CN111597609A - Basic operation unit containing solar radiation and building energy consumption rapid simulation method applying same - Google Patents
Basic operation unit containing solar radiation and building energy consumption rapid simulation method applying same Download PDFInfo
- Publication number
- CN111597609A CN111597609A CN202010351293.3A CN202010351293A CN111597609A CN 111597609 A CN111597609 A CN 111597609A CN 202010351293 A CN202010351293 A CN 202010351293A CN 111597609 A CN111597609 A CN 111597609A
- Authority
- CN
- China
- Prior art keywords
- rad
- wall
- area
- energy consumption
- solar radiation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/06—Power analysis or power optimisation
Abstract
The invention relates to the technical field of energy consumption prediction, in particular to a basic operation unit containing solar radiation and a building energy consumption rapid simulation method applying the same. A basic arithmetic Unit containing solar radiation, wherein the basic arithmetic Unit is one of 5 basic types, and the 5 basic types are unitsground‑ceiling、Unitground‑roof、Unitfloor‑ceiling、Unitfloor‑roofAnd Unitexposed‑ceilingThe basic operation unit comprises geometric parameters and sunA radiation dose parameter. A building energy consumption rapid simulation method comprises the step that parameters of the basic operation unit are used for machine learning algorithm generation and building energy consumption rapid simulation feedback. The invention brings the annual radiation calculation into the machine learning training parameters for the first time, and solves the defect that the influence factors of the environment outside the building cannot be considered when the current machine learning algorithm simulates the building energy consumption; the defect is solved, so that the accuracy of the machine learning algorithm is greatly improved, and the practicability of the machine learning algorithm is improved.
Description
Technical Field
The invention relates to the technical field of energy consumption prediction, in particular to a basic operation unit containing solar radiation and a building energy consumption rapid simulation method applying the same.
Background
The building energy consumption accounts for an important proportion in the energy consumption of China, and the optimization aiming at the building energy consumption in the building design stage has important significance for realizing the sustainable development of cities under the large background of smart cities, energy conservation and emission reduction. However, the current building energy consumption simulation accuracy and the operation speed cannot be simultaneously considered:
on one hand, the traditional physical simulation method is extremely time-consuming in operation, and for a monomer building or a building group with a complex form, one accurate energy consumption prediction has to take hours or even days; a part of operation time can be reduced through combined calculation of similar floors and rooms, complicated external environment radiation is caused due to shadow of an external building or self-shielding of a building body type in a real environment, and the model without considering the external environment radiation is simplified and reduces the energy consumption prediction accuracy.
On the other hand, machine learning algorithms typified by neural network techniques offer the possibility of rapid energy consumption simulation. The machine learning algorithm extracts knowledge from existing sample data and predicts the performance of the newly input sample with acceptable accuracy. Due to the complexity of building form control parameters and the complex interaction between the building form and the surrounding environment, previous researches on energy consumption simulation by means of machine learning are only suitable for parameter optimization of simple block quantity or optimization of building envelope structure under the premise of fixed quantity (CN109255472A, CN104134097A and CN 109409605A). These simulation methods are not suitable for comparing a large number of building schemes with complex shapes in the actual building design process. The only research (CN109033595A) trying to solve this problem is to add the energy consumption of the simple rectangular parallelepiped units by using the shape cutting method to obtain the total energy consumption of the building, but this method only guesses the energy consumption through the basic geometric dimensions (length, width, height, window-wall ratio) of the rectangular parallelepiped units, ignores the influence of external short and long wavelength radiation factors on the energy consumption, and potentially affects the energy consumption prediction accuracy.
The neglect of the radiation of the external environment by the machine learning algorithm can cause large-scale deviation of energy consumption, and a distorted energy consumption simulation result cannot effectively guide a designer to optimize a building scheme, so that the feasibility of introducing actual project practice by the machine learning algorithm is directly influenced.
Disclosure of Invention
The invention provides a basic operation unit containing solar radiation to solve the problem that the accuracy is low due to the fact that external environment radiation cannot be considered in a building energy consumption prediction model in the prior art.
In order to solve the technical problems, the invention adopts the technical scheme that: a basic arithmetic Unit containing solar radiation, wherein the basic arithmetic Unit is one of 5 basic types, and the 5 basic types are unitsground-ceiling、Unitground-roof、Unitfloor-ceiling、Unitfloor-roofAnd Unitexposead-ceiling(ii) a Wherein Unitground-ceilingThe lower surface of the basic operation unit is a terrace surface which generates heat exchange with the ground, and the upper surface is an interlayer ceiling; unitground-roofThe lower surface of the basic operation unit is a terrace surface which generates heat exchange with the ground, and the upper surface is a roof surface exposed to the external environment; unitfloor-ceilingThe lower surface of the basic operation unit is a floor surface between layers, and the upper surface is a ceiling surface between layers; unitfloor-roofThe lower surface of the basic operation unit is a floor surface between floors, and the upper surface is a roof surface exposed to the external environment; unitexposed-ceilingThe lower surface of the basic operation unit is an overhead layer surface exposed to the external environment, and the upper surface is an interlayer floor surface; the basic operation unit comprises geometric parameters and solar radiation quantity parameters.
Preferably, the basic operation unit is a cuboid, and the geometric parameters of the basic operation unit comprise the length X, the width Y and the height Z of the cuboid; taking the plane where the length and the height of the cuboid are located as a reference plane, wherein the azimuth angle between the reference plane and the local north direction is Rotate; an inner wall portion of each face of the basic operation unit andthe ratio of the total area of the faces is AdiabaliciThe ratio of the window to the wall of the outer wall part on each side of the basic operation unit is WWRiI.e. the ratio of the window area to the exterior wall area, where i denotes each side of the basic arithmetic unit, i is 1 to 4.
