A method for analyzing heat demand in severe cold regions based on building digitization technology

Abstract

The application relates to a heat demand analysis method based on building digitization technology in a severe cold area and relates to the technical field of heat supply analysis. The application is aimed at solving the problem of low analysis accuracy of the existing heat demand analysis method. The application comprises the following steps: obtaining outdoor meteorological data at different time points; constructing an indoor BIM model to be analyzed; setting the design temperature, personnel density and equipment use condition of each functional area in the indoor model to be analyzed; obtaining the required water supply flow of each functional area in the indoor model to be analyzed; obtaining the valve opening degree corresponding to the required water supply flow by using the required water supply flow of each functional area in the indoor model to be analyzed, adjusting the valve opening degree of the indoor model to be analyzed according to the valve opening degree, obtaining the temperature of each functional area in the indoor model to be analyzed by using a room temperature sensor, replacing the abnormal value with a correction value, obtaining the temperature of each functional area in the indoor model to be analyzed after processing, and adjusting and controlling the valve opening degree again based on the temperature of each functional area in the indoor model to be analyzed after processing. The application is used for heat demand analysis and heat regulation.

Description

Technical Field

[0001] This invention relates to the field of heating analysis technology, and in particular to a method for analyzing heat demand in frigid regions based on building digitalization technology. Background Technology

[0002] With the national strategic goal of "peak carbon and carbon neutrality" and the increasing emphasis on environmental protection, building energy conservation has become an important component of the country's low-carbon transformation. In recent years, with continuous urbanization and rising expectations for indoor comfort, the proportion of total building energy consumption in China has reached 45.5% of the country's total energy consumption. Furthermore, the long winters in frigid regions, such as Heilongjiang Province with a heating season lasting 180-210 days, lead to an increased proportion of heating energy consumption, with heating energy consumption accounting for over 60% of total building energy consumption in these areas. Therefore, reducing heating energy consumption in frigid regions and minimizing energy waste is crucial for achieving the national "dual carbon" goals and promoting sustainable development. Consequently, the analysis of heat demand has become a key focus of research in this field.

[0003] Traditional methods for analyzing the thermal demand of heating systems employ static analysis, which relies on empirical formulas, simplified physical models, and past operational experience for energy consumption analysis and control. However, these methods fail to consider the impact of dynamic factors such as actual indoor temperature and outdoor meteorological data (outdoor temperature, outdoor wind speed, and sunlight) on thermal demand, resulting in low accuracy in thermal demand analysis. Summary of the Invention

[0004] The purpose of this invention is to address the problem of low accuracy in existing thermal demand analysis methods, and to propose a thermal demand analysis method for frigid regions based on building digitization technology.

[0005] A method for analyzing thermal demand in frigid regions based on building digitization technology includes:

[0006] Step 1: Obtain outdoor meteorological data to be analyzed at different times in the study area;

[0007] The outdoor meteorological data to be analyzed includes: outdoor temperature, outdoor pressure, and outdoor air density;

[0008] Step 2: Obtain the heat transfer coefficient of the building envelope in the room to be analyzed, and construct a BIM model of the building containing the room using the heat transfer coefficient of the building envelope, building shape, building structure drawing and building drawing of the room to be analyzed;

[0009] Step 3: Set the design temperature, personnel density, and equipment usage for each functional area in the room to be analyzed;

[0010] The functional areas include: a conference room, a tea room, and restrooms;

[0011] Step 4: Using the BIM model of the building where the room to be analyzed is located, conduct thermal demand analysis based on the outdoor meteorological data to be analyzed and the design temperature, personnel density and equipment usage of each functional area of ​​the room to be analyzed, to obtain the required water supply flow rate for each functional area of ​​the room to be analyzed.

[0012] Step 5: Using the required water supply flow rate of each functional area in the room to be analyzed obtained in Step 4, obtain the valve opening corresponding to the required water supply flow rate, and adjust the valve opening of the room to be analyzed according to the valve opening. Then, use the room temperature sensor to obtain the temperature of each functional area in the room to be analyzed at different times, and then obtain the abnormal values ​​in the temperature of each functional area in the room to be analyzed at different times, and replace the abnormal values ​​with correction values ​​to obtain the processed temperature of each functional area in the room to be analyzed at different times. Based on the processed temperature of each functional area in the room to be analyzed at different times, adjust the valve opening again.

[0013] Furthermore, in step four, the BIM model of the building containing the room to be analyzed is used to conduct a thermal demand analysis based on the outdoor meteorological data and the design temperature, personnel density, and equipment usage of each functional area within the room to be analyzed, in order to obtain the required water supply flow rate for each functional area within the room to be analyzed. Specifically:

[0014] Step 4.1: Obtain the heat load of the i-th functional zone in the room to be analyzed at the j-th heating time. Specifically:

[0015] Q ij =Q1 ij +Q2 ij +Q3 ij -Q4 ij

[0016] Q1 ij =αKA(t) ij -t w j )

[0017] Q2 ij =0.278NVc p ρ w (t ij -t w j )

[0018] Q3 ij =Q1 ij (1+β c +β f +β q +β m (1+β) fg (1+β) j )

[0019] Q4 ij=η(zq) r +W)

[0020] Among them, Q ij Q1 represents the heat load of the i-th functional zone in the room at the j-th heating time, where i ∈ [1, m], i is the functional zone number to be evaluated, m is the total number of functional zones to be evaluated, and j ∈ [1, n], j is the heating time number, and n is the total number of heating times. ij Q2 is the basic heat loss of the building envelope in the i-th functional zone at the j-th heating moment. ij Q3 is the heat loss due to cold air infiltration in the i-th functional zone at the j-th heating moment. ij Q4 represents the additional heat loss within the building at the j-th heating time in the i-th functional zone. ij This represents the total heat dissipation from people and equipment in the room to be analyzed at the j-th heating moment in the i-th functional zone. α is the temperature difference correction coefficient, K represents the heat transfer coefficient of the building envelope, A represents the heat transfer area of ​​the building envelope, and t... ij t represents the design temperature of the i-th functional zone in the room at the j-th heating time. w j Let N represent the average outdoor temperature at the j-th heating moment, N represent the indoor air exchange rate, V represent the net room volume, and c represent the average outdoor temperature at the j-th heating moment. p It is the specific heat capacity of outdoor air at constant pressure, ρ w β represents outdoor air density. c Indicates the preset orientation correction rate, β f Indicates the preset wind correction rate, β q This indicates the preset correction rate for the two exterior walls, β. m This indicates the correction rate for an excessively large window-to-wall ratio, β. fg This represents the preset room height correction rate, β. j η represents the preset intermittent correction rate, z represents the preset correction rate for heat dissipation by people and equipment, and q represents the number of people in the room to be analyzed. r W represents the heat dissipation of the human body, and W represents the heat dissipation of the equipment.

