respectively. There are more data to consider
for three-day prediction, so having 14 neurons in the hidden layer allows the model to
train more efficiently and handle the complexities and size of the data. The output layer for
the one- and the three-day model is 24 and 72 neurons, respectively, which corresponds
to the number of hours for which we want to obtain predictions. The adaptive moment
estimation (Adam) optimizer is utilized during the compilation of each model. This ensures
enhanced performance and the minimization of the model’s loss function.
The results from the comparison in Table 1 show that the NN model produced the
most accurate results. It received the lowest values for each testing error metric compared
with ARIMA and SVR. This high performance can be attributed to the model’s architecture
comprising hidden layers that utilize activation functions, ensuring that complex and
non-linear patterns are captured. The results further highlight the advantages of a machine
learning approach, as three different prediction models were evaluated, and accurate
results were obtained (Table 4). While an analytical approach may provide predictions
Algorithms 2024, 17, 528 10 of 19

through methodical calculations, it was not selected due to the added complexity. Selecting
machine learning as a means to obtain predictions proved to be a more feasible alternative,
as accurate results are obtainable and varying datasets can be used. This tradeoff allowed
for the balance between simplicity and predictive accuracy, making the machine learning
approach the more practical option for the project scope.

Table 4. Testing error comparison across three models.

Testing Error (RMSE)

Models Three-Day Seven-Day
One-Day Prediction
Prediction Prediction
ARIMA 244.09 182.37 194.16
SVR 291.23 200.03 191.97
FNN 63.07 61.37 195.93
Testing Error (MAE)
Models
Algorithms 2024, 17, x FOR PEER REVIEW 1 day 3 days 7 days 11 of 19
ARIMA 143.25 103.64 119.5
SVR 206.59 103.4 107.13
FNN
1 day 43.62 3 days 41.37 113.45
7 days
Testing Error (nRMSE)
ARIMA Models
0.163 0.110 0.066
1 day 3 days 7 days

SVR ARIMA
0.194 0.163 0.120 0.110 0.0660.066
SVR 0.194 0.120 0.066
FNN 0.042 0.037 0.067
FNN 0.042 0.037 0.067

3.2.1.
3.2.1. One-Day
One-Day Prediction
Prediction
To
To further validatethe
further validate themodel’s
model’s training
training results,
results,a training andand
a training validation loss is
validation pre-
loss is
sented with the model loss over epochs. Training loss is an indication
presented with the model loss over epochs. Training loss is an indication of how well a of how well a model
is fittingistraining
model data, ordata,
fitting training how oreffectively it is learning
how effectively patterns.patterns.
it is learning Validation loss displays
Validation loss
how wellhow
displays the model
well thefitsmodel
new, unseen
fits new,data.
unseenAs shown
data. As in shown
Figure in7, both
Figure training
7, bothand vali-
training
dation loss haveloss
and validation an initial
have an decrease. A decreasing
initial decrease. trend fortrend
A decreasing training loss indicates
for training that the
loss indicates
model
that the is model
learningis and showing
learning and improvement in the training
showing improvement in thedata. A decreasing
training validation
data. A decreasing
loss shows loss
validation thatshows
the model
that is
theimproving its predictions
model is improving on unseenon
its predictions data. We can
unseen also
data. Weinfer
can
that the model has an optimal fit because both losses decrease and they
also infer that the model has an optimal fit because both losses decrease and they stabilize atstabilize at a cer-
tain point.
a certain Figure
point. 8 illustrates
Figure thatthat
8 illustrates the frequency
the frequency of error forfor
of error thethe
one-day
one-day prediction
prediction has a
has
a high
high concentration
concentration ofof values
values surroundingzero,
surrounding zero,indicating
indicatinghigh highlevels
levelsofofmodel
modelaccuracy.
accuracy.

Figure 7.
Figure Training and validation loss for one-day prediction.
7. Training
 Algorithms 2024, 17, 528
Figure 7. Training and validation loss for one-day prediction. 11 of 19

Figure 8. Frequency of errors for one-day prediction.

Figure 8. Frequency of errors for one-day prediction.
Figure 9 depicts the actual versus predicted results of PV system output over a span
Algorithms 2024, 17, x FOR PEER REVIEW
of 24 h. The actual results are obtained from PVWatts, while the predicted output comes
Figure
from the NN model.9 depicts
As depictedthe actual
in the versus
graph, the predicted
predictions results
shown are highly of PV sys
accurate,
proving
of 24 h.the The
effectiveness
actual of the model. are obtained from PVWatts, while the
results
from the NN model. As depicted in the graph, the predictions sho
proving the effectiveness of the model.

Figure
Figure 9. Actual vs. predicted
9. Actual output over
vs. predicted 24 h.
output over 24 h.
3.2.2. Three-Day Prediction
3.2.2. Three-Day
Similarly Prediction
to the one-day time frame, the performance of the model was also analyzed
over a three-day period. The significance of performing predictions over a wider horizon is
Similarly
to strengthen to theofone-day
the validity how well the time frame,
model the over
performs performance
different timeof the model
ranges. This was
over athat
proves three-day
the modelperiod. The
can provide significance
accurate offor
predictions performing
different timepredictions
ranges, whichover a
will assist in future operational planning and energy management.
is to strengthen the validity of how well the model performs over differen
In Figure 10, both training and validation loss also have an initial steep decrease. A
This proves
decreasing trend that the model
for training can learning
loss indicates provide andaccurate
improvement predictions for differen
toward the training
which
data. Thewill
modelassist
is alsoin future operational
improving planning
its predictions on and
unseen data for aenergy management.
three-day range, as
evidentInbyFigure
the decreasing validation loss.
10, both training and validation loss also have an initial stee
decreasing trend for training loss indicates learning and improvement towa
data. The model is also improving its predictions on unseen data for a three
evident by the decreasing validation loss.
 This proves that the model can provide accurate predictions for different time ranges,
which will assist in future operational planning and energy management.
In Figure 10, both training and validation loss also have an initial steep decrease. A
decreasing trend for training loss indicates learning and improvement toward the training
Algorithms 2024, 17, 528 data. The model is also improving its predictions on unseen data for a three-day range,
12 ofas
19
evident by the decreasing validation loss.

Figure
Figure 10.
10. Training
Training and validation loss for three-day prediction.

Figure 11
Figure 11 showcases
showcases that
that the
the model
model can
can still
still perform
perform at high levels
levels of accuracy
accuracy over
over
three days because of the high concentration of zeros for the frequency of
three days because of the high concentration of zeros for the frequency of error. The 13 error. The
results
Algorithms 2024,
Algorithms 2024, 17,
17, xx FOR
FOR PEER
PEER REVIEW
REVIEW 13 of
of 19
19
results in
shown shown in Figure
Figure 12 for
12 for the the three-day
three-day predictions
predictions are similar
are similar to those
to those forone-day
for the the one-day
pre-
predictions,
dictions, where
where the the
modelmodel produces
produces highly
highly accurate
accurate results.
results.

Figure 11.
Figure 11.
Figure Frequency
11. Frequency of
Frequency of errors
of errors for
errors for three-day
for three-day prediction.
three-day prediction.
prediction.

Figure 12.
Figure Actual vs.
12. Actual vs. predicted output over
predicted output over 72
72 h.
h.

3.3. Day-Ahead
3.3. Day-Ahead
3.3. Scheduling
Day-Ahead Scheduling
Scheduling
The results
The results
The from
results from the
from the optimization
the optimization model
optimization model show
model show how
show how the
how the building’s
the building’s energy
building’s energy demands
energy demands
demands
for cooling, heating, and electricity are met. As shown in Figure 13, the cooling energy
for cooling,
for cooling, heating,
heating, and
and electricity
electricity are
are met.
met. AsAs shown
shown in in Figure
Figure 13,
13, the
the cooling
cooling energy
energy
balance is always maintained, with the energy output precisely meeting the building’s
balance is
balance is always
always maintained,
maintained, with
with the
the energy
energy output
output precisely
precisely meeting
meeting the the building’s
building’s
cooling requirements. The energy supply is primarily provided by the AC rather than the
cooling requirements.
cooling requirements. The The energy
energy supply
supply isis primarily
primarily provided
provided by by the
the ACAC rather
rather than
than the
the
EC due to the former being more cost-effective.
EC due
EC due to
to the
the former
former being
being more
more cost-effective.
cost-effective.
In Figure
In Figure 14,
14, the
the optimization
optimization model
model ensures
ensures that
that the
the heating
heating supply
supply fromfrom the
the boiler
boiler
meets the
meets the building’s
building’s heat
heat load.
load. The
The heat
heat pump
pump is is not
not utilized,
utilized, as
as the
the model
model adopts
adopts the
the
most cost-effective
most cost-effective approach
approach to to meet
meet the
the building’s
building’s energy
energy needs.
needs. ItIt is
is important
important to to dis-
dis-
tinguish the
tinguish the heating
heating shown
shown inin Figure
Figure 14
14 from
from that
that in
in Figure
Figure 5d.
5d. The
The difference
difference arises
arises be-
be-
cause Figure
cause Figure 1414 is
is not
not representative
representative of
of the
the heat
heat load
load itself
itself but
but rather
rather the
the heat
heat after
after aa por-
por-
rithms 2024, 17, x FOR PEER REVIEW
Algorithms 2024, 17, 528 13 of 19

rithms 2024, 17, x FOR PEER REVIEW

Figure 13. Optimization model cooling output over 24 h.

Figure 13. Optimization model cooling output over 24 h.
In Figure 14, the optimization model ensures that the heating supply from the boiler
meets the building’s heat load. The heat pump is not utilized, as the model adopts the most
cost-effective approach to meet the building’s energy needs. It is important to distinguish
the heating shown in Figure 14 from that in Figure 5d. The difference arises because
Figure 14 is not representative of the heat load itself but rather the heat after a portion of the
AC’s energy has been added. In the optimization model, the total heat load is calculated by
adding the
Figure 13.heating demand (Figure
Optimization 5d) with
model a portion
cooling of the AC
output energy
over 24 output,
h. explaining
why the heating demand prior to optimization appears lower.

Figure 14. Optimization model heating output over 24 h.

Figure 14. Optimization model heating output over 24 h.

Figure 14. Optimization model heating output over 24 h.
As given in Equation (6), the electric power provided by the energy sources should
equal the power demanded by the building. Figure 15 depicts this balance, showing that
the sum of the building’s energy supply at every time step equals the demand, represented
in the graph as the total electric load. To show the effectiveness of the proposed method,
Table 5 presents a comparison with and without the proposed optimization.

Figure 15. Optimization model electricity outputs over 24 h.

 Algorithms 2024, 17, 528 14 of 19
Figure 14. Optimization model heating output over 24 h.

Algorithms 2024, 17, x FOR PEER REVIEW

Figure 15. Optimization model electricity outputs over 24 h.
Figure 15. Optimization model electricity outputs over 24 h.
Table 5. Non-optimized versus optimized costs.
Table 5. Non-optimized versus optimized costs.
Operational Cost Non-Optimized ($) Optimized ($) Cost Reduction (%)
In Figure 17, the optimization model guarantees that the building’s hea
Operational
Power GenerationCost Non-Optimized
1729.71 ($)
1489.71 Optimized 13.88($) Cost R
ments are being met by the boiler’s heating supply. The heat pump is no
Power Generation
Cooling Generation 1778.25 1729.71 1489.71 1489.7116.23
model recognizes that it would be a more costly option. The same expla
Heating Generation
Cooling Generation 1656.14 1778.25 1489.71 10.05
heating load difference applies to the three-day results1489.71
for the optimization m
Heatingdemand
heating Generation
3.4. Three-Day-Ahead for three days
Scheduling
1656.14
(Figure 6d). 1489.71
Finally,
Similarly to the electricresults,
the one-day power the supplied by the energy
three-day optimization sources must
model demonstrates howmatch
3.4.
powerThree-Day-Ahead
the building’s energy Figure
demand. Scheduling
demands18 aredepicts
met. As shown in Figure 16,
this balance, the cooling energy
showcasing thatbalance
the total sup
is maintained, with no surplus/shortage in energy output. Cost-effectiveness is achieved
at every time
Similarly
by sourcing
step is
to the
the cooling
equal
one-day
energy
to the demand.
supplyresults,
from the the
AC. three-day optimization model dem
the building’s energy demands are met. As shown in Figure 16, the cooling e
is maintained, with no surplus/shortage in energy output. Cost-effectivene
by sourcing the cooling energy supply from the AC.

Figure 16. Optimization model of cooling output over 72 h.

Figure 16. Optimization model of cooling output over 72 h.
In Figure 17, the optimization model guarantees that the building’s heat load require-
ments are being met by the boiler’s heating supply. The heat pump is not used, as the
model recognizes that it would be a more costly option. The same explanation for the
heating load difference applies to the three-day results for the optimization model and the
heating demand for three days (Figure 6d).
 Algorithms 2024, 17, 528 15 of 19
Figure 16. Optimization model of cooling output over 72 h.

Figure 17. Optimization model of heating output over 72 h.

Figure 17. Optimization model of heating output over 72 h.
ithms 2024, 17, x FOR PEER REVIEW
Finally, the electric power supplied by the energy sources must match the building’s
power demand. Figure 18 depicts this balance, showcasing that the total supply of energy
3.5. GUI
at every timeDevelopment
step is equal to the demand.

The development of a GUI significantly enhances the user exp

visualization of data and results. Users can interact with the int
input and output of the optimization model through graphical
study, different tabs are designed, as shown in Figure 19, which a
easily through the results of the optimization model for both one-
Additionally, access to a dedicated tab enables users to retrain th
and run the PV predictions for both one-day and three-day mod
basic visual representation of the neural network model along wi
the number of neurons in each layer. To showcase model efficien
training and validation loss is displayed along with a histogram o

Figure 18. Optimization model of electricity output over 72 h.

Figure 18. Optimization model of electricity output over 72 h.
3.5. GUI Development
The development of a GUI significantly enhances the user experience by enabling
the visualization of data and results. Users can interact with the interface to view both
the input and output of the optimization model through graphical representations. In this
study, different tabs are designed, as shown in Figure 19, which allows users to navigate
easily through the results of the optimization model for both one-day and three-day data.
Additionally, access to a dedicated tab enables users to retrain the neural network model
and run the PV predictions for both one-day and three-day models. This tab includes a
basic visual representation of the neural network model along with a summary detailing
the number of neurons in each layer. To showcase model efficiency, a plot depicting the
training and validation loss is displayed along with a histogram of the frequency of error.
Algorithms 2024, 17, 528 16 of 19
Figure 18. Optimization model of electricity output over 72 h.

Figure 19. Developed GUI.19. Developed GUI.

Figure

The optimization The modeloptimization model tabs hold key equations to further facilitate user comprehen-
tabs hold key equations to further facilitate user comprehen-
sion of how the system optimizes building energy use. The visual of the project’s BES is
sion of how the system optimizes
also provided, building
illustrating howenergy
power isuse. Theand
sourced visual of thethroughout
distributed project’sthe
BES is
building.
also provided, illustrating how power
The comprehensive is sourced
design of the GUI and distributed
ensures that users throughout the building.
can interact seamlessly with the
The comprehensive design of and
optimization the prediction
GUI ensures that users
framework can interact
and directly view the seamlessly
results. Thewith the
future devel-
opment of this GUI will allow for the integration of additional
optimization and prediction framework and directly view the results. The future devel- building energy demands,
enabling the adaptation to more complex energy systems. The extended functionality will
opment of this GUI will allow for the integration of additional building energy demands,
provide the visualization of energy solutions tailored to the unique requirements of each
enabling the adaptation
building.to This
more willcomplex energy
aim to further systems.
enhance The extended
the scalability, allowing functionality will
the GUI to accommodate
provide the visualization
different of energy
building solutions
types and energytailored
demands. to the unique requirements of each
Ultimately, the GUI serves
building. This will aim to further enhance the scalability, as a critical integration
allowing point
theforGUI
the prediction
to accommo-and opti-
mization model within the BES. By providing a user-friendly interface, it allows operators
date different building types and energy demands.
to utilize predictive models and algorithms without low-level knowledge and technical
Ultimately, the GUI serves
expertise. as a this
Providing critical integration
accessibility pointasfor
is crucial, the prediction
it encourages and
users to opti-with
engage
mization model within
energythe BES. By providing
management a user-friendly
tools, fostering interface,
a proactive approach to it allows operators
optimizing energy usage
through renewable resources.
to utilize predictive models and algorithms without low-level knowledge and technical
expertise. Providing this accessibility is crucial, as it encourages users to engage with en-
4. Discussion and Future Direction
ergy management tools, fostering a proactive approach to optimizing energy usage
Due to the unpredictable nature of renewable energy generation, this study acknowl-
through renewableedgesresources.
the limitations related to demand price uncertainty. The analysis and framework
provided in this study are focused on short-term evaluations, specifically one- and three-day
4. Discussion andperiods.
FutureThis limited scope may not fully capture the dynamic nature of renewable energy
Direction
generation. Additionally, the study is conducted within the time range of a single season,
Due to the unpredictable nature
capturing fairly of renewable
consistent energy The
weather conditions. generation, this study
project’s future acknowl-
enhancements plan to
edges the limitations related
capture a moretocomprehensive
demand price uncertainty.
analysis The analysis
that spans multiple seasons and framework
and diverse conditions,
providing a deeper understanding of potential challenges associated with the utilization of
renewable energy systems in buildings.
Moving forward, we aim to combat the limitations by enhancing the framework to
include long-term operation analyses and integrate datasets that hold diverse weather
conditions. Future expansions could focus on the real-time operation of NZEBs using IoT
Algorithms 2024, 17, 528 17 of 19

devices and a data-driven control approach. Furthermore, incorporating multi-objective op-

timization that simultaneously addresses cost minimization and emission reduction could
significantly enhance the economic and environmental sustainability of NZEB operations.

5. Conclusions
In this study, we developed a comprehensive optimization framework for the optimal
operation of NZEBs, consisting of a prediction model and an optimization model. The
framework’s focus was placed on forming predictions based on a provided dataset, which
highlights its adaptability to form predictions given new buildings and energy input
variables. Initially, a neural network-based prediction model was trained with PV data
to provide a highly accurate estimated output power of PV systems, which was then
fed into the optimization model. This model determines the optimal set points for each
component in the network and can accommodate various operational horizons, from one to
three days. The framework not only minimizes operational costs but also reduces NZEBs’
dependence on external grids. Additionally, we developed a detailed GUI to visualize
optimization outputs and monitor the system’s operational status, helping operators to
track changes easily, particularly in emergency operation modes. The GUI serves as a
vital integration point, making sophisticated prediction and optimization tools accessible
for building operators, thereby encouraging the use of energy management systems with
renewable resources.

Author Contributions: Conceptualization, V.-H.B.; methodology, R.K. and V.-H.B.; software, R.K.,
G.V.H. and A.H.; validation, V.-H.B. and W.S.; formal analysis, R.K., G.V.H., and A.H.; investigation,
R.K. and V.-H.B.; resources, G.V.H. and A.H.; data curation, R.K. and W.S.; writing—original draft
preparation, R.K. and V.-H.B.; writing—review and editing, R.K., G.V.H., A.H. and W.S.; visualization,
R.K., G.V.H. and A.H.; supervision, V.-H.B. and W.S.; project administration, V.-H.B.; funding
acquisition, V.-H.B. All authors have read and agreed to the published version of the manuscript.
Funding: The authors’ work was supported by the University of Michigan-Dearborn’s Summer
Undergraduate Research Experience (SURE) program.
Data Availability Statement: Data are provided upon request.
Conflicts of Interest: The authors declare no conflicts of interest.

Nomenclature
AC Absorption chiller
EC Electrical chiller
ESS Energy storage system
GUI Graphical user interface
HP Electric heat pump
NZEB Net-zero energy buildings
PV photovoltaic
t Time interval
T Operation horizon
PV1 , PV2 , . . . , PVn Photovoltaic sites
DG1 , DG2 , . . . , DGn Distributed generators
PPV (t) Output of PV system at t
P1 (t), P2 (t) Output of DGs 1 and 2 at t
Pimin , Pimax Minimum and maximum output of DG
PESS (t) Power exchanged with system and ESS
α, β The weight of sub-objectives
buy sell ( t )
PGrid (t), PGrid Buying and selling amount with utility grid at t
max
PESS Capacity of ESS
dis ( t ), Pchar ( t )
PESS Charging and discharging amount of ESS at t
ESS
SOC (t) State of charge of ESS at t
ηd , ηc Discharging and charging losses of ESS
PL (t) Electric load amount at t
Algorithms 2024, 17, 528 18 of 19

PHP (t) Power consumption by electric heat pump at t

PBl (t), PEC (t) Power consumption by boiler and electric chiller at t
HHP (t) Heat energy output of heat pump at t
HBl (t) Heat energy output of boiler at t
HL (t) Heat load at t
H AC (t) Heat consumption by absorption chiller at t
C AC (t) Cooling output of absorption chiller at t
CEC (t) Cooling output of electric chiller at t
CL (t) Cooling load at t
C1 , C2 Operation cost of DGs 1 and 2 at t
pGrid (t) Trading price with the utility grid at t

References
1. Harish, V.; Kumar, A. A review on modeling and simulation of building energy systems. Renew. Sustain. Energy Rev. 2016, 56,
1272–1292. [CrossRef]
2. Mariano-Hernández, D.; Hernández-Callejo, L.; Zorita-Lamadrid, A.; Duque-Pérez, O.; García, F.S. A review of strategies for
building energy management systems: Model predictive control, demand side management, optimization, and fault detection &
diagnosis. J. Build. Eng. 2021, 33, 101692.
3. Ahmed, A.; Ge, T.; Peng, J.; Yan, W.C.; Tee, B.T.; You, S. Assessment of the renewable energy generation towards net-zero energy
buildings: A review. Energy Build. 2022, 256, 111755. [CrossRef]
4. Wells, L.; Rismanchi, B.; Aye, L. A review of Net Zero Energy Buildings with reflections on the Australian context. Energy Build.
2018, 158, 616–628. [CrossRef]
5. Sartori, I.; Napolitano, A.; Marszal, A.; Pless, S.; Torcellini, P.; Voss, K. Criteria for definition of net zero energy buildings.
In Proceedings of the International Conference on Solar Heating, Cooling and Buildings (EuroSun 2010), Graz, Austria, 28
September–1 October 2010.
6. Wu, W.; Skye, H.M. Residential net-zero energy buildings: Review and perspective. Renew. Sustain. Energy Rev. 2021, 142, 110859.
[CrossRef]
7. Panagiotidou, M.; Aye, L.; Rismanchi, B. Optimisation of multi-residential building retrofit, cost-optimal and net-zero emission
targets. Energy Build. 2021, 252, 111385. [CrossRef]
8. Liu, J.; Zhou, Y.; Yang, H.; Wu, H. Net-zero energy management and optimization of commercial building sectors with hybrid
renewable energy systems integrated with energy storage of pumped hydro and hydrogen taxis. Appl. Energy 2022, 321, 119312.
[CrossRef]
9. Ohene, E.; Chan, A.P.; Darko, A. Review of global research advances towards net-zero emissions buildings. Energy Build. 2022,
266, 11214. [CrossRef]
10. Bataille, C.G.F. Physical and policy pathways to net-zero emissions industry. WIREs Clim. Chang. 2020, 11, e633. [CrossRef]
11. Miller, S.A.; Habert, G.; Myers, R.J.; Harvey, J.T. Achieving net zero greenhouse gas emissions in the cement industry via value
chain mitigation strategies. One Earth 2021, 4, 1398–1411. [CrossRef]
12. Heendeniya, C.B.; Sumper, A.; Eicker, U. The multi-energy system co-planning of nearly zero-energy districts—Status-quo and
future research potential. Appl. Energy 2020, 267, 114953. [CrossRef]
13. Ruparathna, R.; Hewage, K.; Sadiq, R. Rethinking investment planning and optimizing net zero emission buildings. Clean Technol.
Environ. Policy 2017, 19, 1711–1724. [CrossRef]
14. Richter, M.; Lombardi, P.; Arendarski, B.; Naumann, A.; Hoepfner, A.; Komarnicki, P.; Pantaleo, A. A vision for energy de-
carbonization: Planning sustainable tertiary sites as net-zero energy systems. Energies 2021, 14, 5577. [CrossRef]
15. Zhou, Z.; Feng, L.; Zhang, S.; Wang, C.; Chen, G.; Du, T.; Li, Y.; Zuo, J. The operational performance of ’net zero energy building’:
A study in China. Appl. Energy 2016, 177, 716–728. [CrossRef]
16. Klein, K.; Kalz, D.; Herkel, S. Grid impact of a net zero energy building with BiPV using different energy management strategies.
In Proceedings of the International Conference CISBAT 2015 Future Buildings and Districts Sustainability from Nano to Urban
Scale, LESO-PB, EPFL, Lausanne, Switzerland, 9–11 September 2015; pp. 579–584.
17. Deng, S.; Wang, R.; Dai, Y. How to evaluate performance of net zero energy building—A literature research. Energy 2014, 71, 1–16.
[CrossRef]
18. Sun, Y.; Huang, P.; Huang, G. A multi-criteria system design optimization for net zero energy buildings under uncertainties.
Energy Build. 2015, 97, 196–204. [CrossRef]
19. Zhou, Y.; Zhou, N.; Gong, L.; Jiang, M. Prediction of photovoltaic power output based on similar day analysis, genetic al-gorithm
and extreme learning machine. Energy 2020, 204, 117894. [CrossRef]
20. Kabilan, R.; Chandran, V.; Yogapriya, J.; Karthick, A.; Gandhi, P.P.; Mohanavel, V.; Rahim, R.; Manoharan, S. Short-Term power
prediction of building integrated photovoltaic (BIPV) system based on machine learning algorithms. Int. J. Photoenergy 2021, 2021,
5582418. [CrossRef]
Algorithms 2024, 17, 528 19 of 19

21. Pan, Y.; Zhu, M.; Lv, Y.; Yang, Y.; Liang, Y.; Yin, R.; Yang, Y.; Jia, X.; Wang, X.; Zeng, F.; et al. Building energy simulation and its
application for building performance optimization: A review of methods, tools, and case studies. Adv. Appl. Energy 2023, 10,
100135. [CrossRef]
22. Wang, M.; Yu, H.; Jing, R.; Liu, H.; Chen, P.; Li, C. Combined multi-objective optimization and robustness analysis framework for
building integrated energy system under uncertainty. Energy Convers. Manag. 2020, 208, 112589. [CrossRef]
23. PVWatts Calculator. Available online: https://pvwatts.nrel.gov/index.php (accessed on 15 October 2024).
24. Bui, V.H.; Hussain, A.; Im, Y.H.; Kim, H.M. An internal trading strategy for optimal energy management of combined cooling,
heat and power in building microgrids. Appl. Energy 2019, 239, 536–548. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.