Preferably, the solar radiation quantity parameter includes solar radiation quantity Rad _ wall received by any side outer wall of the basic operation unitiAnd the solar radiation amount Rad _ win received by the outer window on the side surfacei。
Preferably, the amount Rad _ wall of solar radiation received by the outer walliIncluding Rad _ walli(low)、Rad_walli(mid) and Rad _ walli(high), wherein Rad _ walli(low) is the amount of solar radiation received by the side surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ walli(mid) is the amount of solar radiation received by the side when the outdoor air temperature is in the indoor set temperature zone; rad _ walli(high) is the amount of solar radiation received by the side when the outdoor air temperature is higher than the indoor set temperature; the Rad _ winiIncluding Rad _ wini(low)、Rad_wini(mid) and Rad _ wini(high), wherein Rad _ wini(low) is the solar radiation quantity received by the external window on the side surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ wini(mid) is the amount of solar radiation received by the external window on the side when the outdoor air temperature is in the indoor set temperature zone; rad _ wini(high) is the amount of solar radiation received by the exterior window on that side when the outdoor air temperature is higher than the indoor set temperature.
Preferably, the basic operation Unit is a Unitground-roofOr Unitfloor-roofIn time, the solar dosimetry parameters further comprise top surface solar radiation Rad-roof, including Rad _ roof (low), Rad _ roof (mid), and Rad _ roof (high); wherein, Rad _ roof (low) is the solar radiation quantity received by the top surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ roof (mid) is the amount of solar radiation received by the top surface when the outdoor air temperature is in the indoor set temperature interval; rad-roof (high) is that when the outdoor air temperature is higher than the indoor set temperature, the top surface is connected withThe amount of solar radiation received.
The invention also provides a building energy consumption rapid simulation method, which comprises the step of using the parameters of the basic operation unit for machine learning algorithm generation and building energy consumption rapid simulation feedback.
Preferably, the machine learning algorithm generation comprises training sample generation, machine learning algorithm training and optimal machine learning algorithm screening;
the training sample is generated by combining the parameters of the basic operation units and randomly supercube sampling, the basic operation units with different geometric parameters and solar radiation quantity parameters are generated in batches by adopting experimental design and are used as the training samples, and the basic operation units are led into energy consumption simulation software to carry out accurate physical simulation so as to obtain energy consumption data of the basic operation units;
the machine learning algorithm training includes two steps:
a: because the selection of the input parameters greatly influences the prediction precision of the machine learning algorithm, the input parameters with the best prediction effect need to be explored, and j groups of input parameters are obtained through a parameter transformation formula;
b: dividing the j training sample combinations after the parameter transformation of the previous step into training sets D according to the proportion0Test set D1Test set D2And will train set D0Respectively inputting the training data into k machine learning algorithms, setting different hyper-parameters p for each machine learning algorithm, and obtaining j x k groups in total, wherein each group has a plurality of machine learning models with hyper-parameter setting combinations
It should be noted that j × K is a numerical value obtained by multiplying j by K;
the optimal machine learning algorithm screening comprises the following steps:
s1: using various types of machine learning algorithm models
Computing test set D1Energy consumption value of, and test set D1Comparing the actual energy consumption to obtain the product of R2Accuracy value of representation
S2: using this accuracy value
Screening optimal hyper-parameter setting model in each type of machine learning algorithm
S3: then using the j x k group candidate machine learning algorithm to the test set D2The prediction accuracy of (2) is calculated to obtain a machine learning algorithm prediction model F with the highest prediction accuracy R2best。
It should be noted that the machine learning algorithm is an artificial neural network, a Gaussian process regression algorithm, a kriging agent model algorithm, a random forest algorithm, a MARS multivariate adaptive regression algorithm, and an SVM support vector machine. Each of the machine learning algorithm prediction models FbestThe branch prediction model comprises 5 basic operation units. The optimal hyper-parameter combinations p obtained by the machine learning algorithm prediction model are different for different basic arithmetic units.
Preferably, in the step a, the parameter information obtained from the basic operation unit can be converted into other associated parameters, and the parameter transformation formula is as follows:
Area＝X*Y
Area_wini+Area_walli＝(Z*Length)*(1-Adiabatici)
(Length＝X，if i＝1，3；Length＝Y，if i＝2，4)
Area_wini＝WWRi*(Area_wini+Area_walli)
Area_walli＝(1-WWRi)*(Area_wini+Area_walli)
Rad-wall_averagei＝Rad_walli/Area_walli；
Rad_win_averagei＝Rad_wini/Area_wini
Rad_roof-average＝Rad_roof/Area
wherein, Area _ winiRepresents the Area of the outer window on the ith side of the basic arithmetic unit, Area _ walliThe area of an outer wall on the ith side of the basic operation unit is shown; rad _ wall _ averageiRad _ win _ average, the amount of solar radiation received by any side wall of the basic arithmetic unitiRepresenting the amount of solar radiation received per unit area of the outer window on the side; rad _ roof _ average represents the amount of solar radiation received per unit area of the top surface; the j sets of input parameters include Configuration1, Configuration2, Configuration3, Configuration4, Configuration5, and Configuration 6; wherein Configuration1 includes X, Y, Z, Area, Rotate, Rad _ walli，Rad_winiThe Configuration2 includes X, Y, Z, Area, Rotate, Adiabalici，Rad_walli，Rad_winiThe Configuration3 includes X, Y, Z, Area, Rotate, Adiabalici，WWRi，Rad_walli，Rad_winiThe Configuration4 includes X, Y, Z, Area, Rotate, Adiabalici，WWRi，Rad_wall_averagei，Rad_win_averageiThe Configuration5 includes X, Y, Z, Area, Rotate, Area _ walli，Area_wini，Rad_walli，Rad_winiThe Configuration6 includes X, Y, Z, Area, Rotate, Area _ walli，Area_wini，Rad_wall_averageiAnd Rad _ win _ averagei。
Preferably, the basic operation Unit is a Unitground-roofOr Unitfloor-roofThe configurations 1, 2, 3 and 5 further include Rad _ roof, the configurations 4 and 6 further include Rad _ roof _ operation, and Rad _ roof _ operation is solar radiation per unit area of the top surface.
