A centralized dispatching method and system for building air conditioning load

Abstract

The present invention discloses a centralized dispatching method and system for building air conditioning load, and relates to the field of air conditioning load dispatching. The method comprises: obtaining the floor correlation between the floor and the air conditioning load data by analyzing the acquired historical air conditioning load data of the building; constructing an air conditioning load prediction model based on the floor correlation and in combination with the pre-acquired user preference data; predicting the air conditioning load of each floor in the building by using the air conditioning load prediction model to obtain a prediction result; and centrally dispatching the air conditioning load based on the prediction result and in combination with the air conditioning power consumption data. The present invention can obtain the floor correlation between the floor and the air conditioning load data by analyzing the historical air conditioning load data of several buildings, so as to deeply understand the difference in air conditioning load between each floor in the building and its changing law, which is helpful to realize the effective saving of electricity and promote the high efficiency and sustainability of building energy use.

Description

Technical Field

[0001] The present invention relates to air conditioning load scheduling, and in particular to a centralized scheduling method and system for building air conditioning load. Background Art

[0002] In recent years, with the rapid growth of economy and electricity demand, electricity consumption and the proportion of electricity consumption have been rising year by year. Among the many power terminal equipment, air conditioning load accounts for a large proportion. Due to the large number and large capacity, air conditioning load usually constitutes an important part of peak load, further exacerbating the peak-to-valley difference, making the proportion of air conditioning load high and showing a significant load peak phenomenon.

[0003] Building air conditioning load mainly refers to the total heat or cold that the air conditioning system in the building needs to process in order to maintain the indoor environment. By implementing reasonable direct load control technology, it can not only quickly respond to power dispatching needs, achieve power conservation, and effectively reduce power load during peak hours, thereby alleviating the contradiction between power supply and demand, but also has almost no impact on the user's power comfort. The number of air conditioning loads after centralized control is considerable, making building air conditioning loads a valuable energy-saving resource for power companies and providing an effective way to save electricity.

[0004] With the deepening of power reform, the power market and power-saving technologies have been widely studied in recent years. In view of the characteristics of buildings and the widespread application of central air-conditioning systems, it is possible to centrally control the air-conditioning load of buildings and show its huge control potential. The existing centralized dispatching method of air-conditioning load, by combining user preferences, has achieved energy-saving effects and alleviated the pressure of electricity consumption to a certain extent. However, the differences in characteristics between different buildings and differences in floors make centralized dispatching random and uncontrollable risks. This uncertainty may lead to an increase in electricity demand, resulting in higher standby costs and risk costs.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the problems in the related art, the present invention proposes a centralized scheduling method and system for building air-conditioning loads to overcome the above technical problems existing in the existing related art.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] According to one aspect of the present invention, a method for centralized scheduling of building air conditioning loads is provided, the method comprising the following steps:

[0009] S1. Analyze the historical air conditioning load data of the building to obtain the floor correlation between the floor and the air conditioning load data;

[0010] S2. Based on the floor correlation and combined with the pre-acquired user preference data, an air conditioning load prediction model is constructed;

[0011] S3. Using the air conditioning load prediction model, predict the air conditioning load of each floor in the building to obtain a prediction result;

[0012] S4. Based on the prediction results and combined with the air conditioning power consumption data, the air conditioning load is centrally scheduled.

[0013] Furthermore, obtaining the floor correlation between the floors and the air conditioning load data by analyzing the acquired historical air conditioning load data of the building includes the following steps:

[0014] S11. Using the building air conditioning dispatching platform, obtain the historical air conditioning load data of several buildings and perform preprocessing, the historical air conditioning load data including historical air conditioning power consumption data, air conditioning operation status data, ambient temperature data, building data and floor data;

[0015] S12. According to the building data, the pre-processed historical air conditioning load data is processed in layers to establish a number of building layered data sets, and the building layered data sets are decomposed into a number of floor layered data sets in combination with the floor data;

[0016] S13, based on the layered data sets of each floor, analyzing the historical air conditioning load data of the same floor in the same building, and obtaining the air conditioning power consumption data on the same floor;

[0017] S14, combining the stratified data sets of each building, analyzing the floor correlation between the floors and the air conditioning power consumption data;

[0018] S15. Based on the floor correlation analysis result, obtain the floor correlation between the floor and the air conditioning load data.

[0019] Furthermore, based on the layered data sets of each floor, analyzing the historical air conditioning load data of the same floor in the same building, obtaining the air conditioning power consumption data on the same floor includes the following steps:

[0020] S131, selecting historical air conditioning load data of the same floor in the same building from the layered data sets of each floor;

[0021] S132, analyzing the changes of air conditioning power consumption data under different air conditioning operation status data and ambient temperature data according to the historical air conditioning load data of the same floor in the same building;

[0022] S133. Based on the analysis result of the change of the air-conditioning power consumption data, obtain the air-conditioning power consumption data on the same floor.

[0023] Furthermore, combining the stratified data sets of each building, analyzing the floor correlation between the floors and the air conditioning power consumption data includes the following steps:

[0024] S141, based on the air conditioning power consumption data on the same floor, combined with the layered data sets of each building, analyzing the air conditioning power consumption data on the same floor in different buildings;

[0025] S142, based on the analysis results of the air conditioning power consumption data on the same floor in different buildings, using a building correlation algorithm to calculate the building correlation between the building data and the air conditioning power consumption data;

[0026] S143. According to the building correlation, a floor correlation algorithm is used to calculate the floor correlation between the air conditioning power consumption data of the same floor and the floor above the same floor in different buildings.