[0021] Step 4.2: Obtain the required water supply flow rate F for the i-th functional zone of the room under analysis at the j-th heating time using the heat load at the j-th heating time. ij .

[0022] Furthermore, in step four-two, the required water supply flow rate F for the i-th functional area of ​​the room under analysis at the j-th heating time is obtained by utilizing the heat load of the i-th functional area of ​​the room under analysis at the j-th heating time. ij Specifically:

[0023] F ij =Q ij / (cρΔT)

[0024] Where c represents the specific heat capacity of water, ρ represents the density of water, ΔT represents the temperature difference between the supply and return water, and F ij It is the water supply flow required for the i-th functional area of ​​the room at the j-th heating moment.

[0025] Furthermore, in step five, the valve opening corresponding to the required water supply flow rate for each functional area in the room to be analyzed, obtained from step four, is specifically as follows:

[0026] If the regulating valve has a quick-opening flow characteristic, the required water supply flow rate corresponding to the valve opening can be obtained in the following way:

[0027]

[0028] Among them, F max L represents the valve's maximum flow rate. ij L represents the required valve opening degree for the i-th functional zone of the room at the j-th heating time. max R represents the maximum valve opening, and R represents the predictive control valve turn-off ratio.

[0029] If the regulating valve has a linear flow characteristic, the required water supply flow rate corresponding to the valve opening can be obtained in the following way:

[0030]

[0031] If the regulating valve has an equal percentage flow characteristic, the required water supply flow rate corresponding to the valve opening can be obtained in the following way:

[0032]

[0033] Further, in step five, the temperature of each functional area in the room to be analyzed is obtained at different times using a room temperature sensor. Then, outliers in the temperatures of each functional area in the room to be analyzed at different times are acquired, and the outliers are replaced with correction values ​​to obtain the processed temperatures of each functional area in the room to be analyzed at different times. Specifically:

[0034] First, obtain the historical sequence of M historical heating times for the i-th functional area of ​​the room to be analyzed. Input the historical sequence of M historical heating times for the i-th functional area of ​​the room to be analyzed into the trained indoor temperature prediction model to obtain the indoor predicted temperature of the i-th functional area at the (M+1)-th heating time, i.e., the correction value.

[0035] The historical sequence of M historical heating times in the i-th functional area includes: the indoor temperature sequence of M historical heating times in the i-th functional area, the water supply temperature sequence of M historical heating times in the i-th functional area, the return water temperature sequence of M historical heating times in the i-th functional area, the flow rate sequence of M historical heating times in the i-th functional area, and the outdoor temperature sequence of M historical heating times in the i-th functional area.

[0036] Then, the predicted indoor temperature of the i-th functional zone at the (M+1)-th heating time is compared with the actual indoor temperature of the i-th functional zone at the (M+1)-th heating time obtained by the room temperature sensor. If the predicted indoor temperature and the actual indoor temperature are within the preset first error range, the valve opening is adjusted again based on the processed temperatures of each functional zone in the room to be analyzed at different times; otherwise, the actual indoor temperature of the i-th functional zone at the (M+1)-th heating time is replaced with the correction value, and the valve opening is adjusted again based on the processed temperatures of each functional zone in the room to be analyzed at different times.

[0037] Furthermore, the trained indoor temperature prediction model is obtained in the following way:

[0038] B1. Obtain the historical sequences of M historical heating times for the i-th functional area of ​​the room to be analyzed, and preprocess each historical sequence. Combine the preprocessed historical sequences with the indoor temperature of the i-th functional area of ​​the room to be analyzed at the (M+1)-th heating time to form a training set.

[0039] The historical sequence of M historical heating times in the i-th functional area includes: the indoor temperature sequence of M historical heating times in the i-th functional area, the water supply temperature sequence of M historical heating times in the i-th functional area, the return water temperature sequence of M historical heating times in the i-th functional area, the flow rate sequence of M historical heating times in the i-th functional area, and the outdoor temperature sequence of M historical heating times in the i-th functional area.

[0040] B2. Use the training set to train the indoor temperature prediction model and obtain a well-trained indoor temperature prediction model.

[0041] Furthermore, the preprocessing of each historical sequence in B1 specifically includes:

[0042] A1. Obtain the location of missing feature data in the m'th historical sequence of the i-th functional area, and then fill the missing feature data location with the average of the feature data values ​​of the previous and next positions.

[0043] Where m' is the historical sequence number, m' = 1, 2, 3, 4, 5 represent the indoor temperature sequence, the water supply temperature sequence, the return water temperature sequence, the flow rate sequence, and the outdoor temperature sequence, respectively;

[0044] A2. Obtain the first quartile Q1, the third quartile Q3, and the interquartile range IQR of the m'-th historical sequence of the i-th functional area after processing by A1.

[0045] The first quartile Q1 is the 25th quartile of the m'th historical sequence of the i-th functional region;

[0046] The third quartile, Q3, is the 75th quartile of the m'-th historical sequence in the i-th functional region;

[0047] The interquartile range (IQR) is obtained in the following way:

[0048] IQR = Q3 - Q1

[0049] A3. Use Q1, Q3, and IQR to obtain the upper and lower bounds of outliers:

[0050] Lower bound of outliers = Q1 - 1.5 × IQR

[0051] Upper bound of outliers = Q3 + 1.5 × IQR

[0052] A4. The feature data values ​​in the m'th historical sequence of the i-th functional area that are not within the range of the lower bound to the upper bound of the outlier are the outliers. Then, the position of the outlier is obtained, and the average of the three feature data values ​​before the outlier position and the three feature data values ​​after the outlier position is replaced in the outlier position to obtain the m'th historical sequence of the i-th functional area after the outlier is filled.

[0053] A5. Normalize the m'-th historical sequence of the i-th functional area after outlier filling to obtain the indoor temperature sequence, supply water temperature sequence, return water temperature sequence, flow rate sequence, and outdoor temperature sequence of the M historical heating times of the i-th functional area after preprocessing.

[0054] Furthermore, the normalization process performed on the m'-th historical sequence of the i-th functional area after outlier imputation in A5 is specifically as follows:

[0055]

[0056] Where, x i,m',d It is the d-th feature data value in the m'-th historical sequence of the i-th functional area after outlier imputation, x′ i,m',d It is the d-th feature data value in the m'-th historical sequence of the i-th functional region after normalization, x i,m',min It is the minimum feature data in the m'-th historical sequence of the i-th functional area after outlier imputation, x i,m',max It is the maximum value of the feature data in the m'th historical sequence of the i-th functional area after outlier imputation.