Preferably, the building energy consumption rapid simulation feedback comprises the following steps:
a: and (3) building decomposition: the building is decomposed into the building volume which is firstly divided into volume slices of different layers according to the actual building floor height, and then each layer of slices is decomposed into a plurality of cuboid basic operation units;
b: calculating the radiation value of the basic operation unit: using Accelard to carry out fast calculation on the sunshine radiation parameters accelerated by the GPU on the basic operation unit obtained in the step A, wherein the operation speed of the calculation algorithm is not in direct proportion to the fineness degree and the surface complexity of a calculation grid;
c: fast calculation of energy consumption: inputting the geometric parameters and solar radiation quantity parameters of the basic operation unit into a machine learning algorithm prediction model F in batchesbestObtaining the energy consumption value E of all basic operation units of the whole buildingidxThe total energy consumption value of the building is
Wherein n is the number of basic arithmetic units constituting the whole building;
d: fast energy consumption feedback: because the machine learning algorithm is close to the real-time energy consumption feedback speed, the calculated total energy consumption value can be quickly displayed in the building model through a visualization method; the energy consumption values of the basic operation units forming the building volume can be directly colored on the surface of each basic operation unit in a pseudo-color graph mode, and architects can conveniently and quickly find out areas with unfavorable building energy consumption to conduct targeted form optimization.
Compared with the prior art, the beneficial effects are:
firstly, the annual radiation calculation is brought into the machine learning training parameters for the first time, and the defect that influence factors of the environment outside a building cannot be considered when the current machine learning algorithm simulates the building energy consumption is overcome; the defect is solved, so that the accuracy of the machine learning algorithm is greatly improved, and the practicability of the machine learning algorithm is improved.
Secondly, by means of the GPU accelerated solar radiation calculation and the quick feedback of a machine learning algorithm, the simulation time of building energy consumption is reduced from hours in the past to seconds, the accuracy is guaranteed, and the feasibility of introducing actual project practice is achieved.
Thirdly, three groups of radiation parameters Rad associated with the indoor design temperature interval are providedlow，Radmid，RadhighThe relation between the building energy consumption and the external solar radiation to which the building energy consumption is subjected can be accurately described.
And fourthly, a building combination concept including a basic operation unit of solar radiation is provided, so that the building energy consumption prediction precision is improved, and meanwhile, an energy consumption numerical value of each combination unit of the building can be provided and visually presented, so that an architect can visually judge an energy consumption unfavorable area and pertinently improve a scheme.
And fifthly, the machine learning algorithm generation module can screen out the optimal machine learning algorithm without manual screening.
Drawings
FIG. 1 is a schematic diagram of a method for rapid simulation of building energy consumption;
FIG. 2 is a schematic diagram of a machine learning algorithm of the present invention;
FIG. 3 is a diagram of a basic operation unit according to the present invention;
FIG. 4 is an experimental design of a basic arithmetic unit training sample in accordance with the present invention;
FIG. 5 is an exploded view of the construction of the present invention;
FIG. 6 is a schematic surface view of the detailed architectural energy consumption of the present invention, which is colored in a pseudo-color manner on each basic operation unit;
FIG. 7 is a schematic view showing the building size in example 3 of the present invention;
FIG. 8 is a graph showing the total annual solar radiation received by a building surface as rapidly calculated in example 3 of the present invention;
FIG. 9 is a regression analysis chart of predicted energy consumption and physical simulation energy consumption under Beijing climate conditions;
FIG. 10 is a regression analysis plot of predicted energy consumption versus physically modeled energy consumption under Guangzhou climate conditions;
fig. 11 is a schematic diagram showing the predicted building energy consumption value of each basic operation unit in the form of a pseudo color map on the surface of a building three-dimensional model.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent; for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted. The positional relationships depicted in the drawings are for illustrative purposes only and are not to be construed as limiting the present patent.
The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "long", "short", etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the drawings, it is only for convenience of description and simplicity of description, but does not indicate or imply that the device or element referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and therefore, the terms describing the positional relationships in the drawings are only used for illustrative purposes and are not to be construed as limitations of the present patent, and specific meanings of the terms may be understood by those skilled in the art according to specific situations.