[0027] Furthermore, based on the analysis results of the air conditioning power consumption data on the same floor in different buildings, the building correlation between the building data and the air conditioning power consumption data is calculated using the building correlation algorithm, including the following steps:

[0028] S1421, scanning the air conditioning power consumption data on the same floor in different buildings to obtain an original data set, and dividing the original data set into a time-level data set according to a time constraint;

[0029] S1422, based on the set support index value, filter out the air-conditioning power consumption data that meets the preset threshold from the time-level data set, and generate the time-level power consumption data set;

[0030] S1423, analyzing the time level one data set, obtaining the time level two data set according to the aggregation degree between the air conditioner power consumption data, and generating the time level two power consumption data set in combination with the set support index value;

[0031] S1424, repeat step S1423, sequentially obtain the time N-level data set and the time N-level electricity consumption data set, until no new time N+1-level data set and time N+1-level electricity consumption data set can be obtained, and all electricity consumption data sets are obtained;

[0032] S1425. Combining the set support and confidence index values, obtain the building correlation between the building data and the air-conditioning power consumption data at different times.

[0033] Furthermore, according to the building correlation, using the floor correlation algorithm to calculate the floor correlation between the air conditioning power consumption data of the same floor and the floor above the same floor in different buildings includes the following steps:

[0034] S1431. Obtain building correlation amplitude according to the building correlation, and use the ratio of the building correlation amplitude within a certain period of time to the building correlation amplitude within the overall period of time as the set amplitude for mode conversion;

[0035] S1432, using the primary scale time window to scan the building correlation amplitude within the overall time, and taking the time window in which the building correlation amplitude exceeds the set amplitude as the time window for mode conversion;

[0036] S1433, using the time window of the secondary scale to detect the time window of the mode conversion;

[0037] S1434, repeating steps S1432-S1433 until no new mode conversion can be detected, obtaining the occurrence point of the mode conversion, and recording the time of several occurrence points;

[0038] S1435, taking the time of each occurrence point as the center, using a preset time window to determine whether a mode conversion occurs within the time neighborhood of each occurrence point, and counting the number of mode conversions at each occurrence point, to form a data set of the number of mode conversions between the same floor and the previous floor in different buildings;

[0039] S1436. Based on the data set of the number of mode conversions, calculate the floor correlation between the same floor and the floor above the same floor in different buildings.

[0040] Furthermore, the floor correlation calculation formula is:

[0041]

[0042] In the formula, R represents the floor correlation value between the same floor and the floor above the same floor in different buildings;

[0043] x represents the number of mode transitions at the occurrence point;

[0044] y represents the building correlation value between building data and air conditioning power consumption data;

[0045] Both t1 and t2 are expressed as arbitrary times;

[0046] l a Indicates the upper floor of the same floor in different buildings;

[0047] l b Indicates the same floor in different buildings.

[0048] Furthermore, based on the floor correlation and combined with the pre-acquired user preference data, the air conditioning load prediction model is constructed, including the following steps:

[0049] S21, analyzing the floor correlation and extracting feature data from the floor correlation;

[0050] S22, using an improved hierarchical clustering algorithm to mine the user's air conditioning power consumption pattern, and combining the air conditioning power consumption pattern with the ambient temperature data to form user preference data;

[0051] S23, pre-processing the feature data and the user preference data to establish a data set;

[0052] S24. Based on the data set and combined with the deep learning model, an air conditioning load prediction model is constructed;

[0053] S25, using historical building air conditioning load data to train an air conditioning load prediction model, and evaluating the trained air conditioning load prediction model;

[0054] S26. Optimize the air conditioning load prediction model based on the evaluation results.

[0055] Furthermore, using the improved hierarchical clustering algorithm to mine the user's air conditioning power consumption pattern, and combining the air conditioning power consumption pattern with the ambient temperature data to form user preference data includes the following steps:

[0056] S221, using historical air conditioning load data as clustering units, and initializing the clustering units;

[0057] S222, based on the initialized clustering units, calculating the covariance of each clustering unit and initializing the similarity matrix;

[0058] S223, selecting two clustering units with the largest similarity values ​​from the similarity matrix, iteratively merging the two clustering units and updating the similarity matrix, and repeating the iterative fusion until all historical air conditioning load data are merged into one unit;

[0059] S225, outputting a hierarchical clustering dendrogram according to the change in harmony of the historical air conditioning load data in the iterative fusion, and selecting a point where the harmony decreases in the dendrogram as a cutting threshold;

[0060] S226. Based on the cutting threshold, the dendrogram is cut into several clusters, the ambient temperature data in the corresponding time period of each cluster is collected, the correlation between the air-conditioning power consumption pattern in the cluster and the ambient temperature data is analyzed, and the user preference data is generated.

[0061] According to another aspect of the present invention, there is also provided a building air conditioning load centralized dispatching system, the system comprising a floor correlation analysis module, a prediction model building module, an air conditioning load prediction module and a centralized dispatching module;

[0062] The floor correlation analysis module, the prediction model building module, the air conditioning load prediction module and the centralized dispatching module are connected in sequence;

[0063] A floor correlation analysis module is used to obtain the floor correlation between the floor and the air conditioning load data by analyzing the acquired historical air conditioning load data of the building;

[0064] A prediction model building module is used to build an air conditioning load prediction model based on floor correlation and pre-acquired user preference data;

[0065] The air conditioning load prediction module is used to predict the air conditioning load of each floor in the building using the air conditioning load prediction model to obtain the prediction result;

[0066] The centralized dispatching module is used to centrally dispatch the air conditioning load based on the prediction results and the air conditioning power consumption data.