[0057] Furthermore, the indoor temperature prediction model in B2 includes: an LSTM module, an efficient additive self-attention mechanism module, and an indoor temperature prediction module;

[0058] The LSTM module is an LSTM network used to obtain the hidden state of the input feature vector, specifically:

[0059] h j' =O j' tanh(C j' )

[0060] O j' =σ(W o [h j'-1 ,x j' ]+b o )

[0061]

[0062] f j' =σ(W f [h j'-1 ,x j' ]+b f )

[0063] i j' =σ(W i” [h j'-1 ,x j' ]+b i” )

[0064] Among them, f j' σ is the output of the forget gate at time step j', σ is the sigmoid activation function, and W is the output of the forget gate at time step j'. f It is the forget gate weight matrix, W i” It is the input gate weight matrix, W c W is the cell state update weight matrix. o It is the output gate weight matrix, h j'-1 It is the hidden state of the previous time step j'-1, x j' b is the input feature vector at the current time step j'. f It is the forgetting gate bias term, b i” It is the input gate bias term, b c It is the cell state update bias term, b o It is the output gate bias term, i j' It is the output of time step j' of the input gate. It represents the candidate cell state, tanh is the hyperbolic tangent activation function, and C... j' Updated cell state, C j'-1 This refers to the cell state at the previous time step j'-1, O j' It is the output of the output gate at time step j', h j' It is the hidden state of time step j' at that time;

[0065] The input feature vector includes: supply water temperature, return water temperature, flow rate, indoor temperature, and outdoor temperature;

[0066] The efficient additive self-attention mechanism module obtains the global context representation and c' based on the hidden state of the input feature vector:

[0067]

[0068] x' j' =Q j' +T(K j' ×q j' )

[0069] q j' =α j' Q j'

[0070]

[0071] Q j' =h j' W Q

[0072] K j' =h j' W K

[0073] Where, x' j' It is the global context representation at the j'-th time step, α j' It is the global attention query vector, d is the dimension of the query matrix, and W is the global attention query vector. Q W K It is the weight matrix for learning, Q j' It is the query vector, K j' It is the key vector, W a It is a preset parameter vector, q j' It is a single global query vector, and T is a linear transformation;

[0074] The indoor temperature prediction module obtains the predicted indoor temperature value based on the hidden state of the input feature vector and the global context representation:

[0075] y′=W'z'+b

[0076] z' = concat[h j' ,c']

[0077] Where W' is the preset weight matrix, b is the bias term, and z' is h j' The concatenated vector of c' and y' is the predicted indoor temperature value.

[0078] Furthermore, in step five, the valve opening is adjusted again based on the temperature of each functional area in the room to be analyzed at different times after processing. Specifically:

[0079] Determine whether the error between the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing and the design temperature of the i-th functional zone in the room to be analyzed at the j-th heating time is within the preset allowable second temperature error range. If it is within the preset allowable second temperature error range, the valve opening is not adjusted. If it is not within the preset allowable second temperature error range, the valve opening is adjusted.

[0080] If the temperature is outside the preset second allowable error range, the valve opening will be adjusted as follows:

[0081] Find the outdoor temperature data P that is closest to the j-th heating time in the historical data, and obtain the valve opening corresponding to the outdoor temperature data P;

[0082] If the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing is higher than the design temperature of the i-th functional zone in the room to be analyzed at the corresponding j-th heating time, then the minimum value of the valve opening corresponding to the outdoor temperature data P will be taken as the current valve opening.

[0083] If the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing is lower than the design temperature of the i-th functional zone in the room to be analyzed at the corresponding j-th heating time, then the maximum value of the valve opening corresponding to the outdoor temperature data P will be taken as the current valve opening.

[0084] The beneficial effects of this invention are as follows:

[0085] This invention proposes a method for analyzing the thermal demand of building heating systems in frigid regions, enabling energy-saving control. By acquiring indoor temperature, outdoor meteorological data, and various dynamic influencing factors, as well as the heat transfer coefficient of the building envelope, this invention achieves precise thermal and flow demand analysis for building heating systems in frigid regions. This effectively improves the accuracy of thermal demand analysis, thereby reducing building energy consumption. Based on the required indoor temperature, actual indoor temperature, and outdoor meteorological data, this invention determines the opening degree of the heating water supply valves. After the first adjustment, it checks whether the actual indoor temperature meets the preset temperature, and then performs a second adjustment of the heating water supply valves to improve the accuracy of indoor heating temperature control. This reduces energy consumption while meeting heating needs. Furthermore, this invention provides real-time feedback on the actual indoor temperature after adjustment and performs a second adjustment of the water supply flow rate, further improving the accuracy of thermal demand analysis. Attached Figure Description

[0086] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0087] Specific implementation method one: as follows Figure 1 As shown in the figure, the specific process of the thermal demand analysis method for severely cold regions based on building digitalization technology in this embodiment is as follows:

[0088] Step 1: Obtain outdoor meteorological data to be analyzed at different times in the frigid region;

[0089] The outdoor meteorological data to be analyzed includes: outdoor temperature, outdoor pressure, and outdoor air density;

[0090] Small meteorological monitoring instruments are installed next to the building to capture real-time outdoor meteorological data. This method is more accurate than obtaining dynamic outdoor meteorological data for the building's location from the National Meteorological Administration's website, leading to more precise subsequent thermal demand analysis.

[0091] The extremely cold regions meet the following conditions:

[0092] Condition 1: The number of days with an average daily temperature below 5℃ is greater than 145.

[0093] The daily average temperature is the average of the highest and lowest temperatures of the day.

[0094] Condition 2: The lowest monthly average temperature of the current year must be less than -10℃, specifically:

[0095] Get the average daily temperature of the current region, sum the daily average temperatures and divide by the number of days in the month to get the average monthly temperature; get the lowest monthly average temperature of the current year, and make sure the lowest monthly average temperature of the current year is less than -10℃.

[0096] Step 2: Obtain the heat transfer coefficient of the building envelope within the room to be analyzed. Using the heat transfer coefficient, building shape, structural drawings, and architectural drawings of the building envelope within the room to be analyzed, construct a BIM model of the building containing the room. Specifically:

[0097] Step 21: Obtain the building pattern of the room to be evaluated. If the building pattern of the room to be evaluated is a drawing, proceed to Step 22; if the building pattern of the room to be evaluated is a renovation, proceed to Step 23.

[0098] The construction modes include: drawing mode and renovation mode;

[0099] The drawing mode is as follows: the types of materials, material thicknesses, and parameters of each material used in BIM digital modeling are all set according to the design drawings. This mode is applicable to buildings constructed according to the design drawings or buildings that have been renovated and whose parameters after the renovation are known.