The technical scheme of the invention is further described in detail by the following specific embodiments in combination with the attached drawings:
example 1
As shown in fig. 3, 4 and 5, a basic arithmetic Unit containing solar radiation is one of 5 basic types, and 5 basic types are unitsground-ceiling、Unitground-roof、Unitfloor-ceiling、Unitfloor-roofAnd Unitexposed-ceiling(ii) a Wherein Unitground-ceilingThe lower surface of the basic operation unit is a terrace surface which generates heat exchange with the ground, and the upper surface is an interlayer ceiling; unitground-roofThe lower surface of the basic operation unit is a terrace surface which generates heat exchange with the ground, and the upper surface is a roof surface exposed to the external environment; unitfloor-ceilingThe lower surface of the basic operation unit is a floor surface between layers, and the upper surface is a ceiling surface between layers; unitfloor-roofThe lower surface of the basic operation unit is a floor surface between floors, and the upper surface is a roof surface exposed to the external environment; unitexposed-ceilingThe lower surface of the basic operation unit is an overhead layer surface exposed to the external environment, and the upper surface is an interlayer floor surface; the basic operation unit comprises geometric parameters and solar radiation quantity parameters.
The basic operation unit is a cuboid, and the geometric parameters of the basic operation unit comprise the length X, the width Y and the height Z of the cuboid; taking the plane where the length and the height of the cuboid are located as a reference plane, wherein the azimuth angle between the reference plane and the local north direction is Rotate; the ratio of the inner wall part of each surface of the basic arithmetic unit to the total area of the surface is AdiabalticiThe ratio of the window to the wall of the outer wall portion on each side of the basic arithmetic unit is WWRiI.e. the ratio of the window area to the exterior wall area, where i denotes each side of the basic arithmetic unit, i is 1 to 4.
In addition, the solar radiation quantity parameter comprises the solar radiation quantity Rad _ wall received by any side outer wall of the basic operation unitiAnd the solar radiation amount Rad _ win received by the outer window on the side surfacei。
Wherein, the solar radiation amount Rad _ wall received by the outer walliIncluding Rad _ walli(low)、Rad_walli(mid) and Rad _ walli(high), wherein Rad _ walli(low) is the amount of solar radiation received by the side surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ walli(mid) is the amount of solar radiation received by the side when the outdoor air temperature is in the indoor set temperature zone; rad _ walli(high) is the amount of solar radiation received by the side when the outdoor air temperature is higher than the indoor set temperature; rad _ winiIncluding Rad _ wini(low)、Rad_wini(mid) and Rad _ wini(high), wherein Rad _ wini(low) is the solar radiation quantity received by the external window on the side surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ wini(mid) is a set temperature region where the outdoor air temperature is in the indoorThe amount of solar radiation received by the outer window on that side at times; rad _ wini(high) is the amount of solar radiation received by the exterior window on that side when the outdoor air temperature is higher than the indoor set temperature.
In addition, the basic arithmetic Unit is a Unitground-roofOr Unitfloor-roofIn time, the solar dosimetry parameters further include top surface solar radiation Rad-roof, including Rad-roof (low), Rad _ roof (mid), and Rad _ roof (high); wherein Rad _ roof (low) is the amount of solar radiation received by the top surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ roof (mid) is the amount of solar radiation received by the top surface when the outdoor air temperature is in the indoor set temperature interval; rad-roof (high) is the amount of solar radiation received by the ceiling when the outdoor air temperature is above the indoor set temperature.
Example 2
As shown in FIG. 1, a building energy consumption rapid simulation method includes that parameters of the basic operation unit are used for machine learning algorithm generation and building energy consumption rapid simulation feedback.
As shown in fig. 2, the machine learning algorithm generation includes training sample generation, machine learning algorithm training, and optimal machine learning algorithm screening;
the training sample is generated by combining the parameters of the basic operation units and randomly sampling the basic operation units in a hypercube way, the basic operation units with different geometric parameters and solar radiation quantity parameters are generated in batches by adopting experimental design and are used as the training samples, and the basic operation units are led into energy consumption simulation software to carry out accurate physical simulation so as to obtain energy consumption data of the basic operation units;
machine learning algorithm training includes two steps:
a: because the selection of the input parameters greatly influences the prediction precision of the machine learning algorithm, the input parameters with the best prediction effect need to be explored, and j groups of input parameters are obtained through a parameter transformation formula;
b: dividing the j training sample combinations after the parameter transformation of the previous step into training sets D according to the proportion0Test set D1Test set D2And will train set D0Respectively input to k machines for learningTraining the algorithm, setting different hyper-parameters p for each machine learning algorithm, and obtaining j x k groups in total, wherein each group has a plurality of machine learning models with hyper-parameter setting combination
The screening of the optimal machine learning algorithm comprises the following steps:
s1: using various types of machine learning algorithm models
Computing test set D1Energy consumption value of, and test set D1Comparing the actual energy consumption to obtain the product of R2Accuracy value of representation
S2: using this accuracy value
Screening optimal hyper-parameter setting model in each type of machine learning algorithm
S3: then using the j x k group candidate machine learning algorithm to the test set D2The prediction accuracy of (2) is calculated to obtain a machine learning algorithm prediction model F with the highest prediction accuracy R2best。
In addition, in step a, the parameter information obtained from the basic operation unit may be converted into other associated parameters, and the parameter transformation formula is as follows: area ═ X × Y
Area_wini+Area_walli＝(Z*Length)*(1-Adiabatici)
(Length＝X，if i＝1，3；Length＝Y，if i＝2，4)
Area_wini＝WWRi*(Area_wini+Area_walli)
Area_walli＝(1-WWRi)*(Area_wini+Area_walli)
Rad_wall_averagei＝Rad_walli/Area_walli；
Rad_win_averagei＝Rad_wini/Area_wini
Rad_roof_average＝Radroof/Area
Wherein, Area _ winiRepresents the Area of the outer window on the ith side of the basic arithmetic unit, Area _ walliThe area of an outer wall on the ith side of the basic operation unit is shown; rad _ wall _ averageiRad _ win _ average, the amount of solar radiation received by any side wall of the basic arithmetic unitiRepresenting the amount of solar radiation received per unit area of the outer window on the side; rad-roof _ average represents the amount of solar radiation received per unit area of the top surface; 6, 6 sets of input parameters including Configuration1, Configuration2, Configuration3, Configuration4, Configuration5, and Configuration 6; wherein Configuration1 includes X, Y, Z, Area, Rotate, Rad _ walli，Rad_winiGonfiguration2 includes X, Y, Z, Area, Rotate, Adiabaltici，Rad_walli，Rad_winiThe Configuration3 includes X, Y, Z, Area, Rotate, Adiabalici，WWRi，Rad_walli，Rad_winiThe Configuration4 includes X, Y, Z, Area, Rotate, Adiabalici，WWRi，Rad_wall_averagei，Rad_win_averageiThe Configuration5 includes X, Y, Z, Area, Rotate, Area _ walli，Area_wini，Rad_walli，Rad_winiThe Configuration6 includes X, Y, Z, Area, Rotate, Area _ walli，Area_wini，Rad_wall_averageiAnd Rad _ win _ averagei。
In addition, the basic arithmetic Unit is a Unitground-roofOr Unitfloor-roofThe configurations 1, 2, 3 and 5 further include Rad _ roof, the configurations 4 and 6 further include Rad _ roof, Rad _ roof _ averagege is the top surface area solar radiation.
