Indoor temperature clustering method and device applied to central heating system

Abstract

This disclosure provides an indoor temperature clustering method and apparatus for centralized heating systems, applicable to the field of precise control of centralized heating systems. The method includes: generating initial room temperature sequence data based on room temperature data acquired from multiple room temperature sensors; constructing room temperature observation vectors of a predetermined duration based on target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; determining a Bayesian information criterion value based on the number of room temperature observation vectors of the predetermined duration, and determining the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as the target number of clusters; performing cluster analysis on the room temperature observation vectors of the predetermined duration according to the target number of clusters, and displaying the clustering results obtained when the cluster analysis converges.

Description

Technical Field

[0001] This disclosure relates to the field of precise control of centralized heating systems, specifically to an indoor temperature clustering method and apparatus for use in centralized heating systems. Background Technology

[0002] Currently, in compliance with legal regulations and with user permission, many centralized heating systems have installed room temperature sensors at the user end to monitor indoor temperatures in real time and adjust the heating system accordingly. However, in actual heating adjustments, simply relying on indoor temperatures reveals that there is no fundamental difference in indoor temperatures among different users, thus reducing the accuracy of heating adjustments. This not only makes it difficult to achieve effective heating but also diminishes the user experience. Summary of the Invention

[0003] In view of this, the present disclosure provides an indoor temperature clustering method and apparatus for centralized heating systems that improves heating regulation accuracy, facilitates effective heating, and enhances user experience.

[0004] One aspect of this disclosure provides an indoor temperature clustering method for a centralized heating system. The method includes: generating initial room temperature sequence data based on room temperature data acquired from multiple room temperature sensors; constructing room temperature observation vectors of a predetermined duration based on target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; determining a Bayesian information criterion value based on the number of room temperature observation vectors of the predetermined duration, and determining the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as the target number of clusters; performing cluster analysis on the room temperature observation vectors of the predetermined duration according to the target number of clusters, and displaying the clustering results obtained when the cluster analysis converges.

[0005] According to embodiments of this disclosure, the predetermined duration includes multiple sub-durations; the construction of a room temperature observation vector of the predetermined duration based on the target room temperature sequence data includes: processing the target room temperature sequence data to obtain multiple sub-duration room temperature data in units of the sub-durations; in response to the presence of missing sub-duration room temperature data in the multiple sub-duration room temperature data, filling in the missing values ​​of the missing sub-duration room temperature data; and generating the room temperature observation vector of the predetermined duration using the filled multiple sub-duration room temperature data.

[0006] According to an embodiment of this disclosure, the above-mentioned processing of the target room temperature sequence data to obtain multiple sub-duration room temperature data in units of the above-mentioned sub-duration includes: constructing a sub-duration room temperature function based on the number of the above-mentioned sub-durations and the number of room temperature data in each sub-duration; and using the above-mentioned sub-duration room temperature function to process the target room temperature sequence data to obtain the above-mentioned sub-duration room temperature data.

[0007] According to an embodiment of this disclosure, the sub-durations have a time order; the missing value filling for the missing sub-duration room temperature data includes: sorting multiple sub-duration room temperature data according to the time order to obtain a sorting result; determining the target position of the missing sub-duration room temperature data in the sorting result; and filling the missing value with sub-duration room temperature data of a preset number of sub-durations preceding the target position.

[0008] According to an embodiment of this disclosure, the above-mentioned method of generating a room temperature observation vector of a predetermined duration using multiple sub-duration room temperature data after filling includes: calling standard room temperature data; generating a deviation degree corresponding to each of the sub-duration room temperature data based on the standard room temperature data and each of the multiple sub-duration room temperature data after filling; and generating a room temperature observation vector of the predetermined duration based on the deviation degree corresponding to each of the sub-duration room temperature data.

[0009] According to an embodiment of this disclosure, the above-mentioned cluster analysis of the room temperature observation vector of the predetermined duration based on the target number of clusters includes: constructing a Gaussian mixture model function based on the target number of clusters, the weight coefficients of the Gaussian distribution, and the characteristic data of the Gaussian distribution; inputting the room temperature observation vector of the predetermined duration into the Gaussian mixture model function, and performing iterative cluster analysis until the cluster analysis converges.

[0010] According to embodiments of this disclosure, the characteristic data of the Gaussian distribution includes the mean and covariance of the Gaussian distribution. The convergence of the clustering analysis is determined as follows: initial values ​​are assigned to the weight coefficients, mean, and covariance of the Gaussian distribution; based on the results obtained by inputting the weight coefficients, mean, and covariance of the Gaussian distribution into the Gaussian mixture model function, a posterior probability of the room temperature vector for a predetermined duration is constructed; based on the posterior probability, the maximum likelihood function value of the mean, the maximum likelihood function value of the covariance, and the maximum likelihood function value of the weight coefficients of the Gaussian distribution are generated; and if the convergence judgment functions constructed based on the maximum likelihood function of the mean, the maximum likelihood function value of the covariance, and the maximum likelihood function value of the weight coefficients satisfy a predetermined convergence threshold, the convergence result of the clustering analysis is obtained.

[0011] According to embodiments of this disclosure, the Gaussian mixture model includes the maximum likelihood function of the model; determining the Bayesian information criterion value based on the number of room temperature observation vectors over the predetermined duration includes: constructing a Bayesian information criterion function based on the number of room temperature observation vectors over the predetermined duration, the maximum likelihood function of the model, and the cluster number parameter; iterating over different cluster number parameters within a predetermined numerical range, and generating the Bayesian information criterion value corresponding to the different cluster number parameters based on the Bayesian information criterion function.

[0012] According to embodiments of this disclosure, the preprocessing includes at least one of the following: in response to the presence of a constant value in the initial room temperature sequence data, deleting data collected by a room temperature acquisition sensor corresponding to the constant value; in response to the ratio between the number of lost room temperature data and the total number of room temperature data in the initial room temperature sequence data exceeding a predetermined ratio, deleting room temperature data collected by a room temperature acquisition sensor corresponding to the lost room temperature data; in response to the presence of abnormal data outside a predetermined normal distribution interval in the initial room temperature sequence data, deleting the abnormal data; and in response to the presence of deleted data in the initial room temperature sequence data, filling it with the average room temperature data over a unit time period.