[0067] The beneficial effects of the present invention are:

[0068] 1. The present invention can obtain the floor correlation between floors and air-conditioning load data by analyzing the historical air-conditioning load data of several buildings, so as to deeply understand the difference in air-conditioning load between floors in the building and its changing rules, and provide basic data support for the subsequent construction of an air-conditioning load prediction model. By mining the floor correlation, the characteristics of the air-conditioning load on each floor can be grasped more accurately, providing a strong basis for subsequent scheduling work, helping to achieve effective electricity conservation, and promoting the efficiency and sustainability of building energy use.

[0069] 2. By comprehensively considering the floor correlation and user preferences, the present invention can construct an air-conditioning load prediction model that is more in line with the actual situation. It not only reflects the personalized needs of different users for air-conditioning use, but also improves the accuracy and reliability of the prediction results, and realizes the accurate prediction of the air-conditioning load on each floor in the building, providing powerful guidance for the subsequent air-conditioning load scheduling, helping managers to better grasp the air-conditioning usage of the building, providing decision-making support for energy saving and consumption reduction, and promoting the optimization and sustainable development of building energy management.

[0070] 3. The present invention analyzes the air-conditioning power consumption data and centrally dispatches the air-conditioning load, thereby achieving optimal distribution and reasonable control of the building air-conditioning load, and formulating a more scientific and reasonable dispatching plan to ensure that the air-conditioning energy consumption is reduced while meeting the user's comfort level; at the same time, centralized dispatching also helps to improve the operating efficiency of the building air-conditioning system, reduce energy waste, and further achieve the goal of energy conservation and emission reduction, and promote the green and intelligent development of building energy management. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0072] Figure 1 is a flow chart of a centralized scheduling method for building air conditioning loads according to an embodiment of the present invention;

[0073] Figure 2 The invention is a block diagram of a centralized building air conditioning load dispatching system according to an embodiment of the invention.

[0074] In the figure:

[0075] 1. Floor correlation analysis module; 2. Prediction model building module; 3. Air conditioning load prediction module; 4. Centralized scheduling module. DETAILED DESCRIPTION

[0076] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. By referring to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0077] According to an embodiment of the present invention, a method and system for centralized scheduling of building air-conditioning loads are provided.

[0078] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to the centralized scheduling method of building air conditioning loads according to an embodiment of the present invention, the method comprises the following steps:

[0079] S1. Analyze the historical air-conditioning load data of the building to obtain the floor correlation between the floor and the air-conditioning load data.

[0080] Specifically, obtaining the floor correlation between the floors and the air conditioning load data by analyzing the acquired historical air conditioning load data of the building includes the following steps:

[0081] S11. Using the building air conditioning dispatching platform, obtain the historical air conditioning load data of several buildings and perform preprocessing. The historical air conditioning load data includes historical air conditioning power consumption data, air conditioning operation status data, ambient temperature data, building data and floor data.

[0082] It should be noted that the acquired historical air-conditioning load data is preprocessed, including data cleaning, data deduplication, outlier processing, missing value processing, etc.; among them, data cleaning is to remove unreasonable data, such as historical air-conditioning load data that exceeds a reasonable range or repeated historical air-conditioning load data; outlier processing is to remove outliers in the historical air-conditioning load data, and missing value processing is to fill in the missing parts of the historical air-conditioning load data.

[0083] S12. According to the building data, the pre-processed historical air-conditioning load data is layered to establish a number of building layered data sets, and combined with the floor data, the building layered data sets are decomposed into a number of floor layered data sets.

[0084] It should be noted that the historical air-conditioning load data is initially stratified based on key information in the building data, such as building type (office building, residential building, commercial building, etc.), construction age, building area, etc. The purpose of this step is to classify buildings with similar characteristics into the same data set for subsequent analysis; for each type of building, an independent building stratification data set is established based on its historical air-conditioning load data (including air-conditioning power consumption data, air-conditioning operating status data, ambient temperature data, etc.). Each data set should contain all relevant information of that type of building to facilitate further analysis and mining in the data set.

[0085] On the basis of each building hierarchical data set, the building hierarchical data set is further combined with the floor data to refine it into the floor hierarchical data set. This step requires data segmentation and classification according to the floor information (such as floor height, floor area, floor function, etc.): For each building, the building hierarchical data set is further decomposed into several floor hierarchical data sets according to its internal floor data; each floor hierarchical data set should contain all air-conditioning load data on the floor, including but not limited to power consumption, operating status, ambient temperature, etc.

[0086] S13. Based on the layered data sets of each floor, analyze the historical air-conditioning load data of the same floor in the same building to obtain the air-conditioning power consumption data on the same floor.

[0087] Specifically, based on the layered data sets of each floor, analyzing the historical air conditioning load data of the same floor in the same building, obtaining the air conditioning power consumption data on the same floor includes the following steps:

[0088] S131. Select historical air conditioning load data of the same floor in the same building from the layered data sets of each floor.

[0089] S132. Analyze changes in air-conditioning power consumption data under different air-conditioning operation status data and ambient temperature data based on historical air-conditioning load data of the same floor in the same building.