[0100] The aforementioned renovation mode is as follows: the renovation mode indicates that after a building is designed and constructed according to the drawings, it is intended to be renovated. This mode can be used to analyze the optimal parameters of the building envelope after renovation, provide guidance for the renovation of old residential areas, and set the material types, material thicknesses and parameters according to the analysis results, which will be used for subsequent dynamic thermal demand analysis.

[0101] Step 22: Obtain the types and thicknesses of the interior materials to be evaluated based on the design drawings. Use these materials to obtain the heat transfer coefficient of the building envelope. Construct a BIM model of the building containing the interior to be analyzed using the building shape, the heat transfer coefficient of the building envelope, the building structure drawings, and the building drawings.

[0102] The BIM model includes detailed information such as building form and materials;

[0103] The same types and thicknesses of materials are used in the same functional areas of the interior, and the same heat transfer coefficients are used.

[0104] Steps 2 and 3: Obtain the heat transfer coefficient of the building envelope by using the types and thicknesses of the materials to be evaluated in the interior after the renovation; and construct a BIM model using the heat transfer coefficient of the building envelope, the building shape, the building structure drawings, and the building drawings.

[0105] The types and thicknesses of materials used in the renovated interior are set based on experience, generally according to cost.

[0106] Step 3: Set the design temperature for each functional area of ​​the room to be analyzed according to the heating operation mode. Simultaneously, set the personnel density and equipment usage in the room. Specifically:

[0107] Step 31: Obtain the heating mode of the room to be analyzed. If the heating mode is the normal mode, proceed to Step 32; if the heating mode is the manual mode, proceed to Step 33.

[0108] The standard mode is as follows: the heating temperature of the room to be analyzed remains unchanged at the initial set value, and the heating time is all-day heating;

[0109] In manual mode, the heating temperature and heating period of each functional area in the room to be analyzed are set by the user and can be changed at any time;

[0110] Step 3.2: Set the preset heating design temperature for each functional area in the room to be analyzed.

[0111] In standard mode, the office area is set to 20 degrees Celsius, the meeting room to 20 degrees Celsius, the tea room to 18 degrees Celsius, and the restrooms to 18 degrees Celsius. The indoor personnel density and equipment usage are also set.

[0112] Step 33: Set the heating design temperature for each functional area of ​​the room to be analyzed at each time period:

[0113] The functional areas include: office area, meeting room, tea room, restrooms, etc.

[0114] Examples of heating design temperatures for different time periods in different functional areas are as follows:

[0115] In manual mode, the heating time can be set according to the working hours. The heating time for the office area is from 8:00 to 18:00, and the heating temperature is 20 degrees Celsius and 12 degrees Celsius from 18:00 to 8:00 respectively.

[0116] The heating schedule for the conference room can be set after the meeting is scheduled, with a heating temperature of 20 degrees Celsius.

[0117] The tea room is heated from 8:00 to 18:00 at a temperature of 18 degrees Celsius, and from 18:00 to 8:00 at a temperature of 12 degrees Celsius.

[0118] The bathroom is heated from 8:00 to 18:00 at a temperature of 18 degrees Celsius, and from 18:00 to 8:00 at a temperature of 12 degrees Celsius.

[0119] Steps three and four: Set the indoor personnel density and equipment usage to be evaluated. In case of special circumstances, temporary settings can be made, such as changing the time to the same day or permanently.

[0120] The device is an electrical appliance, such as a lighting device.

[0121] This invention provides users with a drawing mode and a renovation mode to choose from during the BIM model creation process. The renovation mode offers the optimal renovation plan for old residential areas and is divided into two heating modes: conventional mode and manual mode. Users can select the heating period and heating temperature to meet the different needs of various users.

[0122] Step 4: Using the BIM model of the building containing the room to be analyzed, conduct a thermal demand analysis based on the outdoor meteorological data and the design temperatures of each functional area within the room to obtain the required water supply flow rate for each functional area. Specifically:

[0123] Step 4.1 Obtain the heat load of the i-th functional zone in the room to be analyzed at the j-th heating time:

[0124] Q ij =Q1 ij +Q2 ij +Q3 ij -Q4 ij

[0125] Q1 ij =αKA(t) ij -t w j )

[0126] Q2 ij =0.278NVc p ρ w (t ij-t w j )

[0127] Q3 ij =Q1 ij (1+β c +β f +β q +β m (1+β) fg (1+β) j )

[0128] Q4 ij =η(zq) r +W)

[0129] Among them, Q ij Q1 represents the heat load of the i-th functional zone in the room at the j-th heating time, where i ∈ [1, m], i is the functional zone number to be evaluated, m is the total number of functional zones to be evaluated, and j ∈ [1, n], j is the heating time number, and n is the total number of heating times. ij Q2 is the basic heat loss of the building envelope in the i-th functional zone at the j-th heating moment. ij Q3 is the heat loss due to cold air infiltration in the i-th functional zone at the j-th heating moment. ij Q4 represents the additional heat loss within the building at the j-th heating time in the i-th functional zone. ij This represents the total heat dissipation from people and equipment in the room to be analyzed at the j-th heating moment in the i-th functional zone. α is the temperature difference correction coefficient, K represents the heat transfer coefficient of the building envelope, A represents the heat transfer area of ​​the building envelope, and t... ij t represents the design temperature of the i-th functional zone in the room at the j-th heating time. w j Let N represent the average outdoor temperature at the j-th heating moment, N represent the indoor air exchange rate, V represent the net room volume, and c represent the average outdoor temperature at the j-th heating moment. p It is the specific heat capacity of outdoor air at constant pressure, ρ w β represents outdoor air density. c Indicates the preset orientation correction rate, β f Indicates the preset wind correction rate, β q This indicates the preset correction rate for the two exterior walls, β. m This indicates the correction rate for an excessively large window-to-wall ratio, β. fg This represents the preset room height correction rate, β. j η represents the preset intermittent correction rate, z represents the preset correction rate for heat dissipation by people and equipment, and q represents the number of people indoors. r W represents the heat dissipation of the human body, and W represents the heat dissipation of the equipment.

[0130] The correction rate for the heat dissipation of the people and equipment was obtained from the "Code for Design of Heating, Ventilation and Air Conditioning of Civil Buildings";

[0131] The correction factor β for the indoor orientation (north, northeast, or northwest) to be evaluated c The value is between 0% and 10%; the correction rate β for the east or west orientation of the room to be evaluated. c Take -5%, the correction factor β for the interior orientation (southeast or southwest) to be evaluated. c The value is taken to be between -10% and -15%; the correction factor β for the south-facing orientation of the room to be evaluated. c Take a value between -15% and -30%;

[0132] The buildings to be evaluated are those located at unsheltered elevations, along rivers, coastlines, or in open fields, as well as buildings in towns that are significantly taller than surrounding structures. The wind correction factor β is used. f The value range is between 5% and 10%;

[0133] Correction rate β for two exterior walls q Set to 5%;

[0134] Window-to-wall ratio too large, correction factor β m Set to 10%;

[0135] Room height correction rate β fg The value is set to 2(H-4)%, and β fg When it exceeds 15%, β fg Set to 15%;

[0136] When the room to be evaluated is only used during the day, the intermittent correction rate β j Set to 20%; when the room to be evaluated is not frequently used, the intermittent correction rate β is used. j Set to 30%;

[0137] Indoor spaces to be evaluated that are used only during the day include: offices, schools, and shops;

[0138] Infrequently used interior spaces to be evaluated are those used only for events, meetings, or certain seasons.

[0139] Indoor occupancy density and equipment usage both affect heating energy consumption. Therefore, when calculating the heating load, occupancy and equipment are taken into account, and their heat generation is denoted as Q4. ij The total heat dissipation of the indoor occupants is recorded as zq. r The total heat dissipation of indoor equipment is denoted as W. η is used to correct the heat dissipation of personnel and equipment to ensure a more accurate calculation of the heat load.

[0140] Step 4.2: Obtain the required water supply flow rate F for the i-th functional zone of the room under analysis at the j-th heating time using the heat load at the j-th heating time. ij :

[0141] F ij =Q ij / (cρΔT)

[0142] Where c represents the specific heat capacity of water, ρ represents the density of water, and ΔT represents the temperature difference between the supply and return water.

[0143] This invention acquires various dynamic influencing factors such as indoor temperature and outdoor meteorological data, as well as the heat transfer coefficient of the building envelope, to achieve accurate thermal demand analysis and flow demand analysis of building heating systems in frigid regions, thereby effectively reducing building energy consumption.

[0144] Step 5: Using the required water supply flow rate for each functional area in the analysis chamber obtained in Step 4, determine the valve opening corresponding to the required water supply flow rate. Adjust the valve opening in the analysis chamber according to the valve opening. Then, use a room temperature sensor to obtain the temperature of each functional area in the analysis chamber at different times. Obtain any outliers in the temperature of each functional area in the analysis chamber at different times, and replace the outliers with correction values ​​to obtain the processed temperature of each functional area in the analysis chamber at different times. Based on the processed temperature of each functional area in the analysis chamber at different times, further adjust the valve opening, specifically as follows:

[0145] Step 51: Using the required water supply flow rate for each functional area in the room to be analyzed obtained in Step 4, determine the corresponding valve opening for the required water supply flow rate:

[0146] If the control valve has a quick-opening flow characteristic, the valve opening L is calculated using the required water supply flow rate for each functional area in the room to be analyzed and the formula for the quick-opening flow characteristic opening. Specifically:

[0147]

[0148] Among them, F max L represents the valve's maximum flow rate. ij L represents the required valve opening degree for the i-th functional zone of the room at the j-th heating time. max This indicates the maximum valve opening, and R indicates the control valve's adjustable ratio.

[0149] If the control valve has a linear flow characteristic, the valve opening is calculated using the required water supply flow rate for each functional area of ​​the room to be analyzed and the formula for the opening degree of the linear flow characteristic. Specifically:

[0150]

[0151] If the control valve exhibits a parabolic flow characteristic, the valve opening is calculated using the required water supply flow rate for each functional area of ​​the room under analysis and the formula for the parabolic flow characteristic opening. Specifically:

[0152]

[0153] If the regulating valve has an equal percentage flow characteristic, the valve opening is calculated using the required water supply flow rate for each functional area in the room to be analyzed and the equal percentage flow characteristic opening formula. Specifically:

[0154]

[0155] Step 52: Adjust the valve opening in the room to be analyzed according to the required water supply flow rate. Then, use the room temperature sensor to obtain the temperature of each functional area in the room to be analyzed at different times. Then, obtain the abnormal values ​​in the temperature of each functional area in the room to be analyzed at different times, and replace the abnormal values ​​with correction values ​​to obtain the processed temperature of each functional area in the room to be analyzed at different times.

[0156] The process of obtaining outlier values ​​in the temperatures of each functional area within the room to be analyzed at different times, replacing these outlier values ​​with corrected values, and obtaining the processed temperatures of each functional area within the room to be analyzed at different times is achieved through the following method:

[0157] Step 521: Obtain the historical sequence of M historical heating times for the i-th functional area of ​​the room to be analyzed. Input the historical sequence of M historical heating times for the i-th functional area of ​​the room to be analyzed into the trained indoor temperature prediction model to obtain the indoor predicted temperature of the i-th functional area at the (M+1)-th heating time, i.e., the correction value.

[0158] The historical sequence of M historical heating times for the i-th functional area includes: the indoor temperature sequence, supply water temperature sequence, return water temperature sequence, flow rate sequence, and outdoor temperature sequence for the i-th functional area at M historical times;

[0159] The trained indoor temperature prediction model is obtained through the following method:

[0160] B1. Obtain the historical sequences of M historical times for the i-th functional area in the room to be analyzed, and preprocess each historical sequence. Combine the preprocessed historical sequences and the historical temperature of the i-th functional area in the room to be analyzed at time M+1 to form a training set.

[0161] The preprocessing of each historical sequence specifically involves:

[0162] A1. Obtain the location of missing feature data in the m'th historical sequence of the i-th functional area, and then fill the missing feature data location with the average of the feature data values ​​of the previous and next positions.

[0163] Where m' is the historical sequence number, m' = 1, 2, 3, 4, 5 represent the indoor temperature sequence, supply water temperature sequence, return water temperature sequence, flow rate sequence, and outdoor temperature sequence, respectively;

[0164] A2. Obtain the first quartile Q1, median Q2, third quartile Q3, and interquartile range IQR of the m'-th historical sequence of the i-th functional area after processing A1.

[0165] The first quartile Q1 is the 25th quartile of the m'th historical sequence of the i-th functional region;

[0166] The third quartile, Q3, is the 75th quartile of the m'-th historical sequence in the i-th functional region;

[0167] The interquartile range (IQR) is obtained in the following way:

[0168] IQR = Q3 - Q1

[0169] A3. Use Q1, Q2, Q3, and IQR to obtain the upper and lower bounds of outliers:

[0170] Lower bound of outliers = Q1 - 1.5 × IQR

[0171] Upper bound of outliers = Q3 + 1.5 × IQR

[0172] A4. The feature data values ​​in the m'th historical sequence of the i-th functional area that are not within the range of the lower bound to the upper bound of the outlier are the outliers. Then, the position of the outlier is obtained, and the average of the three feature data values ​​before the outlier position and the three feature data values ​​after the outlier position is replaced in the outlier position to obtain the m'th historical sequence of the i-th functional area after the outlier is filled.

[0173] A5. Normalize the m'-th historical sequence of the i-th functional area after outlier imputation to obtain the preprocessed indoor temperature sequence, supply water temperature sequence, return water temperature sequence, flow rate sequence, and outdoor temperature sequence for M historical heating times of the i-th functional area:

[0174]

[0175] Where, x i,m',d It is the d-th feature data value in the m'-th historical sequence of the i-th functional area after outlier imputation, x i ′ ,m',d It is the d-th feature data value in the m'-th historical sequence of the i-th functional region after normalization, x i,m',min It is the minimum feature data in the m'-th historical sequence of the i-th functional area after outlier imputation, x i,m',max It is the maximum value of the feature data in the m'th historical sequence of the i-th functional area after outlier imputation.

[0176] B2. Use the training set to train the indoor temperature prediction model and obtain a well-trained indoor temperature prediction model.

[0177] The indoor temperature prediction model includes: an LSTM module, an efficient additive self-attention mechanism module, and an indoor temperature prediction module;

[0178] The LSTM module is an LSTM network used to obtain the hidden state of the input feature vector:

[0179] h j' =O j' tanh(C j' )

[0180] O j' =σ(W o [h j'-1 ,x j' ]+b o )

[0181]

[0182] f j' =σ(W f [h j'-1 ,x j' ]+b f )

[0183] i j' =σ(W i” [h j'-1 ,x j' ]+b i” )

[0184] Among them, f j' σ is the output of the forget gate at time step j', σ is the sigmoid activation function, and W is the output of the forget gate at time step j'. f W i” W c W o These are the weight matrices for the forget gate, input gate, cell state update, and output gate, respectively. j'-1 It is the hidden state of the previous time step j'-1, x j' The input feature vector at the current time step j includes supply water temperature, return water temperature, flow rate, indoor temperature, and outdoor temperature. f b i” b c b o These are the bias terms for the forget gate, input gate, cell state update, and output gate, respectively. j' It is the output of time step j' of the input gate. It represents the candidate cell state, tanh is the hyperbolic tangent activation function, and C... j' Updated cell state, C j'-1 This refers to the cell state at the previous time step j'-1, O j'It is the output of the output gate at time step j', h j' It is the hidden state of time step j' at that time;

[0185] The efficient additive self-attention mechanism module obtains the global context representation and c' based on the hidden state of the input feature vector:

[0186]

[0187] x' j' =Q j' +T(K j' ×q j' )

[0188] q j' =α j' Q j'

[0189]

[0190] Q j' =h j' W Q

[0191] K j' =h j' W K

[0192] Where, x' j' Here, α is the global context representation of the j'-th time step, H is the total number of time steps, and α is the global context representation of the j'-th time step. j' It is the global attention query vector, d is the dimension of the query matrix, and W is the global attention query vector. Q W K It is the weight matrix for learning, Q j' It is the query vector, K j' It is the key vector, W a It is a learnable parameter vector, q j' It is a single global query vector, and T is a linear transformation;

[0193] The indoor temperature prediction module obtains the predicted indoor temperature value based on the hidden state of the input feature vector and the global context representation:

[0194] y′=W'z'+b

[0195] z' = concat[h j' ,c']

[0196] Where W' is the learnable weight matrix, b is the bias term, and z' is the weight of h. j' The concatenated vector of c and c'.

[0197] Step 522: Compare the predicted indoor temperature of the i-th functional area at the (M+1)-th heating time with the actual indoor temperature of the i-th functional area at the (M+1)-th heating time obtained from the room temperature sensor. If the predicted indoor temperature and the actual indoor temperature are within the preset first error range, proceed to step 53; otherwise, it indicates that the actual indoor temperature of the i-th functional area at the (M+1)-th heating time is an outlier. Replace the actual indoor temperature of the i-th functional area at the (M+1)-th heating time with the correction value, and then proceed to step 53.

[0198] Step 53: Based on the processed temperatures of each functional area in the room to be analyzed at different times, adjust the opening of the heating water supply valves again:

[0199] The system determines whether the temperature error between the functional areas of the room to be analyzed and the design temperature at different times after processing is within the preset second allowable temperature error range. If it is within the preset second allowable temperature error range, the valve opening is not adjusted. If it is not within the preset second allowable temperature error range, the valve opening is adjusted.

[0200] The preset allowable second temperature error is ±1 degree Celsius;

[0201] If the temperature is outside the preset second allowable error range, the valve opening will be adjusted as follows:

[0202] Find the outdoor temperature data P that is closest to the j-th heating time in the historical data, and obtain the valve opening corresponding to the outdoor temperature data P;

[0203] If the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing is higher than the design temperature of the i-th functional zone in the room to be analyzed at the corresponding j-th heating time, then the minimum value of the valve opening corresponding to the outdoor temperature data P will be taken as the current valve opening.

[0204] If the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing is lower than the design temperature of the i-th functional zone in the room to be analyzed at the corresponding j-th heating time, then the maximum value of the valve opening corresponding to the outdoor temperature data P will be taken as the current valve opening.

[0205] This invention acquires various dynamic influencing factors such as indoor temperature, outdoor meteorological data, and population density to achieve precise dynamic thermal demand analysis and dynamic control of buildings. After dynamic control, the actual indoor heating temperature is detected and adjusted again, thereby effectively reducing building energy consumption while improving user comfort. The BIM model can be built based on drawings or after analysis of old community renovations, referring to the optimal renovation plan for further modeling. It also provides two heating modes for users to choose from, thus meeting the needs of different users in multiple ways.

Claims

1. A method for analyzing thermal demand in frigid regions based on building digitalization technology, characterized in that... The specific process of the method is as follows: Step 1: Obtain outdoor meteorological data to be analyzed at different times in the study area; The outdoor meteorological data to be analyzed includes: outdoor temperature, outdoor pressure, and outdoor air density; Step 2: Obtain the heat transfer coefficient of the building envelope in the room to be analyzed, and construct a BIM model of the building containing the room using the heat transfer coefficient of the building envelope, building shape, building structure drawing and building drawing of the room to be analyzed; Step 3: Set the design temperature, personnel density, and equipment usage for each functional area in the room to be analyzed; The functional areas include: office area, meeting room, tea room, and restrooms; Step 4: Using the BIM model of the building where the room to be analyzed is located, conduct thermal demand analysis based on the outdoor meteorological data to be analyzed and the design temperature, personnel density and equipment usage of each functional area of ​​the room to be analyzed, to obtain the required water supply flow rate for each functional area of ​​the room to be analyzed. Step 5: Using the required water supply flow rate of each functional area in the room to be analyzed obtained in Step 4, obtain the valve opening corresponding to the required water supply flow rate, and adjust the valve opening in the room to be analyzed according to the valve opening. Then, use the room temperature sensor to obtain the temperature of each functional area in the room to be analyzed at different times, and then obtain the abnormal values ​​in the temperature of each functional area in the room to be analyzed at different times, and replace the abnormal values ​​with correction values ​​to obtain the processed temperature of each functional area in the room to be analyzed at different times. Based on the processed temperature of each functional area in the room to be analyzed at different times, adjust the valve opening again. In step five, the valve opening corresponding to the required water supply flow rate for each functional area in the room to be analyzed, obtained from step four, is specifically as follows: If the regulating valve has a quick-opening flow characteristic, the required water supply flow rate corresponding to the valve opening can be obtained in the following way: Among them, F max L represents the valve's maximum flow rate. ij L represents the required valve opening degree for the i-th functional zone of the room at the j-th heating time. max R represents the maximum valve opening, R represents the predictive control valve's adjustable ratio, and F represents the maximum valve opening. ij It is the water supply flow required for the i-th functional area of ​​the room at the j-th heating moment; If the regulating valve has a linear flow characteristic, the required water supply flow rate corresponding to the valve opening can be obtained in the following way: If the regulating valve has an equal percentage flow characteristic, the required water supply flow rate corresponding to the valve opening can be obtained in the following way:

2. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 1, characterized in that: In step four, the BIM model of the building containing the room to be analyzed is used to conduct a thermal demand analysis based on the outdoor meteorological data and the design temperature, personnel density, and equipment usage of each functional area within the room to be analyzed, in order to obtain the required water supply flow rate for each functional area within the room to be analyzed. Specifically: Step 4.1: Obtain the heat load of the i-th functional zone in the room to be analyzed at the j-th heating time. Specifically: Q ij Q1 ij +Q2 ij +Q3 ij -Q4 ij Q1 ij =αKA(t ij -t w j ) Q2 ij =0.278NVc p ρ w (t ij -t w j ) Q3 ij =Q1 ij (1+b c +b f +b q +b m )(1+b fg )(1+b j ) Q4 ij =η(zq r +W) Among them, Q ij Q1 represents the heat load of the i-th functional zone in the room at the j-th heating time, where i ∈ [1, m], i is the functional zone number to be evaluated, m is the total number of functional zones to be evaluated, and j ∈ [1, n], j is the heating time number, and n is the total number of heating times. ij Q2 is the basic heat loss of the building envelope in the i-th functional zone at the j-th heating moment. ij Q3 is the heat loss due to cold air infiltration in the i-th functional zone at the j-th heating moment. ij Q4 represents the additional heat loss within the building at the j-th heating time in the i-th functional zone. ij This represents the total heat dissipation from people and equipment in the room to be analyzed at the j-th heating moment in the i-th functional zone. α is the temperature difference correction coefficient, K represents the heat transfer coefficient of the building envelope, A represents the heat transfer area of ​​the building envelope, and t... ij t represents the design temperature of the i-th functional zone in the room at the j-th heating time. w j Let N represent the average outdoor temperature at the j-th heating moment, N represent the indoor air exchange rate, V represent the net room volume, and c represent the average outdoor temperature at the j-th heating moment. p It is the specific heat capacity of outdoor air at constant pressure, ρ w β represents outdoor air density. c Indicates the preset orientation correction rate, β f Indicates the preset wind correction rate, β q This indicates the preset correction rate for the two exterior walls, β. m This indicates the correction rate for an excessively large window-to-wall ratio, β. fg This represents the preset room height correction rate, β. j η represents the preset intermittent correction rate, z represents the preset correction rate for heat dissipation by people and equipment, and q represents the number of people in the room to be analyzed. r W represents the heat dissipation of the human body, and W represents the heat dissipation of the equipment. Step 4.2: Obtain the required water supply flow rate F for the i-th functional zone of the room under analysis at the j-th heating time using the heat load at the j-th heating time. ij .

3. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 2, characterized in that: In step four-two, the required water supply flow rate F for the i-th functional area of ​​the room under analysis at the j-th heating time is obtained by using the heat load of the i-th functional area of ​​the room under analysis at the j-th heating time. ij Specifically: F ij =Q ij / (cρΔT) Where c represents the specific heat capacity of water, ρ represents the density of water, ΔT represents the temperature difference between the supply and return water, and F ij It is the water supply flow required for the i-th functional area of ​​the room at the j-th heating moment.

4. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 3, characterized in that: Step five involves using a room temperature sensor to obtain the temperature of each functional area within the room to be analyzed at different times, then acquiring outliers in the temperatures of each functional area at different times, and replacing these outliers with corrected values ​​to obtain the processed temperatures of each functional area within the room to be analyzed at different times. Specifically: First, obtain the historical sequence of M historical heating times for the i-th functional area of ​​the room to be analyzed. Input the historical sequence of M historical heating times for the i-th functional area of ​​the room to be analyzed into the trained indoor temperature prediction model to obtain the indoor predicted temperature of the i-th functional area at the (M+1)-th heating time, i.e., the correction value. The historical sequence of M historical heating times in the i-th functional area includes: the indoor temperature sequence of M historical heating times in the i-th functional area, the water supply temperature sequence of M historical heating times in the i-th functional area, the return water temperature sequence of M historical heating times in the i-th functional area, the flow rate sequence of M historical heating times in the i-th functional area, and the outdoor temperature sequence of M historical heating times in the i-th functional area. Then, the predicted indoor temperature of the i-th functional zone at the (M+1)-th heating time is compared with the actual indoor temperature of the i-th functional zone at the (M+1)-th heating time obtained by the room temperature sensor. If the predicted indoor temperature and the actual indoor temperature are within the preset first error range, the valve opening is adjusted again based on the processed temperatures of each functional zone in the room to be analyzed at different times; otherwise, the actual indoor temperature of the i-th functional zone at the (M+1)-th heating time is replaced with the correction value, and the valve opening is adjusted again based on the processed temperatures of each functional zone in the room to be analyzed at different times.

5. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 4, characterized in that: The trained indoor temperature prediction model is obtained through the following method: B1. Obtain the historical sequences of M historical heating times for the i-th functional area of ​​the room to be analyzed, and preprocess each historical sequence. Combine the preprocessed historical sequences with the indoor temperature of the i-th functional area of ​​the room to be analyzed at the (M+1)-th heating time to form a training set. The historical sequence of M historical heating times in the i-th functional area includes: the indoor temperature sequence of M historical heating times in the i-th functional area, the water supply temperature sequence of M historical heating times in the i-th functional area, the return water temperature sequence of M historical heating times in the i-th functional area, the flow rate sequence of M historical heating times in the i-th functional area, and the outdoor temperature sequence of M historical heating times in the i-th functional area. B2. Use the training set to train the indoor temperature prediction model and obtain a well-trained indoor temperature prediction model.

6. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 5, characterized in that: The preprocessing of each historical sequence in B1 specifically includes: A1. Obtain the location of missing feature data in the m'th historical sequence of the i-th functional area, and then fill the missing feature data location with the average of the feature data values ​​of the previous and next positions. Where m' is the historical sequence number, m' = 1, 2, 3, 4, 5 represent the indoor temperature sequence, the water supply temperature sequence, the return water temperature sequence, the flow rate sequence, and the outdoor temperature sequence, respectively; A2. Obtain the first quartile Q1, the third quartile Q3, and the interquartile range IQR of the m'-th historical sequence of the i-th functional area after processing by A1. The first quartile Q1 is the 25th quartile of the m'th historical sequence of the i-th functional region; The third quartile, Q3, is the 75th quartile of the m'-th historical sequence in the i-th functional region; The interquartile range (IQR) is obtained in the following way: IQR = Q3 - Q1 A3. Use Q1, Q3, and IQR to obtain the upper and lower bounds of outliers: Lower bound of outliers = Q1 - 1.5 × IQR Upper bound of outliers = Q3 + 1.5 × IQR A4. The feature data values ​​in the m'th historical sequence of the i-th functional area that are not within the range of the lower bound to the upper bound of the outlier are the outliers. Then, the position of the outlier is obtained, and the average of the three feature data values ​​before the outlier position and the three feature data values ​​after the outlier position is replaced in the outlier position to obtain the m'th historical sequence of the i-th functional area after the outlier is filled. A5. Normalize the m'-th historical sequence of the i-th functional area after outlier filling to obtain the indoor temperature sequence, supply water temperature sequence, return water temperature sequence, flow rate sequence, and outdoor temperature sequence of the M historical heating times of the i-th functional area after preprocessing.

7. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 6, characterized in that: The normalization process for the m'-th historical sequence of the i-th functional area after outlier imputation in A5 is as follows: Where, x i,m',d It is the d-th feature data value in the m'-th historical sequence of the i-th functional area after outlier imputation, x′ i,m',d It is the d-th feature data value in the m'-th historical sequence of the i-th functional region after normalization, x i,m',min It is the minimum feature data in the m'-th historical sequence of the i-th functional area after outlier imputation, x i,m',max It is the maximum value of the feature data in the m'th historical sequence of the i-th functional area after outlier imputation.

8. The method for analyzing thermal demand in frigid regions based on building digitalization technology according to claim 7, characterized in that: The indoor temperature prediction model in B2 includes: an LSTM module, an efficient additive self-attention mechanism module, and an indoor temperature prediction module. The LSTM module is an LSTM network used to obtain the hidden state of the input feature vector, specifically: h j' =O j' fishy(C) j' ) The j' =σ(W o [h j'-1 ,x j' ]+b o ) f j' =σ(W f [h j'-1 ,x j' ]+b f ) i j' =σ(W i” [h j'-1 ,x j' ]+b i” ) Among them, f j' σ is the output of the forget gate at time step j', σ is the sigmoid activation function, and W is the output of the forget gate at time step j'. f It is the forget gate weight matrix, W i” It is the input gate weight matrix, W c W is the cell state update weight matrix. o It is the output gate weight matrix, h j'-1 It is the hidden state of the previous time step j'-1, x j' b is the input feature vector at the current time step j'. f It is the forgetting gate bias term, b i” It is the input gate bias term, b c It is the cell state update bias term, b o It is the output gate bias term, i j' It is the output of time step j' of the input gate. It represents the candidate cell state, tanh is the hyperbolic tangent activation function, and C... j' Updated cell state, C j'-1 This refers to the cell state at the previous time step j'-1, O j' It is the output of the output gate at time step j', h j' It is the hidden state of time step j' at that time; The input feature vector includes: supply water temperature, return water temperature, flow rate, indoor temperature, and outdoor temperature; The efficient additive self-attention mechanism module obtains the global context representation and c' based on the hidden state of the input feature vector: x' j' =Q j' +T(K j' ×q j' ) q j' =α j' Q j' Q j' =h j' W Q K j' =h j' W K Where, x' j' It is the global context representation at the j'-th time step, α j' It is the global attention query vector, d is the dimension of the query matrix, and W is the global attention query vector. Q W K It is the weight matrix for learning, Q j' It is the query vector, K j' It is the key vector, W a It is a preset parameter vector, q j' It is a single global query vector, T is the linear transformation, and H is the total number of time steps; The indoor temperature prediction module obtains the predicted indoor temperature value based on the hidden state of the input feature vector and the global context representation: y′=W'z'+b z'=concat[h j' ,c'] Where W' is the preset weight matrix, b is the bias term, and z' is h j' The concatenated vector of c' and y' is the predicted indoor temperature value.

9. A method for analyzing thermal demand in frigid regions based on building digitalization technology as described in claim 8, characterized in that: The step five, which involves further adjusting the valve opening based on the processed temperatures of each functional area within the room to be analyzed at different times, specifically involves: Determine whether the error between the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing and the design temperature of the i-th functional zone in the room to be analyzed at the j-th heating time is within the preset allowable second temperature error range. If it is within the preset allowable second temperature error range, the valve opening is not adjusted. If it is not within the preset allowable second temperature error range, the valve opening is adjusted. If the temperature is outside the preset second allowable error range, the valve opening will be adjusted as follows: Find the outdoor temperature data P that is closest to the j-th heating time in the historical data, and obtain the valve opening corresponding to the outdoor temperature data P; If the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing is higher than the design temperature of the i-th functional zone in the room to be analyzed at the corresponding j-th heating time, then the minimum value of the valve opening corresponding to the outdoor temperature data P will be taken as the current valve opening. If the temperature of the i-th functional zone in the room to be analyzed at the j-th heating time after processing is lower than the design temperature of the i-th functional zone in the room to be analyzed at the corresponding j-th heating time, then the maximum value of the valve opening corresponding to the outdoor temperature data P will be taken as the current valve opening.

Images