[0013] Another aspect of this disclosure provides an indoor temperature clustering device for a centralized heating system. The device includes: a temperature acquisition sensor for acquiring room temperature data; a generation module for generating initial room temperature sequence data based on room temperature data acquired from multiple room temperature acquisition sensors; a construction module for constructing room temperature observation vectors of a predetermined duration based on target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; a determination module for determining a Bayesian information criterion value based on the number of room temperature observation vectors of the predetermined duration, and determining the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as the target number of clusters; and a clustering module for performing cluster analysis on the room temperature observation vectors of the predetermined duration according to the target number of clusters, and displaying the clustering results obtained when the cluster analysis converges.

[0014] According to embodiments of this disclosure, initial room temperature sequence data is generated based on room temperature data; target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data, and a room temperature observation vector of a predetermined duration is constructed; the Bayesian information criterion value and the target number of clusters are determined according to the number of room temperature observation vectors; cluster analysis is performed on the room temperature observation vector of the predetermined duration according to the target number of clusters, and the results of the cluster analysis are displayed. Since a room temperature observation vector of a predetermined duration, such as a 24-hour room temperature observation vector, is generated based on the room temperature data during the analysis, it can be analyzed from the perspective of daily room temperature variation patterns. Furthermore, cluster analysis is performed using the target number of clusters determined according to the Bayesian information criterion value to deeply analyze the daily variation patterns of room temperature, thereby providing a basis for the heating regulation of the heating system. This at least partially overcomes the problem of low heating regulation accuracy in related technologies, thus achieving the technical effects of improving heating regulation accuracy, realizing effective heating, and improving user experience. Attached Figure Description

[0015] The above and other objects, features and advantages of this disclosure will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0016] Figure 1 An exemplary system architecture of an indoor temperature clustering method and apparatus for a centralized heating system according to embodiments of the present disclosure is illustrated schematically;

[0017] Figure 2 A flowchart illustrating an indoor temperature clustering method applied to a centralized heating system according to an embodiment of the present disclosure is shown schematically.

[0018] Figure 3 A flowchart illustrating cluster analysis according to embodiments of the present disclosure is shown schematically;

[0019] Figure 4 A schematic diagram illustrating clustering analysis results according to embodiments of the present disclosure is shown.

[0020] Figure 5 A flowchart illustrating an indoor temperature clustering method applied to a centralized heating system according to another embodiment of the present disclosure is shown schematically.

[0021] Figure 6 This schematically illustrates the architecture of a clustering system for daily indoor temperature variation patterns applied to a centralized heating system according to an embodiment of the present disclosure; and

[0022] Figure 7 A block diagram of an indoor temperature clustering device applied to a centralized heating system according to an embodiment of the present disclosure is shown schematically. Detailed Implementation

[0023] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0024] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0025] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0026] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0027] In the embodiments of this disclosure, the collection, updating, analysis, processing, use, transmission, provision, disclosure, and storage of data (e.g., including but not limited to user personal information) comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures have been taken to prevent unauthorized access to user personal information data and to safeguard user personal information security, network security, and national security. In the embodiments of this disclosure, user authorization or consent has been obtained before acquiring or collecting user personal information.

[0028] Currently, in compliance with legal regulations and with user permission, many centralized heating systems have installed room temperature sensors at the user end to monitor indoor temperatures in real time. These sensors are not installed on all users; instead, a subset of users is selected to infer the overall heating situation of the centralized heating system from the limited room temperature data. Typically, 2-10% of users are chosen, and the sensors are selectively installed on the top floor, ground floor, and edges of buildings. Common data upload intervals are 15 minutes, 20 minutes, 30 minutes, or 1 hour, resulting in a large amount of indoor temperature data.

[0029] The purpose of collecting indoor temperature data is twofold: first, to understand the heating status of the centralized heating system in order to address various user demands; and second, to apply it to the regulation and control of the centralized heating system. By analyzing the distribution and variation patterns of room temperature data, more precise control directions and objectives can be determined, enabling a more accurate match between heat supply and demand. Preliminary investigations and analyses of multiple room temperature data points revealed significant fluctuations in room temperature throughout the heating season, with inconsistent fluctuation trends among different users. Regarding influencing factors, there is no clear linear relationship between indoor and outdoor temperatures; indoor temperature is influenced by a combination of weather conditions, heating parameters, building structure, and the activities of heat users.

[0030] In heating regulation, it is difficult to achieve precise heating control by simply adjusting the indoor temperature of users. To achieve precise heating temperature control, further analysis of the indoor temperature is needed to obtain the distribution pattern of the indoor temperature. The inventors discovered that, from the perspective of heat users, there is no essential difference in the room temperature of different users. Therefore, further analysis of the room temperature distribution pattern is difficult to conduct from the perspective of heat users. However, by analyzing the daily variation pattern of room temperature, the accuracy of heating regulation can be improved, thereby achieving effective heating and improving the user experience.

[0031] In view of this, embodiments of this disclosure provide an indoor temperature clustering method and apparatus for use in centralized heating systems to achieve effective heating. Specifically, the method includes: generating initial room temperature sequence data based on room temperature data acquired from multiple room temperature acquisition sensors; constructing room temperature observation vectors of a predetermined duration based on target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; determining a Bayesian information criterion value based on the number of room temperature observation vectors of the predetermined duration, and determining the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as the target number of clusters; performing cluster analysis on the room temperature observation vectors of the predetermined duration according to the target number of clusters, and displaying the clustering results obtained when the cluster analysis converges.

[0032] Figure 1 An exemplary system architecture 100 of an indoor temperature clustering method and apparatus for a centralized heating system according to embodiments of this disclosure is illustrated. It should be noted that... Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.

[0033] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, a server 105, and a room temperature sensor 106. The network 104 serves as a medium for providing communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105, and between the server 105 and the room temperature sensor 106. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0034] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, such as receiving requests for indoor temperature cluster analysis or receiving the results of indoor temperature cluster analysis. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as heating regulation applications, applications for processing indoor temperature data, shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social media platforms (for example only).

[0035] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0036] The room temperature sensor 106 can be used to collect the indoor temperature of a user.

[0037] Server 105 can be a server that provides various services, such as a backend management server that supports requests sent by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process the received user requests and other data, such as acquiring the indoor temperature from the room temperature sensor 106 and performing cluster analysis on the indoor temperature, and feeding back the processing results (such as cluster analysis results, web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0038] It should be noted that the indoor temperature clustering method for centralized heating systems provided in this disclosure can generally be executed by server 105. Correspondingly, the indoor temperature clustering device for centralized heating systems provided in this disclosure can generally be installed in server 105. The indoor temperature clustering method for centralized heating systems provided in this disclosure can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, the room temperature sensor 106, and / or server 105. Correspondingly, the indoor temperature clustering device for centralized heating systems provided in this disclosure can also be installed in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, the room temperature sensor 106, and / or server 105. Alternatively, the indoor temperature clustering method for centralized heating systems provided in this disclosure can also be executed by the first terminal device 101, the second terminal device 102, or the third terminal device 103, or by other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103. Correspondingly, the indoor temperature clustering device for centralized heating systems provided in this disclosure can also be installed in the first terminal device 101, the second terminal device 102, or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103.