[0090] It should be noted that, based on the historical air conditioning load data of the same floor in the same building, analyzing the changes in air conditioning power consumption data under different air conditioning operation status data and ambient temperature data includes the following steps:

[0091] The historical air conditioning load data of specific floors in the same building are screened out from the floor-stratified data set. The historical air conditioning load data includes detailed records of air conditioning power consumption, operating status and ambient temperature; the historical air conditioning load data are classified according to the air conditioning operating status data. The air conditioning operating status usually includes on, off, cooling, heating, air supply and other modes. These data are classified according to different operating statuses for further analysis; according to the ambient temperature data, the screened historical air conditioning load data are subdivided into different temperature ranges to observe the changes in air conditioning power consumption in different temperature ranges; for each air conditioning operating status classification, the average air conditioning power consumption in that state is calculated, and the average power consumption of the air conditioning in that state is compared with the average power consumption of the air conditioning in different operating states. The power consumption data under different operating modes is collected to obtain the basic power consumption information under different operating modes, and the possible power consumption differences and patterns are identified; in each temperature range, the changes in air-conditioning power consumption with ambient temperature are analyzed, and by drawing a scatter plot or trend line between power consumption and temperature, the distribution and trend of data points are observed and analyzed to identify whether there is an obvious positive or negative correlation, and whether there is an inflection point or mutation point; considering the two factors of air-conditioning operating status and ambient temperature, a comprehensive analysis is conducted on the changes in air-conditioning power consumption data, and by comparing the power consumption data under different operating states and temperature ranges, possible regularities or anomalies are analyzed, and the analysis results are organized into tables or reports for subsequent use and reference.

[0092] S133. Based on the analysis result of the change of the air-conditioning power consumption data, obtain the air-conditioning power consumption data on the same floor.

[0093] S14. Analyze the floor correlation between floors and air-conditioning power consumption data in combination with the stratified data sets of each building.

[0094] Specifically, combining the stratified data sets of each building, analyzing the floor correlation between the floors and the air conditioning power consumption data includes the following steps:

[0095] S141. Based on the air conditioning power consumption data on the same floor, combined with the stratified data sets of each building, analyze the air conditioning power consumption data on the same floor in different buildings.

[0096] It should be noted that based on the air conditioning power consumption data on the same floor, combined with the layered data sets of each building, the analysis of the air conditioning power consumption data on the same floor in different buildings includes the following steps:

[0097] The air conditioning power consumption data of a specific floor is screened out from the stratified data sets of each building, and the screened air conditioning power consumption data is time-aligned to ensure that the data of different buildings are compared on the same time scale; the air conditioning power consumption data of the same floor in different buildings can be compared by drawing visualization methods such as bar charts, line charts or box plots, and the difference in air conditioning power consumption on the same floor of different buildings is analyzed to observe whether there is an obvious trend or pattern; combined with other information of the building (such as building type, age, area, etc.), a preliminary analysis of the possible impact of these factors on the air conditioning power consumption on the same floor is conducted, and factors such as design differences, equipment configuration or usage habits between different buildings are considered to analyze their potential impact on air conditioning power consumption; the results of the analysis are organized into a table or report form, including comparison charts, statistical data and conclusions of the preliminary analysis, and the results are output for use in subsequent steps (such as calculating building correlation, floor correlation, etc.).

[0098] S142. Based on the analysis results of the air-conditioning power consumption data on the same floor in different buildings, a building correlation algorithm is used to calculate the building correlation between the building data and the air-conditioning power consumption data.

[0099] Specifically, based on the analysis results of the air conditioning power consumption data on the same floor in different buildings, using the building correlation algorithm to calculate the building correlation between the building data and the air conditioning power consumption data includes the following steps:

[0100] S1421, scanning the air conditioning power consumption data on the same floor in different buildings to obtain an original data set, and dividing the original data set into a time-level data set according to a time constraint;

[0101] S1422, based on the set support index value, filter out the air-conditioning power consumption data that meets the preset threshold from the time-level data set, and generate the time-level power consumption data set;

[0102] S1423, analyzing the time level one data set, obtaining the time level two data set according to the aggregation degree between the air conditioner power consumption data, and generating the time level two power consumption data set in combination with the set support index value;

[0103] S1424, repeat step S1423, sequentially obtain the time N-level data set and the time N-level electricity consumption data set, until no new time N+1-level data set and time N+1-level electricity consumption data set can be obtained, and all electricity consumption data sets are obtained;

[0104] S1425. Combining the set support and confidence index values, obtain the building correlation between the building data and the air-conditioning power consumption data at different times.

[0105] It should be noted that if the building data is recorded as X and the air conditioning power consumption data is recorded as Y, the support index is:

[0106] Association Rules The support of represents the probability of X and Y item sets appearing at the same time:

[0107]

[0108] In the formula, Association rules The support index of

[0109] P(X, Y) represents the probability value between building data X and air conditioning power consumption data Y;

[0110] X represents building data;

[0111] Y represents the power consumption data of air conditioner;

[0112] n(X, Y) represents the number of times building data X and air conditioning power consumption data Y appear simultaneously;

[0113] n(A) represents the total number of building data X and air conditioning power consumption data Y.

[0114] The confidence index is:

[0115] Association Rules The confidence level indicates the probability of another item set Y appearing after an item set X appears:

[0116]

[0117] In the formula, Association rules Confidence index of

[0118] P(X|Y) represents the probability value of air conditioning power consumption data Y appearing after building data X appears;

[0119] X represents building data;

[0120] Y represents the power consumption data of air conditioner;

[0121] n(X, Y) represents the number of times building data X and air conditioning power consumption data Y appear simultaneously;

[0122] n(X) represents the number of occurrences of building data X.

[0123] S143. According to the building correlation, a floor correlation algorithm is used to calculate the floor correlation between the air conditioning power consumption data of the same floor and the floor above the same floor in different buildings.

[0124] Specifically, according to the building correlation, using the floor correlation algorithm to calculate the floor correlation between the air conditioning power consumption data of the same floor and the floor above the same floor in different buildings includes the following steps:

[0125] S1431. Obtain building correlation amplitude according to building correlation, and use the ratio of the building correlation amplitude within a certain period of time to the building correlation amplitude within the overall period of time as the set amplitude for mode conversion.

[0126] It should be noted that mode conversion means that when the amplitude of data in a certain time window is greater than or equal to a certain proportion of the overall amplitude of the data, it can be determined that mode conversion has occurred in the time window.

[0127] The calculation formula for the building correlation amplitude over the entire time period is:

[0128] D=D max -D min

[0129] In the formula, D represents the maximum building correlation amplitude that occurs in a complete time period;

[0130] D max It is expressed as the maximum value of the building correlation amplitude obtained during the whole time;

[0131] D min It is expressed as the minimum building correlation amplitude obtained over the entire time.

[0132] The amplitude calculation formula that occurs within a certain time window is:

[0133] d=d max -d min

[0134] In the formula, d represents the maximum building correlation amplitude that occurs within a certain time window;

[0135] d max It is expressed as the maximum value of the building correlation amplitude obtained within a certain time window;

[0136] d min It is expressed as the minimum building correlation amplitude obtained within a certain time window.

[0137] Judgment of mode conversion: When the following formula is satisfied, it can be determined that a mode conversion has occurred, and K is the proportionality coefficient.

[0138] d ≥ K*D

[0139] According to the characteristics of different buildings, time windows of different scales and different proportional coefficients are used to detect whether the mode conversion occurs. For transient variables (current, voltage, etc.) whose building correlation changes rapidly when the mode conversion occurs, a small-scale time window can be used to determine whether the mode conversion occurs within the amplitude range.

[0140] S1432: Scan the building correlation amplitude within the overall time using the primary scale time window, and use the time window where the building correlation amplitude exceeds the set amplitude as the time window for mode conversion.

[0141] S1433. Use the time window of the secondary scale to detect the time window of the mode conversion.

[0142] S1434. Repeat steps S1432-S1433 until no new mode conversion can be detected, obtain the occurrence point of the mode conversion, and record the time of several occurrence points.

[0143] S1435. Taking the time of each occurrence point as the center, use the preset time window to determine whether a mode conversion occurs within the time neighborhood of each occurrence point, and count the number of mode conversions at each occurrence point to form a data set of the number of mode conversions on the same floor and the previous floor in different buildings.

[0144] S1436. Based on the data set of the number of mode conversions, calculate the floor correlation between the same floor and the floor above the same floor in different buildings.

[0145] Specifically, the floor correlation calculation formula is:

[0146]

[0147] In the formula, R represents the floor correlation value between the same floor and the floor above the same floor in different buildings;

[0148] x represents the number of mode transitions at the occurrence point;

[0149] y represents the building correlation value between building data and air conditioning power consumption data;

[0150] Both t1 and t2 are expressed as arbitrary times;

[0151] l a Indicates the upper floor of the same floor in different buildings;

[0152] l b Indicates the same floor in different buildings.

[0153] S15. Based on the floor correlation analysis result, obtain the floor correlation between the floor and the air conditioning load data.

[0154] It should be noted that, based on the floor correlation analysis results, obtaining the floor correlation between the floor and the air conditioning load data includes the following steps:

[0155] The obtained floor correlation is matched with the corresponding air-conditioning load data, which includes matching the correlation value of each floor with the historical data of the air-conditioning power consumption, operating status, etc. of the floor, and analyzing the relationship between the floor correlation and the air-conditioning load data; analyzing the combined data, observing the correlation between the floor correlation and the air-conditioning load data, and judging whether the floor correlation has a significant impact on the air-conditioning load by comparing the correlation values ​​of different floors and the corresponding air-conditioning load data; using statistical analysis methods (such as correlation analysis, regression analysis, etc.) or machine learning algorithms (such as cluster analysis, classification algorithms, etc.) to identify the relationship pattern between the floor correlation and the air-conditioning load data, and through these analysis methods, deeply understand the relationship between the floor correlation and the air-conditioning load data. Positive correlation, negative correlation or other complex relationships between air-conditioning loads; explain the relationship patterns between the identified floor correlations and air-conditioning load data, explain the significance of these patterns in actual building air-conditioning load scheduling, evaluate the impact of floor correlations on air-conditioning loads, and determine which floor combinations require special attention or coordinated scheduling in air-conditioning load scheduling; integrate the relationship patterns between the identified floor correlations and air-conditioning load data and the impact assessment results to form a complete floor correlation analysis result, which will include information such as the correlation values, relationship patterns, and impact levels between each floor; output the floor correlation analysis results in the form of a report or chart to show the relationship between floor correlations and air-conditioning load data.

[0156] S2. Based on the floor correlation and combined with the pre-acquired user preference data, an air conditioning load prediction model is constructed.

[0157] Specifically, based on the floor correlation and combined with the pre-acquired user preference data, building an air conditioning load prediction model includes the following steps:

[0158] S21. Analyze the floor correlation and extract feature data from the floor correlation.

[0159] It should be noted that the characteristic data include statistical indicators such as the maximum value, minimum value, mean value, median, and standard deviation of the floor correlation. At the same time, it is also possible to consider extracting dynamic characteristics such as the changing trend and volatility of the correlation to more comprehensively reflect the characteristics of the floor correlation.

[0160] S22. Use an improved hierarchical clustering algorithm to mine the user's air conditioning power consumption pattern, and combine the air conditioning power consumption pattern with the ambient temperature data to form user preference data.

[0161] Specifically, using the improved hierarchical clustering algorithm to mine the user's air conditioning power consumption pattern, and combining the air conditioning power consumption pattern with the ambient temperature data to form user preference data includes the following steps:

[0162] S221, using historical air conditioning load data as clustering units, and initializing the clustering units;

[0163] S222, based on the initialized clustering units, calculating the covariance of each clustering unit and initializing the similarity matrix;

[0164] S223, selecting two clustering units with the largest similarity values ​​from the similarity matrix, iteratively merging the two clustering units and updating the similarity matrix, and repeating the iterative fusion until all historical air conditioning load data are merged into one unit;

[0165] S225, outputting a hierarchical clustering dendrogram according to the change in harmony of the historical air conditioning load data in the iterative fusion, and selecting a point where the harmony decreases in the dendrogram as a cutting threshold;

[0166] S226. Based on the cutting threshold, the dendrogram is cut into several clusters, the ambient temperature data in the corresponding time period of each cluster is collected, the correlation between the air-conditioning power consumption pattern in the cluster and the ambient temperature data is analyzed, and the user preference data is generated.

[0167] S23. Preprocess the feature data and user preference data to establish a data set.

[0168] It should be noted that the feature data should be standardized or normalized to eliminate the dimensional differences between different features so that the data can be compared and analyzed on the same scale; the user preference data should be encoded or converted into a format suitable for processing by machine learning models, such as converting categorical variables into numerical variables.

[0169] According to the user ID or related fields, the feature data is matched with the user preference data to ensure that each user has corresponding feature data and preference data. If there is a mismatch or missing, appropriate filling or interpolation processing is required to ensure the consistency and integrity of the data. The preprocessed feature data is merged with the user preference data to form a complete data set.

[0170] S24. Based on the data set and combined with the deep learning model, an air conditioning load prediction model is constructed.

[0171] It should be noted that based on the recurrent neural network model, according to the characteristic dimension of the data set, the number of nodes in the input layer is defined, the recurrent unit of the RNN (such as LSTM or GRU) is set, and the number of units in the hidden layer (that is, the size of the hidden state) is determined; after the RNN layer, an additional fully connected layer can be added to extract more advanced features or perform dimensionality transformation; according to the requirements of the prediction task, the number of nodes in the output layer and the format of the output data are determined; the hyperparameters of the recurrent neural network model are set, and the API provided by the deep learning framework is used to build a complete recurrent neural network model, that is, the air conditioning load prediction model, according to the defined input layer, hidden layer and output layer structure.

[0172] S25. Use historical building air-conditioning load data to train an air-conditioning load prediction model, and evaluate the trained air-conditioning load prediction model.

[0173] It should be noted that the historical data is divided into a training set, a validation set, and a test set. The training set is used to train the air-conditioning load prediction model. During the training process, the air-conditioning load prediction model will learn the ability to extract features from the input data and predict the air-conditioning load. Among them, appropriate hyperparameters such as the number of training rounds, batch size, and learning rate are set to ensure that the model can fully learn and avoid overfitting or underfitting. After each training round, the validation set is used to verify the air-conditioning load prediction model, and the performance of the air-conditioning load prediction model is evaluated by calculating the performance indicators on the validation set.

[0174] S26. Optimize the air conditioning load prediction model based on the evaluation results.

[0175] It should be added that the hyperparameters of the model, such as the learning rate and the number of neurons in the hidden layer, are adjusted according to the evaluation results to optimize the performance of the air conditioning load prediction model.

[0176] S3. Use the air conditioning load prediction model to predict the air conditioning load of each floor in the building and obtain the prediction result.

[0177] It should be noted that real-time air-conditioning load data is collected, and the real-time air-conditioning load data, floor correlation and user preference data are preprocessed and converted into a prediction data set suitable for the input of the air-conditioning load prediction model, and the prediction data set is divided into a training set and a test set; according to the characteristic dimension of the prediction data set, the number of nodes in the input layer is determined, the recurrent unit of the RNN is selected, and the number of units in the hidden layer is set; according to the requirements of the prediction task, the number of nodes in the output layer is determined, and after the RNN layer, an additional fully connected layer can be added to extract more advanced prediction data features or perform dimensionality transformation; the API provided by the deep learning framework is used to set the parameters of the model, and the RNN model, i.e., the air-conditioning load prediction model, is constructed according to the defined input layer, hidden layer and output layer; the air-conditioning load prediction model is trained using the training set, the model parameters are updated through the back-propagation algorithm and the gradient descent optimizer, and the test set is used for evaluation; the prediction data set is input into the air-conditioning load prediction model, and the forward propagation calculation is performed to obtain the prediction result.

[0178] S4. Based on the prediction results and combined with the air conditioning power consumption data, the air conditioning load is centrally scheduled.

[0179] It should be noted that, according to the prediction results, combined with the air-conditioning power consumption data, the air-conditioning usage pattern and energy consumption trend of each floor are analyzed, and the scheduling strategy is formulated, including adjusting the opening or closing time and setting temperature of the air-conditioning system, analyzing the balance between comfort and energy efficiency, ensuring that energy consumption is reduced as much as possible while meeting the comfort of the indoor environment, setting priorities, and giving priority to energy-saving scheduling for floors or time periods with high energy consumption and low comfort requirements; for special situations, such as high temperature warnings and changes in building traffic, the scheduling strategy is dynamically adjusted. In the event of a high temperature warning, the set temperature can be appropriately increased or the air-conditioning operation time can be extended to ensure the comfort of the indoor environment. When the building traffic changes greatly, the air-conditioning operation strategy of each floor can be adjusted according to the real-time traffic data; the scheduling strategy is converted into a specific scheduling table, and the air-conditioning operation parameters of each floor and each time period are clarified. The scheduling table should include but is not limited to information such as floor, time period, opening or closing time, and set temperature. The scheduling table should be flexible and operational, and convenient for adjustment according to actual conditions; the formulated scheduling table is imported and applied in the building scheduling platform to ensure that the scheduling platform can automatically adjust the operating parameters of the air-conditioning system to achieve centralized scheduling.

[0180] like Figure 2 As shown, according to another embodiment of the present invention, a building air conditioning load centralized scheduling system is also provided, the system comprising a floor correlation analysis module 1, a prediction model construction module 2, an air conditioning load prediction module 3 and a centralized scheduling module 4;

[0181] The floor correlation analysis module 1, the prediction model building module 2, the air conditioning load prediction module 3 and the centralized dispatching module 4 are connected in sequence;

[0182] The floor correlation analysis module 1 is used to obtain the floor correlation between the floor and the air conditioning load data by analyzing the acquired historical air conditioning load data of the building;

[0183] Prediction model building module 2, used to build an air conditioning load prediction model based on floor correlation and in combination with pre-acquired user preference data;

[0184] The air conditioning load prediction module 3 is used to predict the air conditioning load of each floor in the building by using the air conditioning load prediction model to obtain the prediction result;

[0185] The centralized dispatching module 4 is used to centrally dispatch the air conditioning load based on the prediction results and the air conditioning power consumption data.

[0186] In summary, by means of the above technical solution of the present invention, by analyzing the historical air-conditioning load data of several buildings, the floor correlation between the floors and the air-conditioning load data can be obtained, so as to deeply understand the difference in air-conditioning load between the floors in the building and its changing rules, and provide basic data support for the subsequent construction of the air-conditioning load prediction model. By mining the floor correlation, the characteristics of the air-conditioning load on each floor can be more accurately grasped, providing a strong basis for the subsequent scheduling work, which is helpful to achieve effective electricity conservation and promote the efficiency and sustainability of building energy use. By comprehensively considering the floor correlation and user preferences, an air-conditioning load prediction model that is more in line with the actual situation can be constructed, which not only reflects the personalized needs of different users for air-conditioning use, but also improves the accuracy and reliability of the prediction results, realizes the accurate prediction of the air-conditioning load on each floor in the building, and provides strong guidance for the subsequent air-conditioning load scheduling, helping managers to better grasp the air-conditioning use of the building, and provide decision support for energy saving and consumption reduction, and promote the optimization and sustainable development of building energy management. By analyzing the air-conditioning power consumption data and centrally dispatching the air-conditioning load, we can achieve optimal distribution and reasonable control of the building's air-conditioning load, formulate a more scientific and reasonable dispatching plan, and ensure that the air-conditioning energy consumption is reduced while meeting the user's comfort level. At the same time, centralized dispatching can also help improve the operating efficiency of the building's air-conditioning system, reduce energy waste, and thus achieve the goal of energy conservation and emission reduction, and promote the green and intelligent development of building energy management.

[0187] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A centralized dispatching method for building air conditioning load, characterized in that: The method comprises the following steps: S1. Analyze the historical air conditioning load data of the building to obtain the floor correlation between the floor and the air conditioning load data; S2. Based on the floor correlation and combined with the pre-acquired user preference data, an air conditioning load prediction model is constructed; S3. Using the air conditioning load prediction model, predict the air conditioning load of each floor in the building to obtain a prediction result; S4. Based on the prediction results and combined with the air conditioning power consumption data, the air conditioning load is centrally dispatched; The process of obtaining the floor correlation between the floor and the air conditioning load data by analyzing the historical air conditioning load data of the building comprises the following steps: S11, using the building air conditioning scheduling platform, obtaining the historical air conditioning load data of several buildings and preprocessing them, wherein the historical air conditioning load data includes historical air conditioning power consumption data, air conditioning operation status data, ambient temperature data, building data and floor data; S12. According to the building data, the pre-processed historical air conditioning load data is processed in layers to establish a number of building layered data sets, and the building layered data sets are decomposed into a number of floor layered data sets in combination with the floor data; S13, based on the layered data sets of each floor, analyzing the historical air conditioning load data of the same floor in the same building, and obtaining the air conditioning power consumption data on the same floor; S14, combining the stratified data sets of each building, analyzing the floor correlation between the floors and the air conditioning power consumption data; S15. Based on the floor correlation analysis result, obtain the floor correlation between the floor and the air conditioning load data; The analysis of the floor correlation between the floors and the air-conditioning power consumption data in combination with the layered data sets of each building includes the following steps: S141, based on the air conditioning power consumption data on the same floor, combined with the layered data sets of each building, analyzing the air conditioning power consumption data on the same floor in different buildings; S142, based on the analysis results of the air conditioning power consumption data on the same floor in different buildings, using a building correlation algorithm to calculate the building correlation between the building data and the air conditioning power consumption data; S143, calculating the floor correlation between the air conditioning power consumption data of the same floor and the floor above the same floor in different buildings using a floor correlation algorithm according to the building correlation; The method of calculating the floor correlation between the air conditioning power consumption data of the same floor and the floor above the same floor in different buildings by using the floor correlation algorithm according to the building correlation comprises the following steps: S1431. Obtain building correlation amplitude according to the building correlation, and use the ratio of the building correlation amplitude within a certain period of time to the building correlation amplitude within the overall period of time as the set amplitude for mode conversion; S1432, using the primary scale time window to scan the building correlation amplitude within the overall time, and taking the time window in which the building correlation amplitude exceeds the set amplitude as the time window for mode conversion; S1433, using the time window of the secondary scale to detect the time window of the mode conversion; S1434, repeating steps S1432-S1433 until no new mode conversion can be detected, obtaining the occurrence point of the mode conversion, and recording the time of several occurrence points; S1435, taking the time of each occurrence point as the center, using a preset time window to determine whether a mode conversion occurs within the time neighborhood of each occurrence point, and counting the number of mode conversions at each occurrence point, to form a data set of the number of mode conversions between the same floor and the previous floor in different buildings; S1436. Calculate the floor correlation between the same floor and the floor above the same floor in different buildings based on the data set of the number of mode conversions; The floor correlation calculation formula is: In the formula, R It is expressed as the floor correlation value between the same floor and the floor above the same floor in different buildings; x It is expressed as the number of mode transitions at the occurrence point; y It is expressed as the building correlation value between the building data and the air-conditioning power consumption data; t 1 and t 2 All are expressed as arbitrary time; l a Indicates the upper floor of the same floor in different buildings; l b Indicates the same floor in different buildings.

2. A centralized dispatching method for building air conditioning load according to claim 1, characterized in that: The method of analyzing the historical air conditioning load data of the same floor in the same building based on the layered data set of each floor to obtain the air conditioning power consumption data on the same floor includes the following steps: S131, selecting historical air conditioning load data of the same floor in the same building from the layered data sets of each floor; S132, analyzing the changes of air conditioning power consumption data under different air conditioning operation status data and ambient temperature data according to the historical air conditioning load data of the same floor in the same building; S133. Based on the analysis result of the change of the air-conditioning power consumption data, obtain the air-conditioning power consumption data on the same floor.

3. A centralized dispatching method for building air conditioning load according to claim 1, characterized in that: The method of calculating the building correlation between the building data and the air conditioning power consumption data based on the analysis results of the air conditioning power consumption data on the same floor in different buildings by using the building correlation algorithm comprises the following steps: S1421, scanning the air conditioning power consumption data on the same floor in different buildings to obtain an original data set, and dividing the original data set into a time-level data set according to a time constraint; S1422, based on the set support index value, filter out the air-conditioning power consumption data that meets the preset threshold from the time-level data set, and generate the time-level power consumption data set; S1423, analyzing the time level one data set, obtaining the time level two data set according to the aggregation degree between the air conditioner power consumption data, and generating the time level two power consumption data set in combination with the set support index value; S1424, repeat step S1423, sequentially obtain the time N-level data set and the time N-level electricity consumption data set, until no new time N+1-level data set and time N+1-level electricity consumption data set can be obtained, and all electricity consumption data sets are obtained; S1425. Combining the set support and confidence index values, obtain the building correlation between the building data and the air-conditioning power consumption data at different times.

4. A centralized dispatching method for building air conditioning load according to claim 1, characterized in that: The air conditioning load prediction model is constructed based on the floor association and in combination with the pre-acquired user preference data, and includes the following steps: S21, analyzing the floor correlation and extracting feature data from the floor correlation; S22, using an improved hierarchical clustering algorithm to mine the user's air conditioning power consumption pattern, and combining the air conditioning power consumption pattern with the ambient temperature data to form user preference data; S23, pre-processing the feature data and the user preference data to establish a data set; S24. Based on the data set and combined with the deep learning model, an air conditioning load prediction model is constructed; S25, using historical building air conditioning load data to train an air conditioning load prediction model, and evaluating the trained air conditioning load prediction model; S26. Optimize the air conditioning load prediction model based on the evaluation results.

5. A centralized dispatching method for building air conditioning load according to claim 4, characterized in that: The method of using the improved hierarchical clustering algorithm to mine the user's air conditioning power consumption pattern and combining the air conditioning power consumption pattern with the ambient temperature data to form the user preference data includes the following steps: S221, using historical air conditioning load data as clustering units, and initializing the clustering units; S222, based on the initialized clustering units, calculating the covariance of each clustering unit and initializing the similarity matrix; S223, selecting two clustering units with the largest similarity values ​​from the similarity matrix, iteratively merging the two clustering units and updating the similarity matrix, and repeating the iterative fusion until all historical air conditioning load data are merged into one unit; S225, outputting a hierarchical clustering dendrogram according to the change in harmony of the historical air conditioning load data in the iterative fusion, and selecting a point where the harmony decreases in the dendrogram as a cutting threshold; S226. Based on the cutting threshold, the dendrogram is cut into several clusters, the ambient temperature data in the corresponding time period of each cluster is collected, the correlation between the air-conditioning power consumption pattern in the cluster and the ambient temperature data is analyzed, and the user preference data is generated.

6. A building air conditioning load centralized dispatching system, used to implement the building air conditioning load centralized dispatching method according to any one of claims 1 to 5, characterized in that: The system includes a floor correlation analysis module, a prediction model building module, an air conditioning load prediction module and a centralized dispatching module; The floor correlation analysis module, the prediction model construction module, the air conditioning load prediction module and the centralized scheduling module are connected in sequence; The floor correlation analysis module is used to obtain the floor correlation between the floor and the air conditioning load data by analyzing the acquired historical air conditioning load data of the building; The prediction model building module is used to build an air conditioning load prediction model based on the floor association and in combination with the pre-acquired user preference data; The air conditioning load prediction module is used to predict the air conditioning load of each floor in the building using the air conditioning load prediction model to obtain a prediction result; The centralized dispatching module is used to centrally dispatch the air conditioning load based on the prediction result and the air conditioning power consumption data.

Images