The quick simulation feedback of the building energy consumption comprises the following steps:
a: and (3) building decomposition: the building is decomposed into slices of the building volume which are divided into different layers according to the actual building layer height, and the slices of each layer are further decomposed into cuboid basic operation units;
b: calculating the radiation value of the basic operation unit: using Accelard to carry out fast calculation on the sunshine radiation parameters accelerated by the GPU on the basic operation unit obtained in the step A, wherein the operation speed of the calculation algorithm is not in direct proportion to the fineness degree and the surface complexity of a calculation grid; compared with a CPU (central processing unit) calculation method, the calculation speed and the model complexity are not in a linear relation, so that the calculation time is shortened from minutes to seconds when the radiation value containing a large number of sampling points is calculated;
c: fast calculation of energy consumption: inputting the geometric parameters and solar radiation quantity parameters of the basic operation unit into a machine learning algorithm prediction model F in batchesbestObtaining the energy consumption value E of all basic operation units of the whole buildingidxThe total energy consumption value of the building is
Wherein n is the number of basic arithmetic units constituting the whole building;
d: fast energy consumption feedback: because the machine learning algorithm is close to the real-time energy consumption feedback speed, the calculated total energy consumption value can be quickly displayed in the building model through a visualization method; the energy consumption values of the basic operation units forming the building volume can be directly colored on the surface of each basic operation unit in a pseudo-color graph mode, and architects can conveniently and quickly find out areas with unfavorable building energy consumption to conduct targeted form optimization.
Example 3:
as shown in fig. 7, energy consumption prediction is performed on a city block with a side length of about 140 × 95 m, the total building area of an office building is 33500 square meters, the buildings are all oriented in the north-south direction, the height of the first floor is 4.5m, the height of the standard floor is 4.2m, the north window-wall ratio of all buildings is 0.4, the south window-wall ratio is 0.6, and the window-wall ratio of the east-west sides is 0.2. The building is formed by combining a plurality of high floors and a plurality of floors, the lowest floor of each building group is 3, the highest floor is 15, the complex forms such as block connection, overhead bottom, staggered roof and the like are provided, and the building has a more complex energy consumption simulation environment compared with a single building. Assuming that all building spaces inside the building are in an open office function and have the same building physical properties (surface thermal coefficients), a model is built in a Rhinoceros3D platform, weather files of Beijing and Guangzhou are respectively used for prediction accuracy verification under different climate zones, and building shape decomposition and calculation of related parameters of basic operation units are carried out in a grasshopper platform.
First, 15000 training samples are generated by using weather files of Guangzhou and Beijing, and two machine learning prediction algorithms are trained. And then, importing the relevant parameters of the basic operation unit acquired by the grasshopper platform into a machine learning algorithm and EnergyPlus respectively to calculate and count the running time of the basic operation unit.
Comparing the machine learning predicted energy consumption with the energy consumption obtained by the Energyplus simulation shows that the prediction errors of the refrigeration and heating energy consumption in the two climate partitions of Guangzhou and Beijing are mostly within the range of 5%, wherein the deviation of the heating energy consumption is slightly larger, but the reference significance is still provided, as shown in FIG. 9 and FIG. 10. The typical prediction time of the machine learning algorithm is 40s (including the annual radiation calculation time), the total annual solar radiation received by the building surface is shown in fig. 8, and an accurate energy consumption simulation takes 2.5h, and the speed difference is 220 times. The previous machine learning prediction model can only estimate the whole building energy consumption value, and the rapid simulation algorithm is used for respectively predicting the corresponding energy consumption after the building is decomposed into a large number of basic operation units, so that the annual heating and cooling energy consumption of each building unit can be displayed on the model, as shown in fig. 11, an architect can rapidly and accurately determine the unfavorable position of energy consumption optimization, make targeted body optimization (including adjustment of window-wall ratio, change of building shape, movement of building blocks and the like) and verify the optimization effect in near real time; and the method of calculating the energy consumption by physical simulation consumes too much time, and loses the possibility of repeatedly and rapidly optimizing the shape of the building aiming at the energy-saving target.