[0039] For example, the data collected by the room temperature sensor 106 can be stored in any one of the first terminal device 101, the second terminal device 102, or the third terminal device 103 (e.g., the first terminal device 101, but not limited thereto), or stored on an external storage device and imported into the first terminal device 101. Then, the first terminal device 101 can locally execute the indoor temperature clustering method for centralized heating systems provided in this disclosure embodiment, or send the room temperature data to other terminal devices, servers, or server clusters, and have the other terminal devices, servers, or server clusters that receive the room temperature data execute the indoor temperature clustering method for centralized heating systems provided in this disclosure embodiment.

[0040] It should be understood that Figure 1 The number of terminal devices, networks, servers, and room temperature sensors shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, servers, and room temperature sensors can be included.

[0041] Figure 2 A flowchart illustrating an indoor temperature clustering method applied to a centralized heating system according to an embodiment of the present disclosure is shown.

[0042] like Figure 2 As shown, the method includes operations S201~S204. It is understood that this method can be run on a schedule, i.e., once a day, to perform clustering processing on the daily variation patterns of room temperature data collected during the current heating season up to the operating time.

[0043] In operation S201, initial room temperature sequence data is generated based on room temperature data acquired from multiple room temperature acquisition sensors.

[0044] In operation S202, a room temperature observation vector of a predetermined duration is constructed based on the target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data.

[0045] In operation S203, based on the number of room temperature observation vectors over a predetermined duration, the Bayesian information criterion value is determined, and the number of model parameters corresponding to the minimum value of the Bayesian information criterion value is determined as the target cluster number.

[0046] In operation S204, cluster analysis is performed on the room temperature observation vectors for a predetermined duration based on the target number of clusters, and the clustering results obtained when the cluster analysis converges are displayed.

[0047] Optionally, embodiments of this disclosure can uncover hidden patterns by analyzing room temperature data over a long historical period. By acquiring room temperature data from several room temperature sensors, initial room temperature sequence data can be generated, such as time-series data of the initial room temperature. The time-series data of the initial room temperature can be represented as... The unit of measurement for this sequence data can be degrees Celsius (°C). Here, m can be the number of sensors for room temperature data acquisition, and p can be the time step, such as one day, one week, one month, or multiple months, which can be adaptively adjusted according to actual needs.

[0048] Optionally, the initial room temperature sequence data may contain various anomalous room temperature data, such as data loss due to communication interruptions or room temperature fluctuations caused by sudden window opening. Including these anomalous room temperature data in the analysis may affect the accuracy of the results. Therefore, it is necessary to process the anomalous room temperature data, i.e., preprocess the initial room temperature sequence data to obtain target room temperature sequence data free of anomalous room temperature data.

[0049] Optionally, based on the target room temperature sequence data, a room temperature observation vector of a predetermined duration can be constructed, such as a 24-hour room temperature observation vector. The predetermined duration can be adaptively adjusted according to actual needs. For example, when analyzing the diurnal variation pattern of room temperature, the predetermined duration can be set to 24 hours, and when analyzing the hourly variation pattern of room temperature, the predetermined duration can be set to 1 hour, etc.

[0050] Optionally, the Bayesian Information Criterion (BIC) can be a statistical tool for model selection, helping researchers or data analysts determine which model is most suitable when faced with multiple possible models. In this embodiment, the BIC value can be used to determine the optimal number of clusters, i.e., the target number of clusters, when performing cluster analysis using a Gaussian Mixture Model (GMM). That is, the number of clusters corresponding to the minimum BIC value can be considered the target number of clusters.

[0051] Optionally, based on the target number of clusters determined by the BIC value, a Gaussian Mixture Model (GMM) can be used for cluster analysis. A GMM is a statistical model that assumes the data is a superposition of several Gaussian distributions. This model considers not only the mean but also the covariance of the data, thus better capturing the complex structure of the data. GMM can be used for cluster analysis, i.e., to group data points, each group described by a Gaussian distribution. This approach is more flexible than traditional clustering methods such as K-Means because it allows for different variances in each group, thus better adapting to the distribution of the data.

[0052] Optionally, cluster analysis can be performed on the room temperature observation vectors for a predetermined duration using GMM, and the results can be presented in the form of clusters. These cluster analysis results can be displayed to relevant personnel in the heating system, enabling them to adjust the heating supply based on the results and achieve effective heating.

[0053] According to embodiments of this disclosure, initial room temperature sequence data is generated based on room temperature data; target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data, and a room temperature observation vector of a predetermined duration is constructed; the Bayesian information criterion value and the target number of clusters are determined according to the number of room temperature observation vectors; cluster analysis is performed on the room temperature observation vector of the predetermined duration according to the target number of clusters, and the results of the cluster analysis are displayed. Since a room temperature observation vector of a predetermined duration, such as a 24-hour room temperature observation vector, is generated based on the room temperature data during the analysis, it can be analyzed from the perspective of daily room temperature variation patterns. Furthermore, cluster analysis is performed using the target number of clusters determined according to the Bayesian information criterion value to deeply analyze the daily variation patterns of room temperature, thereby providing a basis for the heating regulation of the heating system. This at least partially overcomes the problem of low heating regulation accuracy in related technologies, thus achieving the technical effects of improving heating regulation accuracy, realizing effective heating, and improving user experience.

[0054] Optionally, the process of preprocessing the initial room temperature sequence data to obtain the target room temperature sequence data mentioned in operation S202 above may include the following operations: in response to the presence of a constant value in the initial room temperature sequence data, deleting the data collected by the room temperature acquisition sensor corresponding to the constant value; in response to the ratio between the number of lost room temperature data and the total number of room temperature data in the initial room temperature sequence data exceeding a predetermined ratio, deleting the room temperature data collected by the room temperature acquisition sensor corresponding to the lost room temperature data; in response to the presence of abnormal data in the initial room temperature sequence data outside a predetermined normal distribution interval, deleting the abnormal data; and in response to the presence of deleted data in the initial room temperature sequence data, filling it with the average value of room temperature data within a unit time period.

[0055] Optionally, some indoor temperature data may remain constant throughout the acquisition period, which is usually due to sensor malfunction. For such data, all data from the entire room temperature acquisition sensor will be discarded.

[0056] Optionally, indoor temperature data may be lost due to communication interruption or damage to sensor accessories. Room temperature data collected by sensors whose ratio of lost room temperature data to total room temperature data exceeds a predetermined value (e.g., the number of days with lost room temperature data exceeds 1 / 3 of the total number of days with room temperature data collected) will be discarded entirely.

[0057] Optionally, the 3Sigma criterion based on statistical principles can be used to detect anomalies in the room temperature data collected by the room temperature sensor, and the detected anomalies can be discarded directly.

[0058] The 3Sigma criterion, also known as outlier detection based on normal distribution, can identify and remove outliers containing gross errors. It assumes a historical data sequence... The average value is The standard deviation is Then the formula for calculating 3Sigma can be shown in formulas (1) to (3).

[0059] (1)

[0060] (2)

[0061] (3)

[0062] The data has the following characteristics: the probability of a data value falling within the range of (μ-σ, μ+σ) is 68.27%; the probability of a data value falling within the range of (μ-2σ, μ+2σ) is 95.44%; the probability of a data value falling within the range of (μ-3σ, μ+3σ) is 99.73%, meaning the probability of a data value falling outside the range of (μ-3σ, μ+3σ) is 0.27%.

[0063] The principle of outlier detection based on the 3Sigma criterion is that if a data value falls outside the range (μ-3σ, μ+3σ), it is an outlier. That is, 99.73% of the data are normal values, and 0.27% are outliers. Outliers can be deleted.

[0064] Alternatively, if data is lost at some point in the day, including data lost during the collection process or data discarded due to anomalies detected using the 3Sigma criterion, the average room temperature data over a unit of time, such as the average daily room temperature data, can be used to fill the gaps.

[0065] According to embodiments of this disclosure, by preprocessing the initial room temperature sequence data, outliers in the initial room temperature sequence data can be identified, thereby improving the accuracy of cluster analysis and the precision of heating regulation, so as to achieve effective heating and improve the user experience.

[0066] Optionally, the predetermined duration described above may include multiple sub-durations, such as a 24-hour room temperature observation vector, which may be composed of multiple 1-hour room temperature data. In one embodiment, the process of constructing a room temperature observation vector of a predetermined duration based on target room temperature sequence data as described in operation S202 above may include the following operations: processing the target room temperature sequence data to obtain multiple sub-duration room temperature data in units of sub-durations; in response to the existence of missing sub-duration room temperature data in the multiple sub-duration room temperature data, filling in the missing values ​​of the missing sub-duration room temperature data; and generating a room temperature observation vector of a predetermined duration using the filled multiple sub-duration room temperature data.

[0067] Optionally, the process of processing the target room temperature sequence data to obtain multiple sub-time-length room temperature data in units of sub-time length may include the following operations: constructing a sub-time-length room temperature function based on the number of sub-time lengths and the number of room temperature data in each sub-time length; and using the sub-time-length room temperature function to process the target room temperature sequence data to obtain the sub-time-length room temperature data.

[0068] Optionally, since a centralized heating system may deploy room temperature sensors from multiple manufacturers with inconsistent acquisition periods, the length of daily data sequences may vary, making it difficult to construct a complete 24-hour observation vector. Therefore, it is necessary to convert each acquired room temperature data sequence into an hourly room temperature data value, calculated using an average value.

[0069] Optionally, the target room temperature sequence data can be processed using a sub-duration room temperature function, which can be shown in formula (4).

[0070] (4)

[0071] Where, x ij Let q be the room temperature data sequence, where n can be the number of sub-durations, i.e., the number of hours, and q can be the number of room temperature data points in each sub-duration, i.e., the number of room temperature data points in each hour.

[0072] Optionally, the sub-hours can have a temporal order. For example, in a 24-hour day, a sub-hour can be a 1-hour duration, such as 1 hour from 0:00 to 1:00, 1 hour from 1:00 to 2:00, ..., 1 hour from 23:00 to 0:00; each of these 1-hour durations can have a temporal order determined by the time point. When the sub-hours have a temporal order, the process described above for filling in missing sub-hour room temperature data in response to the presence of missing sub-hour room temperature data in multiple sub-hour room temperature data can include the following operations: sorting the multiple sub-hour room temperature data according to the temporal order to obtain a sorting result; determining the target position of the missing sub-hour room temperature data in the sorting result; and filling the missing value with sub-hour room temperature data of a predetermined number of sub-hours preceding the target position.

[0073] Optionally, the process of filling missing values ​​described above can be a process of filling missing values ​​using forward interpolation. For all sub-hourly room temperature data obtained based on the aforementioned operations, the determined room temperature data sequence is as follows: Where n is the number of hours and m is the number of room temperature acquisition sensors. Due to data omissions or data preprocessing discarding, some sub-hours of room temperature data may be missing values. This embodiment of the present disclosure can use forward interpolation to complete the missing data, that is, fill the missing values ​​of the current hour with the values ​​of the previous hour.

[0074] Specifically, the room temperature data of each sub-hour can be sorted according to the time sequence. For missing room temperature data of each sub-hour, the position of the missing room temperature data of each sub-hour can be determined. This position can be used as the target position. The room temperature data of each sub-hour that is located before the target position (such as the room temperature data of the first hour) can be filled into the target position as the complete room temperature data of each sub-hour.

[0075] Optionally, if there are no missing sub-time room temperature data among the multiple sub-time room temperature data, the operation of generating a room temperature observation vector of a predetermined duration can be performed directly.

[0076] Optionally, the process of generating a room temperature observation vector of a predetermined duration based on multiple sub-duration room temperature data may include the following operations: calling standard room temperature data; generating the degree of deviation corresponding to each sub-duration room temperature data based on the standard room temperature data and the room temperature data of each sub-duration room temperature data after filling; and generating a room temperature observation vector of a predetermined duration based on the degree of deviation corresponding to each sub-duration room temperature data.

[0077] Optionally, the standard room temperature data can be preset and can be adaptively adjusted according to actual needs.

[0078] Optionally, extract hourly measurements of room temperature from X1 to X during the day. 24 Define an initial 24-hour room temperature observation vector, which can be expressed as shown in formula (5).

[0079] (5)

[0080] Since room temperature is mostly within a specific range and fluctuates relatively little, it is relatively difficult to observe. Therefore, the signal strength of each room temperature data can be amplified by using the degree of deviation between the measured value and the standard value of room temperature. The amplification formula can be shown in formula (6).

[0081] (6)

[0082] in, X represents the degree of deviation in room temperature data for each sub-hour. t This is the standard value for room temperature.

[0083] The 24-hour room temperature observation vector determined based on the degree of deviation can be shown in formula (7).

[0084] (7)

[0085] According to embodiments of this disclosure, by constructing a 24-hour room temperature observation vector, it is convenient to analyze the daily variation pattern of room temperature, which is beneficial to the heating system's heating regulation, achieving effective heating and improving the user experience.

[0086] Optionally, before performing cluster analysis on the 24-hour room temperature observation vectors, the optimal number of clusters can be determined first. In this embodiment of the disclosure, the optimal number of clusters can be determined based on the BIC value. Therefore, the process of determining the Bayesian information criterion value based on the number of room temperature observation vectors of a predetermined duration, as mentioned in operation S203 above, can include the following operations: constructing a Bayesian information criterion function based on the number of room temperature observation vectors of a predetermined duration, the maximum likelihood function of the GMM model, and the cluster number parameter; iterating over different cluster number parameters within a predetermined numerical range, and generating Bayesian information criterion values ​​corresponding to different cluster number parameters based on the Bayesian information criterion function.

[0087] Optionally, in this embodiment of the disclosure, the number of groups in the GMM is probabilistically estimated using BIC-based model selection theory, and the optimal number of clusters is obtained step by step through approximation. The Bayesian information criterion function can be as shown in formula (8).

[0088] (8)

[0089] In the formula, K is the cluster number parameter, i.e. the number of Gaussian clusters; N is the number of samples, such as the number of room temperature observation vectors for a predetermined duration; and L is the maximum likelihood function of the GMM model.

[0090] When determining the target number of clusters, i.e., the optimal number of clusters, models with low BIC values ​​are preferred. The specific calculation process involves iterating over different values ​​within a predetermined numerical range (e.g., from 1 to 30). The value is calculated for each iteration's BIC. The minimum BIC is... The value represents the optimal number of clusters, which is the target number of clusters.

[0091] Alternatively, the number of target clusters determined based on the Bayesian information criterion value can be used for cluster analysis using a GMM model.

[0092] Figure 3 A flowchart illustrating cluster analysis according to an embodiment of the present disclosure is shown schematically.

[0093] like Figure 3 As shown, the cluster analysis may include operations S301 to S307.

[0094] In operation S301, initial values ​​are assigned to the weight coefficients, mean, and covariance of the Gaussian distribution.

[0095] In operation S302, the posterior probability is calculated.

[0096] In operation S303, solve for the maximum likelihood function of the mean of the Gaussian distribution.

[0097] In operation S304, solve for the maximum likelihood function value of the covariance of the Gaussian distribution.

[0098] In operation S305, solve for the maximum likelihood function of the weight coefficients of the Gaussian distribution.

[0099] In operation S306, it is determined whether the convergence threshold is met. If the convergence threshold is met, operation S307 can be executed. If the convergence threshold is not met, the process can be repeated starting from operation S302.

[0100] When operating S307, the clustering results are output.

[0101] like Figure 3As shown, in one embodiment, the process of performing cluster analysis on the room temperature observation vector of a predetermined duration based on the target number of clusters described in operation S204 may include the following operations: constructing a Gaussian mixture model function based on the target number of clusters, the weight coefficients of the Gaussian distribution, and the characteristic data of the Gaussian distribution; inputting the room temperature observation vector of a predetermined duration into the Gaussian mixture model function, and performing iterative cluster analysis until the cluster analysis converges.

[0102] Optionally, clustering involves grouping data objects with high similarity into the same cluster and data objects with high dissimilarity into different clusters based on the principle of similarity. The clustering process is unsupervised, meaning that there is no prior knowledge about the data objects to be processed.

[0103] GMM is a probabilistic clustering method that assumes that the input samples follow a known Gaussian distribution of K clusters, and samples that follow the same distribution are clustered into one class. The core theorem of GMM is its probability density function. In one embodiment, the function of the GMM model can be shown as Equation (9).

[0104] (9)

[0105] Where x is the input data to be clustered, namely the aforementioned 24-hour room temperature observation vector, and the sample size is N; K is the number of clusters, i.e. the number of Gaussian clusters; The weight coefficients of the k-th Gaussian distribution satisfy the following condition: and ; It is the k-th Gaussian distribution with a mean of The covariance is .

[0106] The GMM above can have three parameters that need to be calculated to determine whether the cluster analysis has converged, namely the characteristic data of the Gaussian distribution, namely the mean. Weighting coefficients and covariance The Expectation-Maximization (EM) algorithm is used to calculate and determine whether convergence has occurred. EM is an iterative optimization algorithm.

[0107] In one embodiment, cluster analysis convergence is determined as follows: initial values ​​are assigned to the weight coefficients, mean, and covariance of the Gaussian distribution; based on the results obtained by inputting the weight coefficients, mean, and covariance of the Gaussian distribution into the Gaussian mixture model function, a posterior probability of the room temperature vector for a predetermined duration is constructed; according to the posterior probability, the maximum likelihood function of the mean, the maximum likelihood function of the covariance, and the maximum likelihood function of the weight coefficients of the Gaussian distribution are generated; if the convergence judgment functions constructed based on the maximum likelihood function values ​​of the mean, covariance, and weight coefficients of the Gaussian distribution satisfy a predetermined convergence threshold, the cluster analysis convergence result is obtained.

[0108] Optionally, initial values ​​are assigned to the weight coefficients of the Gaussian distribution, the mean of the Gaussian distribution, and the covariance of the Gaussian distribution, respectively, i.e., specified. The initial value can be adaptively adjusted according to actual needs, and this process can correspond to the above operation S301.

[0109] Calculate each data point based on formula (9) Posterior probability belonging to each Gaussian component k As shown in formula (10), this process can correspond to the above operation S302.

[0110] (10)

[0111] Solving based on posterior probability The maximum likelihood function is shown in Equation (11), and this process can correspond to the above operation S303.

[0112] (11)

[0113] in, It is the number of valid samples assigned to the k-th component.

[0114] Solving based on posterior probability The maximum likelihood value is shown in Equation (12), and this process can correspond to the above operation S304.

[0115] (12)

[0116] Solving based on posterior probability The maximum likelihood function is shown in Equation (13), and this process can correspond to the above operation S305.

[0117] (13)

[0118] If satisfied , and (in To determine the convergence threshold, it can be adaptively adjusted according to actual needs. This process corresponds to operation S306 above. The values ​​are respectively taken That is, the kth time The value; otherwise, the formulas (10) to (13) are executed repeatedly and the following steps are performed. , and The judgment is to repeatedly execute operations S302 to S306 until the calculation converges.

[0119] Optionally, the GMM model calculation process considers the weights of each category, and its clustering effect on imbalanced sample categories is better than that of the kmean model or other models. GMM is a generative model that can calculate the joint probability distribution from the distribution of the samples themselves to obtain the classification result. The interpretability of the model is better than that of the kmean model, a discriminative model.

[0120] Figure 4 A schematic diagram illustrating the clustering analysis results according to an embodiment of the present disclosure is shown.

[0121] like Figure 4 As shown, the horizontal axis represents time, and the vertical axis represents the degree of deviation. Figure 4 The daily room temperature curves for each cluster are shown for holidays and weekdays, with the thick line representing the centroid and the thin line representing representative detailed data. Cluster analysis reveals that cluster 1 exhibits relatively small fluctuations on both weekdays and holidays. Cluster 2's curve is second only to cluster 4 in steepness compared to the other clusters. Cluster 3 has the smoothest curve among the four clusters. Cluster 4's curve shows significant amplitude during both its rise and fall, indicating steep rises and falls. Among these four clusters, based on the degree of deviation between indoor temperature and standard room temperature data, although cluster 2's indoor temperature is closer to the standard room temperature, it is not comfortable for heat users. Cluster 3 has the smallest fluctuations, but its indoor temperature is much higher than the standard room temperature, which could easily lead to higher energy consumption. Cluster 4 is far below the target value. Therefore, cluster 1 is the optimal target in practical operation.

[0122] This embodiment of the disclosure preprocesses the raw room temperature data to remove outliers; for multiple room temperature data points from the same acquisition device within one hour, the average value is calculated to obtain hourly room temperature data; a 24-hour observation vector is constructed based on the hourly room temperature data; the sum of squared errors method is used to determine the number of clusters; and a Gaussian mixture model (GMM) is used for clustering. This facilitates a deeper understanding of the daily variation patterns of room temperature, provides a foundation for heating regulation, and supports effective heating and a better user experience.

[0123] Figure 5 A flowchart illustrating an indoor temperature clustering method applied to a centralized heating system according to another embodiment of the present disclosure is shown.

[0124] like Figure 5 As shown, the process may include operations S501 to S509.

[0125] This embodiment's process can employ a timed execution strategy, running once daily to cluster diurnal variation patterns in room temperature data collected during the current heating season up to the execution time. After obtaining the room temperature data, preprocessing is performed, including detecting outliers and missing values. Then, hourly room temperature values ​​are calculated, and missing hourly room temperatures are filled in. Finally, a 24-hour observation vector is constructed. The target cluster number is calculated, and the 24-hour observation vector is clustered. The clustering results are displayed visually. The display method is as follows: Figure 4 As shown, the time is divided into weekdays and holidays, and then clustered into 4 clusters. The thick lines in the figure represent the centroids of the clusters, and the thin lines represent the detailed data.

[0126] Specifically, the process may include operations S501 to S509.

[0127] Use the S501 to obtain room temperature data.

[0128] In operation S502, data preprocessing.

[0129] Calculate the hourly room temperature while operating S503.

[0130] In operation S504, fill in the missing hourly room temperature values.

[0131] Using the S505, construct a 24-hour room temperature observation vector.

[0132] In operation S506, calculate the number of target clusters.

[0133] Cluster analysis was performed on the 24-hour room temperature observation vector using the S507.

[0134] The clustering results are displayed when operating S508.

[0135] In operation S509, determine whether to cycle to the next period. If cycling is determined, operations S501 through S509 can be executed repeatedly starting from a previous period; if cycling is determined not to cycle, the process can be terminated directly. Optionally, determining whether to cycle to the next period can be based on actual circumstances. For example, if cluster analysis needs to be repeated, cycling is necessary; otherwise, cycling is not required. Another example is performing cluster analysis once a day. After the daily cluster analysis is completed, cycling may not be necessary; if the daily cluster analysis is not completed, cycling can continue.

[0136] Optionally, the above operations S501 to S509 can be referred to operations S201 to S204, and will not be repeated here.

[0137] Embodiments of this disclosure also provide a clustering system for daily indoor temperature variation patterns applied to centralized heating systems, which can, as follows: Figure 6 As shown.

[0138] Figure 6 The diagram illustrates the architecture of a clustering system for daily indoor temperature variation patterns applied to a centralized heating system according to an embodiment of the present disclosure.

[0139] like Figure 6 As shown, the system can be composed of a room temperature acquisition sensor 601, a room temperature acquisition system 602, and a room temperature analysis system 603.

[0140] The room temperature acquisition sensor 601 can consist of several sensors deployed in the room of the heat user, responsible for collecting the user's indoor temperature and uploading it to the room temperature acquisition system 602 at regular intervals via a communication network.

[0141] The room temperature acquisition system 602 is responsible for receiving room temperature data and storing the room temperature data according to a certain logical structure.

[0142] The room temperature analysis system 603 can preprocess room temperature data, construct a 24-hour observation vector, perform clustering, and finally display the clustering results.

[0143] This disclosure proposes a clustering method and system for the daily indoor temperature variation pattern in centralized heating systems. Based on room temperature data acquired from room temperature sensors, a 24-hour observation vector is constructed, and then the vector is analyzed using a clustering method. This facilitates a deeper understanding of the daily room temperature variation pattern, provides a foundation for heating regulation, and supports effective heating and improved user experience.

[0144] It should be noted that, unless it is explicitly stated that there is a sequential order of execution between different operations, or that there is a sequential order of execution between different operations in terms of technical implementation, the execution order between multiple operations may not be significant, and multiple operations may be executed simultaneously.

[0145] Figure 7 A block diagram of an indoor temperature clustering device applied to a centralized heating system according to an embodiment of the present disclosure is shown schematically.

[0146] like Figure 7 The indoor temperature clustering device 700 shown for use in a centralized heating system may include a temperature acquisition sensor 710, a generation module 720, a construction module 730, a determination module 740, and a clustering module 750.

[0147] Temperature sensor 710 is used to collect room temperature data.

[0148] The generation module 720 is used to generate initial room temperature sequence data based on room temperature data acquired from multiple room temperature acquisition sensors.

[0149] Module 730 is used to construct a room temperature observation vector of a predetermined duration based on the target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data.

[0150] The determination module 740 is used to determine the Bayesian information criterion value based on the number of room temperature observation vectors for a predetermined duration, and to determine the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as the target number of clusters.

[0151] Clustering module 750 is used to perform cluster analysis on room temperature observation vectors for a predetermined duration based on the target number of clusters, and to display the clustering results obtained when the cluster analysis converges.

[0152] According to embodiments of this disclosure, initial room temperature sequence data is generated based on room temperature data; target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data, and a room temperature observation vector of a predetermined duration is constructed; the Bayesian information criterion value and the target number of clusters are determined according to the number of room temperature observation vectors; cluster analysis is performed on the room temperature observation vector of the predetermined duration according to the target number of clusters, and the results of the cluster analysis are displayed. Since a room temperature observation vector of a predetermined duration, such as a 24-hour room temperature observation vector, is generated based on the room temperature data during the analysis, it can be analyzed from the perspective of daily room temperature variation patterns. Furthermore, cluster analysis is performed using the target number of clusters determined according to the Bayesian information criterion value to deeply analyze the daily variation patterns of room temperature, thereby providing a basis for the heating regulation of the heating system. This at least partially overcomes the problem of low heating regulation accuracy in related technologies, thus achieving the technical effects of improving heating regulation accuracy, realizing effective heating, and improving user experience.

[0153] According to embodiments of this disclosure, the building module may include a processing submodule, a filling submodule, and a first generation submodule.

[0154] The processing submodule is used to process the target room temperature sequence data to obtain multiple sub-time-length room temperature data in units of sub-time.

[0155] The fill submodule is used to fill in missing values ​​for missing sub-duration room temperature data in response to the presence of missing sub-duration room temperature data in multiple sub-duration room temperature data.

[0156] The first generation submodule is used to generate a room temperature observation vector of a predetermined duration using multiple sub-duration room temperature data after padding.

[0157] According to embodiments of this disclosure, the processing submodule may include a first building unit and a processing unit.

[0158] The first building unit is used to construct a sub-time room temperature function based on the number of sub-times and the number of room temperature data in each sub-time.

[0159] The processing unit is used to process the target room temperature sequence data using the sub-time room temperature function to obtain sub-time room temperature data.

[0160] According to embodiments of this disclosure, the filling submodule may include a sorting unit, a first determining unit, and a filling unit.

[0161] The sorting unit is used to sort multiple sub-duration room temperature data according to time order to obtain the sorting result.

[0162] The first determining unit is used to determine the target location of the missing sub-duration room temperature data in the sorting results.

[0163] The filling unit is used to fill the missing values ​​with room temperature data for a preset number of sub-durations prior to the target location.

[0164] According to embodiments of this disclosure, the first generation submodule may include a calling unit, a first generation unit, and a second generation unit.

[0165] The calling unit is used to call standard room temperature data.

[0166] The first generation unit is used to generate the degree of deviation of each sub-time period room temperature data according to the standard room temperature data and each sub-time period room temperature data in the multiple sub-time period room temperature data after filling.

[0167] The second generation unit is used to generate a room temperature observation vector of a predetermined duration based on the degree of deviation of the room temperature data corresponding to each sub-duration.

[0168] According to embodiments of this disclosure, the clustering module may include a first construction submodule and an iterative clustering submodule.

[0169] The first construction submodule is used to construct a Gaussian mixture model function based on the target number of clusters, the weight coefficients of the Gaussian distribution, and the feature data of the Gaussian distribution.

[0170] The iterative clustering submodule is used to input the room temperature observation vectors for a predetermined duration into the Gaussian mixture model function for iterative cluster analysis until the cluster analysis converges.

[0171] According to embodiments of this disclosure, the iterative clustering submodule may include an assignment unit, a second construction unit, a third generation unit, and a second determination unit.

[0172] The assignment unit is used to assign initial values ​​to the weight coefficients, mean, and covariance of the Gaussian distribution.

[0173] The second building unit is used to construct the posterior probability of the room temperature vector for a predetermined duration based on the results obtained by inputting the weight coefficients of the Gaussian distribution, the mean of the Gaussian distribution, and the covariance of the Gaussian distribution into the Gaussian mixture model function.

[0174] The third generation unit is used to generate the maximum likelihood function values ​​of the mean of the Gaussian distribution, the maximum likelihood function values ​​of the covariance of the Gaussian distribution, and the maximum likelihood function values ​​of the weight coefficients of the Gaussian distribution based on the posterior probability.

[0175] The second determining unit is used to obtain the convergence result of cluster analysis when the convergence judgment functions constructed based on the maximum likelihood function of the mean of the Gaussian distribution, the maximum likelihood function of the covariance of the Gaussian distribution, and the maximum likelihood function of the weight coefficients of the Gaussian distribution satisfy a predetermined convergence threshold.

[0176] According to embodiments of this disclosure, the determining module may include a second building submodule and a second generating submodule.

[0177] The second construction submodule is used to construct the Bayesian information criterion function based on the number of room temperature observation vectors for a predetermined duration, the maximum value of the model's likelihood function, and the cluster number parameter.

[0178] The second generation submodule is used to iterate over different cluster number parameters within a predetermined numerical range and generate Bayesian information criterion values ​​corresponding to different cluster number parameters according to the Bayesian information criterion function.

[0179] According to embodiments of this disclosure, the indoor temperature clustering device applied to a centralized heating system may further include a first deletion module, a second deletion module, a third deletion module, and a filling module.

[0180] The first deletion module is used to delete the data collected by the room temperature acquisition sensor corresponding to the constant value in the initial room temperature sequence data.

[0181] The second deletion module is used to delete the room temperature data collected by the room temperature acquisition sensor corresponding to the lost room temperature data in response to the ratio between the number of lost room temperature data in the initial room temperature sequence data and the total number of room temperature data exceeding a predetermined ratio.

[0182] The third deletion module is used to delete abnormal data in response to the presence of abnormal data outside the predetermined normal distribution range in the initial room temperature sequence data.

[0183] The filling module is used to fill in the initial room temperature sequence data with the average room temperature data over a unit of time in response to the presence of deleted data.

[0184] Any one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure, or at least part of the functions of any one or more of them, can be implemented in one module. Any one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure can be implemented by dividing them into multiple modules. Any one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure can be at least partially implemented as hardware circuitry, such as a Field-Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a System-on-Chip, a System-on-a-Substrate, a System-on-Package, an Application-Specific Integrated Circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure can be at least partially implemented as computer program modules, which, when run, can perform corresponding functions.

[0185] For example, any plurality of the temperature acquisition sensor 710, generation module 720, construction module 730, determination module 740, and clustering module 750 can be combined into one module / unit / subunit, or any one of these modules / units / subunits can be split into multiple modules / units / subunits. Alternatively, at least part of the functionality of one or more of these modules / units / subunits can be combined with at least part of the functionality of other modules / units / subunits and implemented in one module / unit / subunit. According to embodiments of the present disclosure, at least one of the temperature acquisition sensor 710, generation module 720, construction module 730, determination module 740, and clustering module 750 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the temperature acquisition sensor 710, generation module 720, construction module 730, determination module 740 and clustering module 750 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0186] It should be noted that the indoor temperature clustering device part applied to the centralized heating system in the embodiments of this disclosure corresponds to the indoor temperature clustering method part applied to the centralized heating system in the embodiments of this disclosure. For a detailed description of the indoor temperature clustering device part applied to the centralized heating system, please refer to the indoor temperature clustering method part applied to the centralized heating system, which will not be repeated here.

[0187] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways, even if such combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments of this disclosure may be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0188] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for clustering indoor temperature applied to a central heating system, characterized in that, The method includes: Initial room temperature sequence data is generated based on room temperature data obtained from multiple room temperature acquisition sensors; Based on the target room temperature sequence data, a room temperature observation vector of a predetermined duration is constructed, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; Based on the number of room temperature observation vectors for the predetermined duration, the Bayesian information criterion value is determined, and the number of model parameters corresponding to the minimum value of the Bayesian information criterion value is determined as the target cluster number. Based on the target number of clusters, cluster analysis is performed on the room temperature observation vectors for the predetermined duration, and the clustering results obtained when the cluster analysis converges are displayed. The step of performing cluster analysis on the room temperature observation vector of the predetermined duration based on the target number of clusters includes: constructing a Gaussian mixture model function based on the target number of clusters, the weight coefficients of the Gaussian distribution, and the characteristic data of the Gaussian distribution; inputting the room temperature observation vector of the predetermined duration into the Gaussian mixture model function, and performing iterative cluster analysis until the cluster analysis converges; The characteristic data of the Gaussian distribution includes the mean and covariance of the Gaussian distribution. The convergence of the clustering analysis is determined as follows: initial values ​​are assigned to the weight coefficients, mean, and covariance of the Gaussian distribution; based on the results obtained by inputting the weight coefficients, mean, and covariance of the Gaussian distribution into the Gaussian mixture model function, the posterior probability of the room temperature vector for the predetermined duration is constructed; according to the posterior probability, the maximum likelihood function values ​​of the mean, covariance, and weight coefficients of the Gaussian distribution are generated; if the convergence judgment functions constructed based on the maximum likelihood functions of the mean, covariance, and weight coefficients of the Gaussian distribution satisfy a predetermined convergence threshold, the convergence result of the clustering analysis is obtained. The Gaussian mixture model includes the maximum likelihood function of the model; determining the Bayesian information criterion value based on the number of room temperature observation vectors over the predetermined duration includes: constructing a Bayesian information criterion function based on the number of room temperature observation vectors over the predetermined duration, the maximum likelihood function of the model, and the number of clusters; iterating over different number of clusters parameters within a predetermined numerical range, and generating the Bayesian information criterion value corresponding to the different number of clusters parameters based on the Bayesian information criterion function; The preprocessing includes at least one of the following: in response to the presence of a constant value in the initial room temperature sequence data, deleting the data collected by the room temperature acquisition sensor corresponding to the constant value; in response to the ratio between the number of lost room temperature data and the total number of room temperature data in the initial room temperature sequence data exceeding a predetermined ratio, deleting the room temperature data collected by the room temperature acquisition sensor corresponding to the lost room temperature data; in response to the presence of abnormal data outside a predetermined normal distribution interval in the initial room temperature sequence data, deleting the abnormal data; and in response to the presence of deleted data in the initial room temperature sequence data, filling it with the average room temperature data over a unit time period.

2. The method of claim 1, wherein, The predetermined duration includes multiple sub-durations; The construction of a room temperature observation vector of a predetermined duration based on the target room temperature sequence data includes: The target room temperature sequence data is processed to obtain multiple sub-time period room temperature data in units of the sub-time period; In response to the presence of missing sub-time room temperature data in the plurality of sub-time room temperature data, the missing sub-time room temperature data is filled with missing values; Using the filled room temperature data from multiple sub-durations, a room temperature observation vector for the predetermined duration is generated.

3. The method of claim 2, wherein, The process of processing the target room temperature sequence data to obtain multiple sub-time period room temperature data in units of the sub-time period includes: Based on the number of sub-durations and the number of room temperature data in each sub-duration, a sub-duration room temperature function is constructed; The target room temperature sequence data is processed using the sub-duration room temperature function to obtain the sub-duration room temperature data.

4. The method of claim 3, wherein, The sub-durations have a temporal order; The missing value imputation for the missing sub-duration room temperature data includes: Based on the time sequence, the multiple sub-duration room temperature data are sorted to obtain the sorting result; In the sorting results, the target location of the missing sub-duration room temperature data is determined; The room temperature data for a predetermined number of sub-durations prior to the target location is used to fill the missing value.

5. The method of claim 2, wherein, The step of generating the room temperature observation vector of the predetermined duration using multiple sub-duration room temperature data after padding includes: Call standard room temperature data; Based on the temperature data of each of the multiple sub-time periods in the filled temperature data and the standard temperature data, a deviation degree corresponding to each of the sub-time periods in the temperature data is generated; Based on the degree of deviation of the room temperature data corresponding to each of the sub-durations, a room temperature observation vector for the predetermined duration is generated.

6. A room temperature clustering device for a district heating system, applied to the room temperature clustering method for a district heating system according to any one of claims 1 to 5, characterized by The device includes: Temperature acquisition sensor, used to collect room temperature data; The generation module is used to generate initial room temperature sequence data based on room temperature data obtained from multiple room temperature acquisition sensors; A construction module is used to construct a room temperature observation vector of a predetermined duration based on target room temperature sequence data, wherein the target room temperature sequence data is obtained by preprocessing the initial room temperature sequence data; A determining module is configured to determine a Bayesian information criterion value based on the number of the room temperature observation vectors of the predetermined time length, and determine the number of model parameters corresponding to the minimum value of the Bayesian information criterion value as a target cluster number; A clustering module is configured to perform cluster analysis on the room temperature observation vectors of the predetermined time length according to the target cluster number, and display the cluster result obtained under the cluster analysis convergence condition.

Images