https://doi.org/10.31449/inf.v48i12.6211 Informatica 48 (2024) 153 –170 153 
 
 
Application of Intelligent Reverse Regulation in Uniform Heating of 
Buildings 
 
Xiaoli Wu 
Department of Architectural Engineering, Shijiazhuang College of Applied Technology, Shijiazhuang 050800, China 
E-mail: wuxiaoli_0319@163.com 
Keywords: intelligent regulation, commutation time ratio, buildings, uniform heating, proportional-integral method 
Received: May 17, 2024 
The increase in heating building area has led to an upward trend in heating energy consumption. 
Using automated intelligent equipment to control the heating system has become an important 
development content to solve the building thermal imbalance and maintain thermal balance. Therefore, 
the study proposes to achieve temperature regulation through intelligent reverse in uniform heating of 
buildings, which calculates the energy balance coefficient based on the heat exchange process and the 
physical model of the heating system. In the control strategy, the focus is to discuss the effects of 
heating and cooling on time regulation. Subsequently, a heating automatic control system is designed 
to achieve deviation control and constraint cost function solving using proportional integral control 
and particle swarm optimization algorithm. Performance testing is conducted on the proposed 
adjustment method. The parameter controller reduced indoor temperature fluctuations by adjusting the 
flow rate. Its predicted temperature difference did not exceed 4%, and the flow consumption was 
relatively low. The study selected a high-rise building in the north of China as the experimental object, 
and installed heating systems and intelligent temperature and humidity sensors on it. The automatic 
control results showed that the intelligent directional adjustment mode reduced the heating volume of 
low floors. The average heating volume of low and high floors was 57.14% and 73.68% of the original 
level. The indoor temperature adjustment of vertical floors was 2℃-3℃ lower than that of manual 
adjustment mode. The intelligent regulation method proposed in the study can provide good reference 
value and significance for the energy-saving and temperature control design of heating systems in 
northern regions. 
Povzetek: Inteligentna povratna regulacija za enakomerno ogrevanje stavb zmanjša nihanja 
temperature in porabo energije. Model dosega do 4 % boljše rezultate, kar predstavlja prispevek k 
energetsko učinkovitemu ogrevanju.
1 Introduction 
With the continuous increase of construction area, the 
demand for building heating in northern cities and towns 
continues to rise. The popularity of building heating 
equipment ensures the living comfort of northern 
residents. However, existing research shows that when 
using traditional building heating equipment, there are 
problems such as uneven heating temperature and air 
humidity inside the building, so that the energy loss 
inside the building accounts for one-fifth of the overall 
loss. These include hot water loss and leakage, heat loss, 
and high heat imbalance loss in the horizontal and 
vertical directions of the building. The uneven heating 
inside the building includes the uneven internal flow of 
the heating pipeline and the uneven temperature of the 
heat source inside the pipeline. Influenced by many 
factors such as outdoor temperature and heating demand 
on building heating system, the application method of 
solving thermal imbalance in a single aspect has little 
effect. Therefore, it is necessary to start with the causes 
of temperature imbalance in the heating system and the 
problems that lead to excessive energy consumption. The 
design of building heating systems usually relies on 
predictive models to achieve their temperature control 
effects, but the lag of indoor and outdoor temperature 
differences and equipment delays are largely ignored. 
Therefore, it is difficult to achieve uniform heating in 
buildings under complex climate change [1]. Therefore, 
on the basis of this research, it is proposed to use 
intelligent reverse regulation for uniform building heating. 
In this study, different connection modes of heating 
system are used to explore the causes that affect 
temperature regulation, and to improve them, so as to 
better improve the application and effectiveness of 
heating system. The aim of the research is to reduce 
heating energy cost and waste on the basis of certain 
thermal comfort. The research mainly analyzes the 
uniform heating in buildings from four aspects. The first 
part is the review and discussion of the relevant literature 
on the current intelligent building heating and heating 
energy consumption. In the second part, on the basis of 
analyzing the differences of different heating system 

154   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
design forms, the intelligent commutation regulation 
method and the heating energy consumption prediction 
under the intelligent algorithm are proposed. The third 
part is the performance test and application analysis of 
the building heating condition under the heating 
regulation method. The last part is the summary of the 
full text. 
2 Related work 
The complexity of current building energy system makes 
it necessary to improve efficiency in strategy selection. 
Artificial intelligence algorithm shows good applicability 
in this field. Scholar Song J proposed an improved long 
and short-term memory prediction network based on 
space-time mixing to analyze heating systems. This 
hybrid model could well describe the heating load, and 
the absolute percentage error evaluation index was 
between 3%-4%, which had good thermal adaptability 
and application [2]. Szul, Tabor, and Pancerz (2021) used 
feature selection algorithm and rough set theory to design 
conditional variables, which had high prediction accuracy 
[3]. Psimopoulos et al. (2019) designed a compact 
heating system for an independent house to improve the 
wind speed limit of the original heating. This algorithm 
reduced the energy cost by more than 5% [4]. Aliferov et 
al. (2020) designed a sensor that could uniformly heat the 
end of the steel pipe, and used a Genetic Algorithm (GA) 
to design an induction heating system [5]. The control 
accuracy of flow control in common valve-controlled 
heating systems was low. Zhang, Lei, and Zhang (2019) 
proposed a reference predictive fuzzy adaptive control 
method [6]. This method could better realize the room 
temperature regulator. Sarbu, Mirza, and Crasmareanu 
(2019) conducted numerical simulation model analysis on 
the heating system of heat distribution network aiming at 
the heating problem [7]. Faced with the rapid 
development of home energy management, Devia, 
Agbossou, and Cardenas (2020) proposed a distributed 
co-evolutionary optimization algorithm for heating 
system design [8]. This system could effectively reduce 
cost by more than 20%. The energy consumption of 
residential space heating was an important part of future 
residential research. The differences in heating systems 
enabled technological innovation, strategic planning, and 
other means to achieve energy goals [9]. 
There is a close relationship between the design of 
building heating and cooling system and energy 
consumption, and strengthening the server control ability 
plays an important role in intelligent buildings. The 
artificial intelligence algorithm has a significant 
energy-saving effect (20%-40%) [10]. 
Esrafilian-Najafabadi and Haghighat (2021) proposed a 
multi-layer perceptron network for residential building 
control decision analysis, and added TOPSIS 
multi-criteria decision method to evaluate temperature 
sensitivity [11]. This prediction method could better 
ensure thermal adaptability. With the help of Gaussian 
mixture model clustering algorithm, Lu et al. (2019) 
better identified heating load patterns based on time and 
energy characteristics [12]. Duan et al. (2021) proposed 
an encoder-decoder model under the attention mechanism 
to realize temperature detection, thereby achieving 
long-term spatial factor state information [13]. This 
model had good temperature detection effects. Accurate 
heating load prediction plays an important role in the 
research of intelligent district heating. On the basis of 
ignoring indoor temperature in traditional prediction 
methods, Xue et al. (2020) used long and short term 
memory network to predict heating load, and built a 
nonlinear prediction model covering multiple factors [14]. 
The results showed that this method had a prediction 
accuracy of more than 95%. Based on the demand 
response problem of heat-supply connected buildings, 
Salo et al. (2019) proposed an optimal demand response 
control strategy under case algorithm simulation [15]. 
The results showed that this method could better achieve 
peak load balance and save cost. To solve the response 
time lag of traditional radiant floor heating systems, Chen 
and Li (2020) used Gaussian process regression algorithm 
for predictive model control [16]. The study conducted 
parameter simulation analysis based on various scenarios 
and configuration conditions. The results showed that this 
strategy could effectively reduce the response time and 
ensure better indoor temperature comfort. Vand et al. 
(2019) proposed a demand response adjustment algorithm 
under real-time pricing to achieve temperature adjustable 
settings [17]. Experiments were conducted with the help 
of dynamic building simulation tools. This method could 
effectively reduce heating energy consumption and cost. 
A summary of previous research on building energy 
consumption and heating related work is presented in 
Table 1. 
 
 
Table 1: Summary of building energy consumption and heating related work 
Scholar Methodology Key results Limitations 
Song et al. [2] 
Analyzed heating systems using an 
improved spatiotemporal mixed long 
short-term memory prediction 
network 
The absolute error 
evaluation index of heat 
load is between 3% and 
4% 
The prediction network 
may have generalization 
issues due to the 
influence of training data 
Szul, Tabor, and 
Pancerz [3] 
Feature selection algorithm and rough 
set theory for designing 
energy-saving variable conditions in 
High accuracy in 
predicting building energy 
consumption after thermal 
Affected by changes in 
conditional variables 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 155 
 
buildings renovation 
Psimopoulos et 
al. [4] 
Designed a compact heating system 
Reduce user energy costs 
by over 5% 
Restricted algorithm 
applicability 
 
Aliferov et al. 
[5] 
 
Designed sensors for uniformly 
heating the end of steel pipes and 
solved the heating system using 
genetic algorithm 
 
The control accuracy of 
flow control in ordinary 
valve-controlled heating 
systems is relatively low 
 
Resource and time 
consumption issues, and 
uniform heating design 
cannot adapt to extreme 
conditions 
Zhang, Lei, and 
Zhang [6] 
Design of flow control for 
valve-controlled heating system using 
fuzzy adaptive control method 
Significant room 
temperature regulation 
effect 
Unable to cope with the 
randomness of traffic 
fluctuations 
Sarbu, Mirza, 
and 
Crasmareanu 
[7] 
Application of numerical simulation 
models and deterministic heuristic 
optimization techniques to regional 
heating system distribution networks 
This method can better 
consider the influencing 
factors of system model 
objectives 
Affected by various 
operating conditions 
Devia, 
Agbossou, and 
Cardenas [8] 
Numerical simulation of heating 
network distribution in heating 
systems 
This system can 
effectively reduce costs 
by more than 20% 
Affected by various 
operating conditions 
Ioakimidis [10] 
Introduced artificial intelligence 
algorithms into intelligent buildings 
/ / 
Esrafilian-Najaf
abadi and 
Haghighat [11] 
Analysis of residential buildings 
using multi-layer perceptron networks 
and TOPSIS multi-criteria decision 
methods 
This method can achieve 
good thermal adaptability 
Unable to consider all 
possible diversity and 
complexity in 
decision-making 
problems 
Lu et al. [12] 
Gaussian mixture model clustering 
algorithm for identifying heat load 
patterns 
High precision pattern 
recognition based on time 
and energy characteristics 
The accuracy of load 
pattern recognition is 
influenced by the quality 
and representativeness 
of input data 
Duan et al. [13] 
Combined attention mechanism and 
long short-term memory network to 
detect temperature models 
Good temperature 
detection performance 
The parameter 
sensitivity characteristics 
of attention mechanism 
are more pronounced 
Xue et al. [14] 
Established nonlinear prediction 
model for building heat load using 
long short term memory network 
Prediction accuracy 
exceeds 95% 
Poor performance in 
handling nonlinear 
problems 
Salo et al. [15] 
Optimal demand response control 
strategy for heating connection 
buildings in case simulation 
This method achieves 
better peak load balancing 
and saves costs 
Over-fitting risk 
Chen and Li 
[16] 
Predictive control of traditional floor 
radiant heating systems using 
Gaussian process regression 
algorithm 
This method can 
effectively shorten 
response time and ensure 
better indoor temperature 
comfort 
The case algorithms are 
difficult to ensure 
comprehensiveness 
Vand et al. [17] 
Demand response adjustment 
algorithm under real-time pricing 
This method can 
effectively reduce heating 
energy consumption and 
costs 
May ignore 
environmental changes 
and conditions 
 
 
 
In the past, scholars mostly used numerical 
simulation methods, fuzzy control methods, intelligent 
algorithms, etc. to study the heating flow rate, load mode, 
and energy consumption of intelligent heating buildings. 
Although they have shown good application effects, they 
ignore the time delay characteristics of internal heating in 
buildings, and the automatic temperature control effect is 
poor. Most of the design ideas for some heating systems 

156   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
are limited to specific types and have limited applicability 
for promotion. Therefore, this study proposes a uniform 
heating system with intelligent directional regulation. It is 
designed from the heating system parameter 
identification and prediction model design. This research 
method can better consider the differences in heating 
characteristics of buildings themselves, and improve the 
automation and intelligence of building heating. 
 
3 Design of uniform building heating 
system under intelligent reverse 
regulation 
The uneven heat distribution in different heating systems 
is more prominent, and there is a certain imbalance in 
building heating heat. Therefore, by analyzing different 
heating systems, the analysis and identification of energy 
balance parameters and switching time are proposed. The 
control method is combined with intelligent methods to 
achieve predictive analysis of building heating energy 
consumption [18-19]. 
3.1 Uniform temperature design of buildings 
under intelligent reverse regulation 
The uniform temperature heating system of the building 
is to transmit the low-temperature heat carrier to the heat 
source after being heated, and then to the user through the 
heating pipe. At the same time, the heat carrier that has 
been dissipated is transported back to the circulation 
system at the heat source. Hot water is generally used as 
heat carrier in civil buildings. Steam is generally used as 
a heat carrier in industrial buildings. Generally, there are 
generally three ways to achieve uniform temperature 
heating: parallel, series, and series-parallel coupling. The 
parallel building heating system can better solve the 
uneven building heating, and its specific transmission 
principle is shown in Figure 1. 
Regulating valve
Water supply
Backwater
 
Figure 1: Schematic diagram of parallel system for uniform heating 
 
In Figure 1, hot water from the heat source is 
distributed to different tributaries, and the flow is 
controlled by valves to ensure uniform heat transfer. In 
the upward transmission process, the hot water flow of 
each layer is controlled to ensure that the upper and lower 
temperatures are uniform. Finally, the cooled water is 
returned to the heat source for recycling. In the series 
heating system, the horizontal single-pipe system ensures 
a relatively constant temperature between users on the 
same floor by adjusting the water supply and return time 
for each floor (Figure 2). The vertical single-pipe system 
mainly solves the uneven heat dissipation at the upper 
and lower heat dissipation ends by changing the direction 
of water supply and return (Figure 2(b)). 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 157 
 
76 ℃ 60 ℃
76 ℃
60 ℃
76 ℃
60 ℃
76 ℃
60 ℃
(a) Horizontal series system for 
uniform heating
78 ℃
76 ℃
62 ℃
60 ℃
62 ℃
76 ℃
(b) Vertical series system for 
uniform heating
 
Figure 2: Schematic diagram of a uniform heating series system 
 
For indoor temperature control, temperature 
adjustment can be achieved better by installing reverse 
device to make the water supply and return directions 
opposite. During the heating system cycle, the heat 
g
Q
 
obtained by the heat dissipation terminal equipment 
transmitted to each household is shown in Equation (1). 
12
()
g
Q CG T T =−
        (1) 
In Equation (1), 

is the water density. C is the 
specific heat capacity of water. G is the volume flow of 
water. 
1
T is the water temperature from the heat source. 
2
T is the cooled water temperature. Since the heat loss in 
the transport process is negligible, the approximate 
indoor temperature can be calculated according to the 
energy conservation law. On this basis, according to the 
calculation results, the heat supply is adjusted to ensure 
that the indoor temperature of the building is relatively 
constant. Assuming that the indoor temperature is 
n
T , the 
indoor heat 
n
Q is shown in Equation (2). 
1 2 1 2
( 2 ) ( )
n n g
Q B T T T CG T T Q  = + − = − =
(2) 
In Equation (2), B is the energy balance 
coefficient. When transferred energy is not balanced, the 
corresponding energy of the room temperature 
N
T is 
N
Q . If 
Ng
QQ 
, the room temperature is higher than the 
predetermined value. If 
Ng
QQ 
, it means that the room 
temperature has not reached the predetermined value. The 
system regulates the flow rate by setting the energy 
balance coefficient B to ensure a constant indoor 
temperature. In the parallel heat exchanger uniform 
heating system, it is necessary to determine the energy 
balance coefficient B . When building the physical 
model, considering the diversity of internal floors and the 
thermal conductivity of different materials, various 
materials will be set up to verify the stability of the 
equilibrium coefficient. The specific physical model is 
shown in Figure 3. 
Ground floor
Mortar leveling layer
Fine aggregate concrete
Insulation layer
Floor level
Elastic insulation material
Plastic fixing clip
PEX pipe
Tinfoil
20 50 30
Expansion zone
 
Figure 3: Physical model of series heating system 

158   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
 
The surface temperature of the floor layer in the 
Figure is difficult to be directly related to the indoor 
temperature because of the coverage of the heat 
dissipation tube. The whole process can be completed 
indirectly through the convection heat transfer inside the 
heat transmission pipe, the heat conduction between the 
pipe wall and the adjacent floor layer, and the radiation 
heat transfer of other objects in the room. Therefore, there 
is a certain heat transfer delay. Based on the heat transfer 
process, a mathematical model is constructed to calculate 
the total thermal resistance of the heat source through the 
floor. Equation (3) displays the final energy balance 
coefficient B. 
1
2
B
R
=             (3) 
 
In Equation (3), R represents the total thermal 
resistance. It includes the thermal resistance 
1
R of 
convective heat transfer in the tube. The specific equation 
is shown in Equation (4). 
1
1
pp
R
hA
=
            (4) 
In Equation (4), 
p
h
 is the convective heat transfer 
coefficient between the heat source and the transfer 
pipeline. 
p
A
 denotes the internal surface area of the 
pipeline. Equation (5) is obtained according to Gnielinski 
equation and Filonenko equation. 
 
1
1 2/3
2
1
8 35.92(1.82lg Re 1.64) (Pr 1)
(1.82lg Re 1.64) (Re 1000) Pr
ff
p
f f f
R
A

−
−
=
+ − −
−−
(5) 
In Equation (5), Re is R ey n o ld ’ s coefficient, 
which is the criterion for judging the fluid flow state. f 
is the Darcy resistance coefficient of turbulent flow in the 
pipe. 
Pr
f
 is the damping factor of the liquid at the 
average temperature in the inlet and outlet of the pipeline. 
1
 denotes thermal conductivity corresponding to 
different materials of the floor layer. The total thermal 
resistance also includes the thermal resistance of tube 
wall heat conduction 
2
R , as shown in Equation (6). 
2
2
p
p
R
A


=
             (6) 
In Equation (6), 
p

 is the wall thickness of the 
tube. 
2
A is the surface area of the tube wall. 
p

 is the 
thermal conductivity of the tube wall. In addition, the 
total thermal resistance also includes the thermal 
conductivity resistance 
3
R of the floor layer and the 
comprehensive thermal resistance 
4
R of the floor heat 
dissipation surface, as shown in Equation (7). 
3
3
i
i
R
A


=
            (7) 
In Equation (7), 
i
 is the thickness corresponding 
to different floor materials. 
3
A is the surface area of the 
floor. The comprehensive thermal resistance 
4
R of the 
floor heat dissipation surface is calculated in Equation 
(8). 
4
3
1
c
R
hA
=
           (8) 
In Equation (8), 
c
h is the heat transfer coefficient 
of the ground surface inside the building. Due to radiation 
differences in building materials and shapes, temperature 
distribution is uneven. Therefore, according to previous 
studies, the mathematical model of Equation (9) is 
proposed to calculate the radiation heat transfer 
f
Q
 
inside the building. 
44
,
273 273
4.98
100 100
pj s f
f
TT
Q

++    
 =−
   

   

 (9) 
In Equation (9), 
pj
T
is the average temperature of 
the heat dissipation surface. 
, sf
T
 is the average 
temperature of a non-radiating surface. When the indoor 
temperature is constant, it can be seen from Equation (13) 
that the radiant heat transfer is related to the floor surface 
temperature. In the actual heating process, the average 
temperature of the floor surface layer is generally about 
25°C within the average supply and return water 
temperature. Therefore, the comprehensive heat transfer 
coefficient can be designed, and its mathematical 
expression is shown in Equation (10). 
c
c
tn
q
h
TT
=
−
           (10) 
In Equation (10), n T represents the room 
temperature. 
c q
 is the comprehensive heat exchange. 
f T is the average surface temperature. The thermal 
conductivity resistance of the northern low-temperature 
floor radiant heating system is mainly composed of 
3
R 
and 
4
R . Compared with the former, the values of 
1
R 
and 
2
R differ by more than two orders of magnitude. 
Therefore, the energy balance coefficient B is shown in 
Equation (11). 
34
1
2( )
B
RR
=
+
         (11) 
The energy balance coefficient B per square meter 
inside the building will be derived from 
m
B . Then, the 
relationship between the flow in the heating pipe and the 
energy balance coefficient B is determined by changing 
the flow value in the heating pipe. The experimental 
design is carried out under different working conditions. 
 
3.2 Predictive control analysis of building 
heating energy consumption under intelligent 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 159 
 
reverse regulation 
Considering the delay of the indoor heating system in 
buildings, the time adjustment of heating and cooling 
variables should be explored when the control strategy is 
designed. In the series heating system, the heat transfer 
model of parallel and vertical single pipe systems is 
shown in Equation (12). 
Kt

 =         (12) 
In Equation (12), K is the heat transmitted by the 
radiator. 

 and 

 are radiator coefficients. t  is the 
temperature obtained by subtracting room temperature 
from the average supply and return water temperatures. 
The water supply temperature is 
s
t , the return water 
temperature is 
r
t , and the heat allocated to each 
household is 
1
Q , 
2
Q ,..., 
1 N
Q
−
, 
N
Q . The water supply 
and return temperature of each household in the 
horizontal and vertical directions can be expressed, as 
shown in Figure 4. 
t
s
t
N
t
N-1 t
2
t
1
(t
r
)
Q
N
Q
N-1
Q
2
Q
1
(a) Horizontal system water 
temperature calculation
t
s
t
N
t
N-1
t
2
t
1
(t
r
)
Q
N
Q
N-1
Q
2
Q
1
(b) Vertical system water 
temperature calculation
 
Figure 4: Schematic diagram of heating system with series heat exchanger 
 
In Figure 4, if the heat loss in the pipeline 
transmission process is ignored, the total heat load of a 
vertical single pipe can be expressed as Equation (13). 
1 2 1
1
N
i N N
i
Q Q Q Q Q
−
=
= + + + +

   (13) 
In Equation (13), 
i
Q represents the heat load of the 
i layer radiator. According to the balance relationship 
between indoor and outdoor heat in the heating system, 
the outlet water temperature 
j
t
 of each radiator can be 
obtained, as shown in Equation (14). 
1
()
N
i
ij
j s s r N
i
i
Q
t t t t
Q
=
=
= − −


      (14) 
In Equation (14), 
s
t is the water supply 
temperature. 
r
t is the return water temperature. In the 
heating process, the heat from the source end of hot water 
is continuously lost during the supply process to users, 
which leads to uneven heating at both ends of the 
building. In the cascade system, the heat imbalance is 
alleviated by changing the flow direction, which often 
ignores the heat loss caused by the mixed water in the 
pipe and the thermal inertia of the building. Therefore, by 
changing the ratio of forward time and reverse time, that 
is, the commutation time ratio Ra , the corresponding 
forward and reverse heating cycle is adjusted to ensure 
that the whole building achieves a relative temperature 
constant state. In the vertical series single-pipe heating 
system, the commutation time ratio of each floor is 
calculated, as shown in Equation (15). 
i
i
i
a
Ra
b
=
           (15) 
In Equation (15), i represents layer i . 
i
a 
represents the thermal cycle time of each layer in the 
up-supply and down-cycle mode. 
i
b denotes the heat 
cycle time of each layer in the down-feed up-return mode. 
The commutation time ratio is related to the commutation 
time and the misalignment of the system. Temperature, 
heat dissipation area and flow rate all affect this variable. 
During the experiment, it should be taken into account 
that the heat and load conditions located in the middle 
floor under the two heating systems are similar, so the 
value should be excluded to reduce the experimental error. 
In the design scheme, the research combines the water 
system and the electric system to form a floor radiant 
heating system with air source heat pump inside the 
building. The internal structure is shown in Figure 5. 
Among them, the water system includes air source heat 
pump, flow meter, thermometer, etc. The electric system 
is composed of electric regulating valve and air source 
heat pump unit. 

160   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
Data acquisition
Throttle mechanism
Air 
source 
heat 
pump Evaporator
Condenser
Compressor
Y-shaped filter
Pump
Flowmeter
Electric control valve
Thermometer
220V
24V
Thermometer
T
2
T
1
Indoor heat 
exchange 
pipe loop
 
Figure 5: Air source heat pump floor radiation heating system 
 
In Figure 5, the heating system can obtain low 
temperature heat source by air energy storage, and form 
high temperature heat source through system heat 
collection and integration, thus providing heating. The 
heat balance is maintained by adjusting coefficient and 
commutation time ratio in series and parallel to achieve 
uniform heating effect. In addition, this study also 
designs a heating automatic control system to collect 
information and complete valve control under flow 
regulation. The adjusting hardware system includes 
temperature sensor, single chip microcomputer and 
connection system. In the part of connection system 
design, the system control of energy balance method is 
studied. The traditional Proportional Integral-Differential 
(PID) controller will inevitably lead to deviation signal 
after factor stabilization and system adjustment. 
Therefore, the research uses Proportional Integral (PI) 
control to design the system. The PI control module 
adjusts the valve position to achieve the output value. In 
addition, it maintains the circulation flow in the system to 
the required flow value. No matter how the external 
pipeline network changes, increasing or decreasing flow, 
PI control only requires fewer parameter variables to 
effectively avoid oscillation situations. When the input 
deviation step occurs, the proportional part of PI control 
acts in a timely manner, first performing coarse tuning to 
suppress the influence of disturbances. Then, the 
integration adjustment effect gradually accumulates and 
enhances, and fine adjustments are made to gradually 
eliminate residual errors. This control method can 
effectively ensure the rapid response and steady-state 
accuracy of building heating problems, which is suitable 
for heating systems with rapid dynamic changes. When 
the room temperature, temperature deviation and 
circulation flow are set, the four-way valve opening can 
be calculated by PI control, and then intelligent 
adjustment can be achieved. The output result of the 
controller is shown in Equation (16). 
( ) 1
( ) ( ) ( ) pd
i
de t
u t K e t T e t dt
dt T

= + +



(16) 
In Equation (16), t represents the set temperature. 
() et indicates input deviation. p K is a proportional 
factor. i T is the integral time constant. d T is the 
differential time constant. Particle Swarm Optimization 
(PSO) is used to solve the constrained cost function in 
predictive control. As an intelligent algorithm to simulate 
predation behavior, PSO achieves the optimal solution 
through fitness update and algorithm iteration. The PSO 
is applied to the intelligent regulation of building heating 
to find the optimal value at the corresponding time and 
avoid sudden changes in system flow during the 
temperature regulation process. The PSO has strong 
global search ability, which can effectively search for the 
global optimal solution in the solution space. It has 
significant advantages for multi-variable and 
multi-constraint optimization problems in building 
heating systems. Compared with other optimization 
algorithms such as GA and simulated annealing, the 
concept of PSO is simple, easy to program and 
implement, with less parameter adjustment, which 
reduces the difficulty of algorithm implementation. This 
algorithm can be adjusted according to specific heating 
systems and needs, adapting to different constraint 
conditions and cost functions, which has good 
adaptability and parallel processing ability. The 
mathematical expression of the fitness function is shown 
in Equation (17). 
 
 
2
2
min max
( ( )
()
seti i J Min q T T r dG
G G G
= − + 

  (17) 
In Equation (17), 
q
 represents the weight 
coefficient affecting the response speed of the control 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 161 
 
system. r represents the weight factor affecting the 
stability of the control system. seti T indicates the set 
temperature. i T indicates the room temperature. , G dG 
are the flow rate and the flow increment. max min , GG and 
max min , GG indicate the maximum and minimum values 
of the set flow. The control device designed by PI has 
less control, and its connection with the sensor can better 
control the temperature, the difference of water supply 
and return pressure, and the valve operation. 
4 Application effect of uniform 
building heating under the 
regulation method 
The thermal comfort inside the building is an important 
index to ensure the residential comfort. The heating 
system under intelligent reverse regulation is designed 
through the system program design to achieve uniform 
heating effect. Firstly, the performance of the proposed 
regulation method is tested, including the change of 
energy consumption and uniform heating. The hardware 
equipment involved in the research mainly includes 
intelligent heating equipment (valves, four-way 
directional valves, ultrasonic flow meters, and 
temperature sensors). The temperature sensors is mainly 
installed on the water supply and return pipelines in the 
hot well. The study uses MATLAB software to calculate 
the actual operating data of the building heating system, 
with a data sampling interval of 1 hour. A total of 24 data 
points are collected within a day and combined with the 
corresponding meteorological temperature at that time. 
The collected normal operation data of the heating system 
is used as training and testing data for predicting the 
heating load of the heating system. The design parameters 
of heating data include primary return water temperature, 
instantaneous load, primary supply and return water flow 
rate, and cumulative load. After that, the identification 
results of changed energy balance coefficient and 
commutation time ratio are analyzed, and then the 
parameter sensitivity under the model predictive control 
is analyzed. The result is shown in Figure 6. 
Time(h)
2 6 4 10 8 14 12 18 16 22 20
Indoor temperature ( ℃)
0
17.0
17.5
18.0
18.5
19.0
19.5
20.0
20.5
21.0
20.0
20.5
21.0
21.5
22.0
22.5
23.0
23.5
24.0
24
Flow (m
3
/h)
Solar Radiant intensity intensity (W/m
2
)
600
500
400
300
200
100
0
(b) Disturbance changes in solar radiation
Indoor temperature
Flow
Solar Radiant intensity intensity
Time(h)
(a) Outdoor temperature disturbance changes
2 6 4 10 8 14 12 18 16 22 20
Indoor temperature ( ℃)
0
17.0
17.5
18.0
18.5
19.0
19.5
20.0
20.5
21.0
21.4
21.6
21.8
22.0
22.2
22.4
22.6
22.8
23.0 2
0
-2
-4
-6
-8
-10
24
Flow (m
3
/h)
Outdoor temperature ( ℃)
Indoor temperature
Flow
Outdoor temperature
 

162   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
Figure 6: Parameters under the change of outdoor temperature and solar radiant intensity 
 
In Figure 6, there is a negative correlation between 
the variation curve of system flow and outdoor 
temperature. When the temperature difference is larger, 
the heat loss rate is accelerated, and the parameter 
controller can adjust the flow rate according to the change. 
The solar radiation intensity from 7:00 to 18:00 first 
increased and then decreased. The controller can better 
adjust the flow rate to achieve indoor temperature 
regulation. Then, the temperature regulation effect of the 
proposed predictive control method is analyzed, and the 
results are shown in Figure 7. 
19.2
19.0
18.8
18.6
18.4
18.2
18.0
17.8
0 12 24 36 48 52 72
0 12 24 36 48
12.5
15.0
17.5
20.0
22.5
Temperature(℃)
Time(h)
Flow (m
3
/h)
Time(h)
PID
Improved PI
Setting temperature
 
Figure 7: Adjustment effects of different methods at set temperatures 
 
In Figure 7, the predictive control method adopted in 
the study showed a lower temperature difference with the 
limited temperature. The maximum temperature 
difference between 0-24 hours and 24-72 hours was 
1.25% and 3.24%. However, the difference between the 
traditional PID algorithm and the limit temperature varied 
greatly, and the maximum error reached 7.38%. The 
improved PI algorithm had lower traffic consumption 
than PID algorithm under the flow variation. When the 
time was 27 hours, the flow rate was 12.6m3/h, which 
had a good response time and anti-interference. 
In Figure 8, when the outdoor temperature fluctuated 
greatly and had obvious node fluctuation, the water 
supply temperature of the heating system was always 
kept between 36°C and 42°C. The overall change 
appeared stages, the volatility was small. The proposed 
heating system had obvious low energy consumption. 
Then, the indoor temperature and water supply 
temperature of the heating system proposed in the study 
are analyzed, and the results are shown in Figure 8. 
-14
-12
32
30
36
34
44
42
48
46
50
38
40
-10
-8
-6
-4
-4
-2
0
2
4
8
6
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Time(weeks)
Outdoor temperature ( ℃)
Heating temperature ( ℃)
Outdoor temperature
Heating temperature
(a)  Water temperature variation under heating system
 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 163 
 
50
45
500 1000 1500 2000 2500
40
35
30
25
20
15
10
5
0
225 0 725 1225 1725 2250 2750
Time(h)
Energy consumption (KW/h)
Energy consumption(KW/h)
Heating temperature
(b) Energy consumption under heating system
 
Figure 8: Water temperature change and energy consumption in hydronics 
 
In Figure 8, the outdoor temperature was higher than 
the indoor temperature on the whole, and its fluctuation 
was relatively large. Under different solar radiation 
intensity, the overall temperature change of the heating 
system was relatively small when the temperature was 
adjusted. The prediction method proposed in the study is 
analyzed, and the results are shown in the Figure 9. 
 
5
0.5
0
0.3
0.4
Sample
0.2
10 15 20 25
Actual value
Predicted value
(a) Temperature measurement
Performance data
0.1
5
3000
0
1500
2000
Sample
1000
10 15 20
Actual value
Predicted value
(b) Energy consumption analysis
Performance data
500
2500
 
Figure 9: Predicted values for temperature measurement and energy consumption analysis 
 
From Figure 9, the prediction method proposed in 
the study showed that the predicted result curve was 
basically consistent with the actual result curve during 
temperature measurement and energy consumption 
analysis, with an overall small deviation. Therefore, the 
method proposed in the study can effectively monitor 
changing temperatures, which is highly feasible. The heat 
load of the building heating system is analyzed and 
compared with algorithms such as Back Propagation 
Neural Network (BPNN), GA, and extreme vector 
machine. The Mean Absolute Error (MAE), Mean 
Squared Error (MSE), Root Mean Square Error (RMSE), 
and Mean Absolute Percentage Error (MAPE) are shown 
in the Table 2. 
 
Table 2: Heating system heat load calculation 
Algorithm 
Mean absolute 
error 
Mean squared 
error 
Root mean 
square error 
Mean absolute 
percentage error/% 
BPNN 0.532 0.415 0.637 1.543 
GA 0.643 0.463 0.596 1.261 
Extreme vector machine 0.579 0.377 0.584 1.126 

164   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
Research algorithm 0.376 0.205 0.432 0.817 
 
The results in Table 2 indicated that the proposed 
prediction method exhibited lower error results, with 
significantly lower values than other comparative 
algorithms in MAE (0.376), MSE (0.205), RMSE (0.432), 
and MAPE (0.817) metrics. Among them, the GA 
exceeded 0.6 in all four indicators, with a maximum 
value of 1.453. The above results indicate that the 
prediction method proposed in the study can effectively 
analyze the thermal load conditions of buildings. The 
thickness of the external wall and roof insulation layer is 
replaced by the external wall heat transfer coefficient and 
the roof heat transfer coefficient, respectively. The 
external window is replaced by its corresponding heat 
transfer coefficient and solar thermal gain coefficient. 
The processed data is subjected to multiple regression 
analysis, as shown in Table 3. 
 
 
Table 3: Multiple regression results 
Variable Standard error t Sig 
Aspect ratio 0.135 4.032 0.000 
Facing 0.253 10.261 0.000 
External window heat transfer coefficient 0.002 17.134 0.000 
The external wall receives sparse heat from the sun 0.072 -8.564 0.000 
External wall heat transfer coefficient 0.431 14.302 0.000 
Roof heat transfer coefficient 0.396 21.365 0.000 
 
The variables in Table 3 all had obvious significance. 
The fitting effect was good when the table was adjusted 
and calculated. The above results indicate that the 
proposed method has good effectiveness. Then, a 
high-rise building in the north is selected as the 
experimental object. The building covers an area of more 
than 8,000m2, with a total of 20 floors (including one 
underground). The building load ranges from 190 to 
210KW. The heating system of the building is low 
temperature floor radiant heating system, according to the 
low zone and high zone heating. In this study, six 
households in two districts are selected to install water 
supply and return temperature measurement points and 
heating equipment. The heating system is placed in the 
pipeline well in the form of household installation. At the 
same time, in order to ensure the normal heating demand 
under the adjustment effect, a smart temperature and 
humidity sensor is installed in the user's home to observe 
the temperature change. In the results of building 
intelligent management, the test is carried out in two 
forms: manual control and automatic control. The test 
time is the end of the heating time in the area (about 120 
days). The application results of the heating system are 
shown in Figure 10. 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 165 
 
Hour(h)
Solar Radiant intensity 
intensity (W/m
2
)
Water supply temperature (℃)
0 24 48 72 96 120 168 144
0 24 48 72 96 120 168 144
Outdoor temperature (℃)
Indoor temperature ( ℃)
0.5
1.0
1.5
2.0
2.5
3.0
0
-14
-12
-10
-8
-6
-4
-2
0
2
4
100
200
300
400
500
600
700
30
35
40
45
50
0
(b) Change of water supply temperature under Radiant intensity intensity
Solar Radiant intensity intensity
Water supply temperature
Outdoor temperature
Indoor temperature
Hour(h)
(a) Outdoor temperature and indoor temperature changes
 
Figure 10 Indoor and outdoor temperature changes 
2 6 4 10 8 14 12 18 16 22 24 20
50
60
40
30
20
10
Quantity of heat Q
g
(kW)
Time(h)
(a) Comparison Chart of Heating Capacity in Low Area
0
70
80
Manual mode
Intelligent adjustment mode
2 6 4 10 8 14 12 18 16 22 24 20
30
35
25
20
15
10
Quantity of heat Q
g
(kW)
Time(h)
(b) Comparison Chart of Heating Capacity in High Area
0
40
45
Manual mode
Intelligent adjustment mode
 
Figure 11: Changes in heating capacity of two zones under different modes 
 
In Figure 11, the heat supply of high floors was 
obviously less than that of low floors. Under manual 
control mode, the heat supply of low and high floors was 
basically maintained between 70kW and 80kW and 
between 35kW and 40kW. Under the intelligent reverse 
regulation mode, the overall heat supply of the low floor 

166   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
showed a decreasing trend from 80kW to 10kW. The 
heating capacity of high floors was basically maintained 
between 20kW and 35kW. Whether on high or low floors, 
the intelligent reverse adjustment mode saved heating for 
over 80% of the time. After the improvement, the average 
heating capacity of the lower and upper floors was 40kW 
and 28kW, which was 57.14% and 73.68% of the energy 
consumption before the improvement, effectively 
realizing the indoor temperature control in the state of 
energy saving. Then, the change of the room temperature 
inside the building under the two heating modes is 
statistically analyzed. The average temperature change is 
expressed, as shown in Figure 12. 
2 6 4 10 8 14 12 18 16 22 24 20
24.0
23.0
23.5
22.5
22.0
21.5
21.0
20.5
20.0
19.5
Temperature (℃)
Time(h)
(a) Room temperature comparison of two modes in the low zone
0
2 6 4 10 8 14 12 18 16 22 24 20
24.0
23.0
23.5
22.5
22.0
21.5
21.0
20.5
Temperature( ℃)
Time(h)
(b) Room temperature comparison of two modes in the high zone
0
Manual mode
Intelligent adjustment mode
Manual mode
Intelligent adjustment 
mode
24.5
25.0
 
Figure 12: Room temperature changes under different control modes 
 
In Figure 12, the average indoor temperature in the 
manual control mode was high, and the low floor was 
between 22.5°C and 23.5°C. The high floor was between 
23.5°C and 24.5°C, which was significantly higher than 
the indoor heating standard temperature stipulated by the 
state. The average temperature of the low floor in the rear 
of the intelligent reverse control system was between 
21°C and 23°C. The temperature of the high floor was 
between 22°C and 23°C. On the whole, the temperature 
was 2°C to 3°C lower than the manual adjustment mode. 
In addition, the day and night temperature changes in the 
manual adjustment mode were more volatile, which 
increased the power consumption of the internal heating 
system to a certain extent. Subsequently, the parameters 
under different disturbances are analyzed to further 
understand the heating system. The results are shown in 
the Figure 13. 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 167 
 
21.0
20.5
20.0
19.5
19.0
18.5
18.0
17.5
17.0
23.0
22.8
22.6
4
0
-4
24.0
23.5
600
500
21.0
20.5
20.0
19.5
19.0
18.5
18.0
17.5
17.0
22.6
22.6
22.6
22.6
22.6
22.6
-8
-12
-16
-20
23.0
22.5
22.0
21.5
21.0
20.5
20.0
400
300
200
100
0
0 4 8 12 16 20 24 0 4 8 12 16 20 24
Indoor temperature/ ℃
Indoor temperature/ ℃
Flow/(m
3
/h)
Outdoor temperature/ ℃
Flow/(m
3
/h)
Solar radiation/(W/m
2
)
Indoor temperature
Outdoor temperature
Flow
Indoor temperature
Flow
Solar radiation
Hour/h
Hour/h
(a) Parameter changes under outdoor 
temperature disturbances
(b) Parameter changes under solar 
radiation disturbance
21.0
20.5
20.0
19.5
19.0
18.5
18.0
17.5
17.0
24.5
24.0
23.5
23.0
22.5
0 4 8 12 16 20 24
Indoor temperature/ ℃
Flow/(m
3
/h)
Hour/h
(c) Parameter changes under wind 
speed disturbance
5
4
3
1
0
2
Wind speed/(m/s)
Indoor temperature
Flow
Wind speed
 
Figure 13: Parameters under disturbance changes of different factors 
 
Figures 13 (a), (b), and (c) show the disturbance 
changes of outdoor temperature, solar radiation, and wind 
speed factors, respectively. Specifically, the system flow 
rate was negatively correlated with outdoor temperature. 
From the perspective of heat transfer theory, the larger 
the temperature difference between indoor and outdoor 
buildings, the faster the heat loss. Therefore, when the 
outdoor temperature drops, the controller will adjust to 
increase flow. When the outdoor temperature rises, the 
controller will adjust to reduce flow. From the indoor 
temperature and flow rate changes, when the solar 
radiation intensity undergoes drastic changes, the thermal 
intelligent controller can predict the indoor temperature 
changes in advance and make adjustments. Overall, there 
is a negative correlation between system flow and solar 
radiation intensity. When calculating the thermal load of 
a building, wind energy consumption is usually added to 
the enclosure structure. When controlling constant solar 
radiation, outdoor temperature, and water supply 
temperature, the system flow rate in building thermal 
intelligent control is positively correlated with wind 
speed. In order to further explore the temperature 
regulation under the intelligent commutation adjustment 
method, the study conducts experimental observation on 
the low floor. The continuous changes in room 
temperature for five days are shown in Figure 14. 

168   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
18.0
17.0
16.0
15.0
14.0
13.0
12.0
19.0
Temperature ( ℃)
Time(h)
2 0 6 4 10 8 14 12 18 16 22 24 20
18.0
17.0
16.0
15.0
14.0
13.0
12.0
19.0
Temperature ( ℃)
Time(h)
2 0 6 4 10 8 14 12 18 16 22 24 20
(a) Original mode room temperature distribution map
First
Second
Third
Fourth
First Second Third Fourth
(b) Room temperature distribution map in intelligent 
adjustment mode
 
Figure 14: Room temperature distribution map before and after intelligent adjustment 
 
In Figure 14, the temperature difference between the 
three floors was large under the positive room 
temperature condition. When the time was less than 10 
hours, the average temperature of floors 1-4 was 13.8°C, 
14.9°C, 16.2°C and 17.9°C, respectively. The four floors 
all showed a fluctuation rise and a slight decline within 
12-16 hours. After more than 18 hours, the average 
temperature of the floor showed a downward trend, and 
the maximum floor temperature difference exceeded 4°C. 
After the commutation temperature adjustment, the 
temperature imbalance between the four floors is 
obviously alleviated. Although there are still temperature 
fluctuations, the overall temperature difference tends to 
be stable, which effectively realizes temperature 
regulation. 
5 Discussion 
On the basis of considering the complexity of building 
energy systems and the variability of heating influencing 
factors, this study proposes a design concept for a 
uniform heating system based on intelligent directional 
control and regulation. It is implemented based on key 
parameter identification, series parallel control system 
design, and algorithm design. Performance testing 
analysis is conducted on the research method. According 
to the results, there was a negative correlation between 
the system flow rate change curve and outdoor 
temperature. The parameter controller could adjust the 
flow rate based on changes in solar radiation intensity to 
achieve indoor temperature control. The method proposed 
in the study predicted a maximum temperature difference 
of 1.25% and 3.24% within 0-24 hours and 24-72 hours, 
which was significantly better than comparison 
algorithms. The improved PI algorithm showed lower 
flow consumption compared with the PID algorithm, with 
a flow rate of 12.6m3/h within 27 hours, and had better 
response time and anti-interference performance. 
Accurate parameter identification can effectively grasp 
various response conditions changes inside and outside 
the building, thereby improving identification accuracy. 
The designed water supply system designed has obvious 
advantages of low energy consumption. The energy 
balance coefficient can effectively consider the impact of 
heat exchangers and material thermal conductivity on 
building heating heat. Compared with the literature [2] 
using an improved long short-term memory prediction 
network, the literature [2] may have some shortcomings 

Application of Intelligent Reverse Regulation in Uniform Heating … Informatica 48 (2024) 153 –170 169 
 
in reading and adapting to real-time changes, while the 
intelligent directional adjustment proposed in the study 
can adjust heating parameters in real-time and better 
maintain temperature balance. Compared with the 
literature [8] that uses the distributed co-evolutionary 
optimization algorithm, the improved approach proposed 
in the study focuses more on real-time control 
performance, which is different from the focus on 
structural efficiency in the literature [8]. 
The application performance results showed that 
under the intelligent directional control mode, the overall 
heating supply of low floors showed a decreasing trend 
from 80kW to 10kW, while the heating supply of high 
floors basically remained between 20kW and 35kW. 
More than 80% of the time, the intelligent directional 
adjustment mode saved more heat supply. The average 
heat supply of the low and high floors in the building 
after improvement was 40kW and 28kW, which was 
57.14% and 73.68% of the energy consumption before 
improvement. Reverse temperature regulation can 
effectively alleviate temperature imbalance on different 
floors, and the overall temperature regulation effect is 
relatively obvious. The reason for this result is that during 
the design of control strategies, the study explores the 
time adjustment of heating and heat dissipation variables, 
effectively considering the delay of indoor heating 
systems in buildings. Compared with the approach 
proposed in literature [14] that combined attention 
mechanism and long short-term memory network to 
predict heating load, the proposed method combines the 
water and electricity system to observe the internal air 
source heat situation of buildings. PI control can calculate 
the opening of four-way valves, which can be better 
connected to sensors. Therefore, it has higher sensitivity 
and adaptability to data reflection. Although literature [14] 
can maintain indoor temperature stability well, its 
approach using algorithmic calculations is difficult to 
timely examine the dynamic response of the system. The 
literature [11] utilizes TOPSIS multi-criteria 
decision-making to evaluate building temperature 
sensitivity issues. Although it can effectively ensure 
thermal adaptability, compared with research methods, its 
impact on the heating system itself is not sufficiently 
considered. 
6 Conclusion 
Aiming at the temperature imbalance in building heating 
system, the intelligent reverse regulation equipment was 
introduced into the heating system. The intelligent 
uniform regulation system was constructed to achieve 
uniform heating. When the indoor temperature remained 
unchanged, the flow curve of the heating system was 
negatively correlated with the outdoor temperature. The 
designed parameter controller had obvious effects in 
temperature regulation. When the flow rate changed, the 
flow rate of the improved PI algorithm was 12.6m3/h at 
27 hours, which had better response time and 
anti-interference. The application results showed that the 
intelligent reverse adjustment mode was better than the 
traditional manual adjustment mode. Its low floor (80kW 
to 10kW) and high floor heating capacity (20kW to 35kW) 
was much lower than the manual control mode between 
70kW to 80kW and 35kW to 40kW. The overall indoor 
temperature adjustment was 2°C~3°C lower than the 
manual adjustment mode, which effectively improved the 
temperature imbalance on the floor. The intelligent 
control method for uniform heating in buildings is 
applicable to all heating systems. The intelligent control 
method for uniform heating in buildings can effectively 
avoid overheating and achieve relatively uniform heating 
distribution in building spaces, which has a positive 
impact on the intelligent temperature regulation and 
energy control of buildings. Intelligent temperature 
regulation can be achieved by installing sensors and 
four-way valves. The architectural layout, orientation, 
body shape coefficient, enclosure structure, shading 
method, ventilation, noise, lighting, and other aspects all 
have impacts on the indoor thermal environment of 
residential buildings before and after the heating period. 
In the actual installation process, design values should be 
considered based on on-site conditions. The difference 
between the deployment range of temperature sensors and 
the selected type will to some extent affect the calculation 
cost. This study proposes a building uniform heating 
system based on intelligent directional adjustment, which 
can effectively regulate temperature and improve thermal 
imbalance. However, there are still certain shortcomings 
in the research, among which the energy regulation 
coefficient in the energy balance method still needs to 
consider the comprehensive factors such as temperature 
and flow rate, and strengthen the study for temperature 
distribution under different commutation time ratios. At 
the same time, the coupling relationship between the 
energy balance method and the reverse process should be 
comprehensively considered to optimize the system 
heating. 
References 
[1] E. Prem, M. Muthtamilselvan, S. Muthukumar, S. 
Sureshkumar, and A. Malleswaran, "Effects of 
uniform or non-uniform heating at bottom wall on 
MHD mixed convection in a porous cavity saturated 
by nanofluid," International Journal of Nonlinear 
Sciences and Numerical Simulation, vol. 23, no. 2, 
pp. 177-196, 2022. 
https://doi.org/10.1515/ijnsns-2017-0258 
[2] J. Song, L. Zhang, G. Xue, Y. Ma, S. Gao, and Q. 
Jiang, "Predicting hourly heating load in a district 
heating system based on a hybrid CNN-LSTM 
model," Energy and Buildings, vol. 234, no. 3, pp. 
110998, 2021. 
https://doi.org/10.1016/j.enbuild.2021.110998 
[3] T. Szul, S. Tabor, and K. Pancerz, "Application of the 
BORUTA algorithm to input data selection for a 

170   Informatica 48 (2024) 153 –170                                                                    X. Wu 
 
model based on rough set theory (RST) to prediction 
energy consumption for building heating," Energies, 
vol. 14, no. 10, pp. 2779, 2021. 
https://doi.org/10.3390/en14102779. 
[4] E. Psimopoulos, E. Bee, J. Widen, and C. Bales, 
"Techno-economic analysis of control algorithms 
for an exhaust air heat pump system for detached 
houses coupled to a photovoltaic system," Applied 
Energy, vol. 249, no. 9, pp. 355-367, 2019. 
https://doi.org/10.1016/j.apenergy.2019.04.080 
[5] A. Aliferov, P. Di Barba, F. Dughiero, M. Forzan, S. 
Lupi, M E. Mognaschi, and E. Sieni, "Optimal 
design methods for the uniform heating of tube ends 
for stress relieving," COMPEL, vol. 39, no. 1, pp. 
12-20, 2020. 
https://doi.org/10.1108/COMPEL-05-2019-0204 
[6] X. Zhang, J. Lei, and M. Zhang, "Research on control 
method of valve-controlled heating system," Journal 
of Physics: Conference Series, no. 1345, no. 3, pp. 
032081-032086, 2019. 
https://doi.org/10.1088/1742-6596/1345/3/032081 
[7] I. Sarbu, M. Mirza, and E. Crasmareanu, "A review of 
modelling and optimisation techniques for district 
heating systems," International Journal of Energy 
Research, vol. 43, no. 13, pp. 6572-6598, 2019. 
https://doi.org/10.1002/er.4600 
[8] W. Devia, K. Agbossou, and A. Cardenas, "An 
evolutionary approach to modeling and control of 
space heating and thermal storage systems," Energy 
and Buildings, vol. 234, no. 3, pp. 
110674.1-110674.18, 2020. 
https://doi.org/10.1016/j.enbuild.2020.110674 
[9] C. Wei, Y. Huang, and A. Loeschel, "Recent 
advances in energy demand for residential space 
heating," Energy and Buildings, vol. 4, no. 261, pp. 
1119651-1119656, 2022. 
https://doi.org/10.1016/j.enbuild.2022.111965 
[10] C. S. Ioakimidis, "Use of AI algorithms in different 
building typologies for energy efficiency towards 
smart buildings," Buildings, vol. 11, no. 12, pp. 613, 
2021. https://doi.org/10.3390/buildings11120613 
[11] M. Esrafilian-Najafabadi and F. Haghighat, 
"Occupancy-based HVAC control using deep 
learning algorithms for estimating online 
preconditioning time in residential buildings," 
Energy and Buildings, vol. 252, no. 12, pp. 
111377.1-111377.15, 2021. 
https://doi.org/10.1016/j.enbuild.2021.111377 
[12] Y. Lu, Z. Tian, P. Peng, J. Niu, W. Li, and H. Zhang, 
"GMM clustering for heating load patterns in-depth 
identification and prediction model accuracy 
improvement of district heating system," Energy 
and Buildings, vol. 190, no. 5, pp. 49-60, 2019. 
https://doi.org/10.1016/j.enbuild.2019.02.014 
[13] S. Duan, W. Yang, X. Wang, S. Mao, and Y. Zhang, 
"Temperature forecasting for stored grain: A deep 
spatiotemporal attention approach," IEEE Internet of 
Things Journal, vol. 8, no. 23, pp. 17147-17160, 
2021. https://doi.org/10.1109/JIOT.2021.3078332 
[14] G. Xue, C. Qi, H. Li, X. Kong, and J. Song, "Heating 
load prediction based on attention long short-term 
memory: A case study of xingtai," Energy, vol. 203, 
no. 15, pp. 117846.1-117846.17, 2020. 
https://doi.org/10.1016/j.energy.2020.117846 
[15] S. Salo, A. Hast, J. Jokisalo, R. Kosonen, S. Syri, J. 
Hirvonen, and K. Martin, "The impact of optimal 
demand response control and thermal energy storage 
on a district heating system," Energies, vol. 12, no. 9, 
pp. 1678, 2019. https://doi.org/10.3390/en12091678 
[16] Q. Chen and N. Li, "Thermal response time 
prediction-based control strategy for radiant floor 
heating system based on gaussian process 
regression," Energy and Buildings, vol. 263, no. 5, 
pp. 112044.1-112044.17, 2022. 
https://doi.org/10.1016/j.enbuild.2022.112044 
[17] B. Vand, K. Martin, J. Jokisalo, R. Kosonen, and A. 
Hast, "Demand Response Potential of District 
Heating and Ventilation in an Educational Office 
Building," Science and Technology for the Built 
Environment, vol. 26, no. 3, pp. 10-21, 2019. 
https://doi.org/10.1080/23744731.2019.1693207 
[18] Y. Guo, Z. Mustafaoglu, and D. Koundal, "Spam 
detection using bidirectional transformers and 
machine learning classifier algorithms," Journal of 
Computational and Cognitive Engineering, vol. 2, 
no. 1, pp. 5-9, 2023. 
https://doi.org/10.47852/bonviewJCCE2202192 
[19] P. Dey and D. K. Jana, "Evaluation of the 
convincing ability through presentation skills of 
pre-service management wizards using AI via T2 
linguistic fuzzy logic," Journal of Computational 
and Cognitive Engineering, vol. 2, no. 2, pp. 
133-142, 2022. 
https://doi.org/10.47852/bonviewJCCE2202158