It should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (10)
1. A basic arithmetic unit comprising solar radiation, characterized by: the basic operation Unit is one of 5 basic types, and the 5 basic types are unitsground-ceiling、Unitground-roof、Unitfloor-ceiling、Unitfloor-roofAnd Unitexposed-ceiling(ii) a Wherein Unitground-ceilingThe lower surface of the basic operation unit is a terrace surface which generates heat exchange with the ground, and the upper surface is an interlayer ceiling; unitground-roofThe lower surface of the basic operation unit is a terrace surface which generates heat exchange with the ground, and the upper surface is a roof surface exposed to the external environment; unitfloor-ceilingThe lower surface of the basic operation unit is a floor surface between layers, and the upper surface is a ceiling surface between layers; unitfloor-roofThe lower surface of the basic operation unit is a floor surface between floors, and the upper surface is a roof surface exposed to the external environment; unitexposed-ceilingThe lower surface of the basic operation unit is an overhead layer surface exposed to the external environment, and the upper surface is an interlayer floor surface; the basic operation unit comprises geometric parameters and solar radiation quantity parameters.
2. A basic arithmetic unit containing solar radiation according to claim 1, characterized in that: the basic operation unit is a cuboid, and the geometric parameters of the basic operation unit comprise the length of the cuboidX, width Y, height Z; taking the plane where the length and the height of the cuboid are located as a reference plane, wherein the azimuth angle between the reference plane and the local north direction is Rotate; the ratio of the inner wall part of each surface of the basic operation unit to the total area of the surface is AdiabaliciThe ratio of the window to the wall of the outer wall part on each side of the basic operation unit is WWRiI.e. the ratio of the window area to the exterior wall area, where i denotes each side of the basic arithmetic unit, i is 1 to 4.
3. A basic arithmetic unit containing solar radiation according to claim 2, characterized in that: the solar radiation quantity parameter comprises solar radiation quantity Rad _ wall received by any side outer wall of the basic operation unitiAnd the solar radiation amount Rad _ win received by the outer window on the side surfacei。
4. A basic arithmetic unit containing solar radiation according to claim 3, characterized in that: solar radiation Rad _ wall received by outer walliIncluding Rad _ walli(low)、Rad_walli(mid) and Rad _ walli(high), wherein Rad _ walli(low) is the amount of solar radiation received by the side surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ walli(mid) is the amount of solar radiation received by the side when the outdoor air temperature is in the indoor set temperature zone; rad _ walli(high) is the amount of solar radiation received by the side when the outdoor air temperature is higher than the indoor set temperature; the Rad _ winiIncluding Rad _ wini(low)、Rad_wini(mid) and Rad _ wini(high), wherein Rad _ wini(low) is the solar radiation quantity received by the external window on the side surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ wini(mid) is the amount of solar radiation received by the external window on the side when the outdoor air temperature is in the indoor set temperature zone; rad _ wini(high) is the amount of solar radiation received by the exterior window on that side when the outdoor air temperature is higher than the indoor set temperature.
5. Root of herbaceous plantA basic arithmetic unit containing solar radiation according to any one of claims 1 to 4, characterized in that: the basic operation Unit is a Unitground-roofOr Unitfloor-roofIn time, the solar dosimetry parameters further comprise top surface solar radiation Rad _ roof, which includes Rad _ roof (low), Rad _ roof (mid), and Rad _ roof (high); wherein, Rad _ roof (low) is the solar radiation quantity received by the top surface when the outdoor air temperature is lower than the lowest indoor set temperature; rad _ roof (mid) is the amount of solar radiation received by the top surface when the outdoor air temperature is in the indoor set temperature interval; rad _ roof (high) is the amount of solar radiation received by the ceiling when the outdoor air temperature is above the indoor set temperature.
6. A building energy consumption rapid simulation method is characterized in that: comprising using the parameters of the basic arithmetic unit of any of claims 1 to 5 for machine learning algorithm generation and building energy consumption fast simulation feedback.
7. The building energy consumption rapid simulation method according to claim 6, characterized in that: the machine learning algorithm generation comprises training sample generation, machine learning algorithm training and optimal machine learning algorithm screening;
the training sample is generated by combining the parameters of the basic operation units and randomly supercube sampling, the basic operation units with different geometric parameters and solar radiation quantity parameters are generated in batches by adopting experimental design and are used as the training samples, and the basic operation units are led into energy consumption simulation software to carry out accurate physical simulation so as to obtain energy consumption data of the basic operation units;
the machine learning algorithm training includes two steps:
a: because the selection of the input parameters greatly influences the prediction precision of the machine learning algorithm, the input parameters with the best prediction effect need to be explored, and j groups of input parameters are obtained through a parameter transformation formula;
b: dividing the j training sample combinations after the parameter transformation of the previous step into training sets D according to the proportion0Test set D1And testing ofCollection D2And will train set D0Respectively inputting the training data into k machine learning algorithms, setting different hyper-parameters p for each machine learning algorithm, and obtaining j x k groups in total, wherein each group has a plurality of machine learning models with hyper-parameter setting combinations
The optimal machine learning algorithm screening comprises the following steps:
s1: using various types of machine learning algorithm models
Computing test set D1Energy consumption value of, and test set D1Comparing the actual energy consumption to obtain the product of R2Accuracy value of representation
S2: using this accuracy value
Screening optimal hyper-parameter setting model in each type of machine learning algorithm
S3: then using the j x k group candidate machine learning algorithm to the test set D2The prediction accuracy of (2) is calculated to obtain a machine learning algorithm prediction model F with the highest prediction accuracy R2best。
8. The building energy consumption rapid simulation method according to claim 7, characterized in that: in the step a, the parameter information obtained from the basic operation unit can be converted into other associated parameters, and the parameter transformation formula is as follows:
Area＝X*Y
Area_wini+Area_walli＝(Z*Length)*(1-Adiabatici)
(Length＝X,if i＝1,3；Length＝Y,if i＝2,4)
Area_wini＝WWRi*(Area_wini+Area_walli)
Area_walli＝(1-WWRi)*(Area_wini+Area_walli)
Rad_wall_averagei＝Rad_walli/Area_walli；
Rad_win_averagei＝Rad_wini/Area_wini
Rad_roof_average＝Rad_roof/Area
wherein, Area _ winiRepresents the Area of the outer window on the ith side of the basic arithmetic unit, Area _ walliThe area of an outer wall on the ith side of the basic operation unit is shown; rad _ wall _ averageiRad _ win _ average, the amount of solar radiation received by any side wall of the basic arithmetic unitiRepresenting the amount of solar radiation received per unit area of the outer window on the side; rad _ roof _ average represents the amount of solar radiation received per unit area of the top surface; the j sets of input parameters include Configuration1, Configuration2, Configuration3, Configuration4, Configuration5, and Configuration 6; wherein Configuration1 includes X, Y, Z, Area, Rotate,
Rad_winiconfiguration2 includes X, Y, ZArea, Rotate, Adiabalici,Rad_walli,Rad_winiThe Configuration3 includes X, Y, Z, Area, Rotate, Adiabalici,WWRi,Rad_walli,Rad_winiThe Configuration4 includes X, Y, Z, Area, Rotate, Adiabalici,WWRi,Rad_wall_averagei,Rad_win_averageiThe Configuration5 includes X, Y, Z, Area, Rotate, Area _ walli,Area_wini,Rad_walli,Rad_winiThe Configuration6 includes X, Y, Z, Area, Rotate, Area _ walli,Area_wini,Rad_wall_averageiAnd Rad _ win _ averagei。
9. The building energy consumption rapid simulation method according to claim 8, characterized in that: the basic operation Unit is a Unitground-roofOr Unitfloor-roofThe configurations 1, 2, 3 and 5 further include Rad _ roof, the configurations 4 and 6 further include Rad _ roof _ operation, and Rad _ roof _ operation is solar radiation per unit area of the top surface.
10. The building energy consumption rapid simulation method according to claim 9, characterized in that: the quick simulation feedback of the building energy consumption comprises the following steps:
a: and (3) building decomposition: the building is decomposed into the building volume which is firstly divided into volume slices of different layers according to the actual building layer height, and the slices of each layer are further decomposed into the cuboid basic operation unit;
b: calculating the radiation value of the basic operation unit: using Accelard to carry out fast calculation on the sunshine radiation parameters accelerated by the GPU on the basic operation unit obtained in the step A, wherein the operation speed of the calculation algorithm is not in direct proportion to the fineness degree and the surface complexity of a calculation grid;
c: fast calculation of energy consumption: inputting the geometric parameters and solar radiation quantity parameters of the basic operation unit into a machine learning algorithm prediction model F in batchesbestObtaining the energy consumption value E of all basic operation units of the whole buildingidxThe total energy consumption value of the building is
Wherein n is the number of basic arithmetic units constituting the whole building;
d: fast energy consumption feedback: because the machine learning algorithm is close to the real-time energy consumption feedback speed, the calculated total energy consumption value can be quickly displayed in the building model through a visualization method; the energy consumption values of the basic operation units forming the building volume can be directly colored on the surface of each basic operation unit in a pseudo-color graph mode, and architects can conveniently and quickly find out areas with unfavorable building energy consumption to conduct targeted form optimization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010351293.3A CN111597609B (en) | 2020-04-28 | 2020-04-28 | Basic operation unit containing solar radiation and building energy consumption rapid simulation method applying same |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010351293.3A CN111597609B (en) | 2020-04-28 | 2020-04-28 | Basic operation unit containing solar radiation and building energy consumption rapid simulation method applying same |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111597609A true CN111597609A (en) | 2020-08-28 |
| CN111597609B CN111597609B (en) | 2023-03-24 |
Family
ID=72189470
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010351293.3A Active CN111597609B (en) | 2020-04-28 | 2020-04-28 | Basic operation unit containing solar radiation and building energy consumption rapid simulation method applying same |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111597609B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112270137A (en) * | 2020-11-24 | 2021-01-26 | 同济大学 | Method and device for dynamically generating sunshade behavior based on building energy consumption simulation software |
| CN112801357A (en) * | 2021-01-21 | 2021-05-14 | 长江慧控科技(武汉)有限公司 | Sunshine radiation amount prediction method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4425983B1 (en) * | 2009-03-18 | 2010-03-03 | 株式会社パスコ | Method and apparatus for evaluating solar radiation |
| WO2012152939A1 (en) * | 2011-05-11 | 2012-11-15 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | Method of predicting the energy consumption of a building |
| CN104133938A (en) * | 2014-07-01 | 2014-11-05 | 哈尔滨工业大学 | Severe cold area office building external window geometric parameter optimization method based on heating energy consumption consideration |
| CN109033595A (en) * | 2018-07-16 | 2018-12-18 | 清华大学 | Build Green performance sunykatuib analysis new method and device |
| CN109255472A (en) * | 2018-08-20 | 2019-01-22 | 天津科技大学 | A kind of Office Building Energy Consumption prediction technique based on Multivariate adaptive regression splines batten |
| CN110046743A (en) * | 2019-03-06 | 2019-07-23 | 上海交通大学 | Energy Consumption of Public Buildings prediction technique and system based on GA-ANN |
-
2020
- 2020-04-28 CN CN202010351293.3A patent/CN111597609B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4425983B1 (en) * | 2009-03-18 | 2010-03-03 | 株式会社パスコ | Method and apparatus for evaluating solar radiation |
| WO2012152939A1 (en) * | 2011-05-11 | 2012-11-15 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | Method of predicting the energy consumption of a building |
| CN104133938A (en) * | 2014-07-01 | 2014-11-05 | 哈尔滨工业大学 | Severe cold area office building external window geometric parameter optimization method based on heating energy consumption consideration |
| CN109033595A (en) * | 2018-07-16 | 2018-12-18 | 清华大学 | Build Green performance sunykatuib analysis new method and device |
| CN109255472A (en) * | 2018-08-20 | 2019-01-22 | 天津科技大学 | A kind of Office Building Energy Consumption prediction technique based on Multivariate adaptive regression splines batten |
| CN110046743A (en) * | 2019-03-06 | 2019-07-23 | 上海交通大学 | Energy Consumption of Public Buildings prediction technique and system based on GA-ANN |
Non-Patent Citations (4)
| Title |
|---|
| 张崎等: "建筑能耗模拟中太阳直射辐射的处理方法对比研究", 《施工技术》 * |
| 朱姝妍 等: ""太阳能热水系统在医疗建筑中的设计应用"", 《建筑节能》 * |
| 李剑琨等: "基于标准气象日的南墙太阳辐射净得热总量预测模型研究", 《四川理工学院学报(自然科学版)》 * |
| 田玮等: "基于机器学习的建筑能耗模型适用性研究", 《天津科技大学学报》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112270137A (en) * | 2020-11-24 | 2021-01-26 | 同济大学 | Method and device for dynamically generating sunshade behavior based on building energy consumption simulation software |
| CN112270137B (en) * | 2020-11-24 | 2022-06-17 | 同济大学 | Method and device for dynamically generating sunshade behavior based on building energy consumption simulation software |
| CN112801357A (en) * | 2021-01-21 | 2021-05-14 | 长江慧控科技(武汉)有限公司 | Sunshine radiation amount prediction method, device, equipment and storage medium |
| CN112801357B (en) * | 2021-01-21 | 2024-04-02 | 长江慧控科技(武汉)有限公司 | Solar radiation quantity prediction method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111597609B (en) | 2023-03-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | An ANN-based fast building energy consumption prediction method for complex architectural form at the early design stage | |
| Yang et al. | Influence of urban morphological characteristics on thermal environment | |
| Nazarian et al. | Predicting outdoor thermal comfort in urban environments: A 3D numerical model for standard effective temperature | |
| Wang et al. | Multi-objective optimization (MOO) for high-rise residential buildings’ layout centered on daylight, visual, and outdoor thermal metrics in China | |
| CN111597609B (en) | Basic operation unit containing solar radiation and building energy consumption rapid simulation method applying same | |
| CN108399470B (en) | Indoor PM2.5 prediction method based on multi-example genetic neural network | |
| CN111310257A (en) | Regional building energy consumption prediction method under BIM environment | |
| You et al. | Improving building facade design using integrated simulation of daylighting, thermal performance and natural ventilation | |
| He et al. | Modeling solar radiation on a human body indoors by a novel mathematical model | |
| Chen et al. | A rapid evaluation method for design strategies of high-rise office buildings achieving nearly zero energy in Guangzhou | |
| Chen et al. | Numerical simulations on atmospheric stability conditions and urban airflow at five climate zones in China | |
| Gholami et al. | A modelling methodology for a solar energy-efficient neighbourhood | |
| CN110334366A (en) | A kind of instantaneous cooling load prediction method of building based on the Monte Carlo Method with Latin Hypercube Sampling | |
| Liu et al. | Building information modelling-enabled multi-objective optimization for energy consumption parametric analysis in green buildings design using hybrid machine learning algorithms | |
| JP2005301487A (en) | Thermal environment prediction method for inside/outside of building | |
| Peters | Defining environments: Understanding architectural performance through modelling, simulation and visualisation | |
| Tayari et al. | Investigating DesignBuilder Simulation Software's Validation in Term of Heat Gain through Field Measured Data of Adjacent Rooms of Courtyard House | |
| Yu et al. | Climatic formations: Evolutionary dynamics of urban morphologies | |
| Kalogeras et al. | Verification and validation of a simulation model for energy use in buildings | |
| Bo et al. | Reconstruction Design of Existing Residential Buildings Based on 3D Simulation Method | |
| Yao et al. | Urban renewal based wind environment at pedestrian level in high-density and high-rise urban areas in Sai Ying Pun, Hong Kong | |
| Chen et al. | Optimization and prediction of energy consumption, light and thermal comfort in teaching building atriums using NSGA-II and machine learning | |
| BEHZADFAR et al. | Generic flows of sustainable urban form | |
| Davis | Development and validation of trombe walls in autodesk simulation CFDTM | |
| Baydoun et al. | EVALUATING THE EFFECT OF MASHRABIYA AND PERFORATED MALAY CARVING WINDOW ON THE INDOOR NATURAL VENTILATION PERFORMANCE OF LIVING SPACE |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |