A Multi-Source Heterogeneous Data Management Method and System Based on Digital Twin

Abstract

The present invention provides a method and system for managing multi-source heterogeneous data based on digital twin, which relates to the field of multi-source heterogeneous data management. The method includes: clustering building units with the same energy consumption characteristics into a cluster according to the basic attributes and historical energy consumption of each building unit, and each cluster corresponds to an initial historical energy consumption label; using the energy consumption prediction model of the digital twin platform to predict the target predicted energy consumption time curve S of the target building unit within T, and then obtaining the target predicted energy consumption label BS; at the same time, obtaining the target measured energy consumption time curve Y of the target building unit within T to obtain the target measured energy consumption label BY, and then determining whether the measured energy consumption of the target building unit within T is normal according to BS and BY; the present invention can determine the abnormality through the prediction method of the digital twin platform when the measured energy consumption is abnormal, avoid the digital twin platform using inaccurate energy consumption data, and ensure the normal analysis and processing results.

Description

Technical Field

[0001] The present invention relates to the field of multi-source heterogeneous data management, and in particular to a multi-source heterogeneous data management method and system based on digital twin. Background Art

[0002] With the rapid development of building intelligence, more and more large-scale parks begin to use digital twin technology to monitor and display the energy consumption of buildings in the park in real time. In the prior art, energy consumption monitoring data of buildings is usually obtained in real time, and then analyzed and processed by a digital twin platform. However, the energy consumption monitoring of buildings is obtained by metering devices, and the metering devices themselves are affected by environmental factors, such as electromagnetic interference, etc. At the same time, as the service life of the metering devices increases, the probability of failure also increases. The above situations will seriously affect the metering accuracy of the energy consumption metering devices. If the digital twin platform uses inaccurate energy consumption data, it will lead to abnormal analysis and processing results. Therefore, how to judge the abnormality of the energy consumption data of buildings through the digital twin platform has become an urgent technical problem to be solved. Summary of the Invention

[0003] For the above technical problems, the technical solution adopted by the present invention is as follows:

[0004] According to the first aspect of the present application, a multi-source heterogeneous data management method based on digital twin is provided, and the method includes the following steps:

[0005] S100, clustering all building units of the target building according to the basic attributes and historical energy consumption corresponding to each building unit in the target area to obtain a cluster list A = (A1, A2,..., A i ,..., A n ), i = 1, 2,..., n; where A i is the i-th cluster obtained by clustering, and n is the number of clusters obtained by clustering; the energy consumption characteristics of the building units within each cluster are the same;

[0006] S200, obtaining the initial historical energy consumption labels corresponding to each cluster in A to obtain an initial historical energy consumption label list B = (B1, B2,..., B i ,..., B n ); where B i is the initial historical energy consumption label corresponding to A i ;

[0007] S300, inputting the basic attributes and historical energy consumption corresponding to the target building unit into the energy consumption prediction model of the preset digital twin platform to obtain the target predicted energy consumption time curve S of the target building unit within the preset time period T; where the start time of T is the current time Tnow ;

[0008] S400. Determine the target predicted energy consumption label BS corresponding to the target building unit within T according to S.

[0009] S500. If T now = T end , then obtain the target measured energy consumption time curve Y of the target building unit within T; where T end is the end time of T.

[0010] S600. Determine the target measured energy consumption label BY corresponding to the target building unit within T according to Y.

[0011] S700. If BS is the same as B' and BY is the same as B', then determine that the measured energy consumption of the target building unit within T is normal; otherwise, proceed to S800.

[0012] S800. Obtain the first measured energy consumption label of each building unit within the cluster corresponding to B' to determine whether the measured energy consumption of the target building unit within T is abnormal; where B' is the initial historical energy consumption label corresponding to the target building unit in B.

[0013] According to another aspect of the present application, there is also provided a digital twin-based multi-source heterogeneous data management system, the system includes: a processor and a storage medium; wherein, at least one instruction or at least one program segment is stored in the storage medium, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the digital twin-based multi-source heterogeneous data management method as described in the first aspect.

[0014] The present invention has at least the following beneficial effects:

[0015] The digital twin-based multi-source heterogeneous data management method of the present invention clusters building units with the same energy consumption characteristics into a cluster according to the basic attributes and historical energy consumption of each building unit, and each cluster corresponds to an initial historical energy consumption label; uses the energy consumption prediction model of the digital twin platform to predict the target predicted energy consumption time curve S of the target building unit within T, and then obtains the target predicted energy consumption label BS; at the same time, obtains the target measured energy consumption time curve Y of the target building unit within T to obtain the target measured energy consumption label BY, and then determines whether the measured energy consumption of the target building unit within T is normal according to BS and BY; thus, when the measured energy consumption is abnormal, the abnormality is determined through the prediction method of the digital twin platform, avoiding the use of inaccurate energy consumption data by the digital twin platform and ensuring the normalcy of the analysis and processing results. Description of the Drawings

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] Figure 1 It is a flowchart of a multi-source heterogeneous data management method based on digital twin provided by an embodiment of the present invention. Specific embodiments

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0019] It should be noted that based on this disclosure, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. In addition, this device and / or practice this method can be implemented using other structures and / or functions in addition to one or more of the aspects described herein.

[0020] The following will refer to Figure 1 the flowchart of the multi-source heterogeneous data management method based on digital twin shown in the figure to introduce a multi-source heterogeneous data management method based on digital twin.

[0021] The multi-source heterogeneous data management method based on digital twin may include the following steps:

[0022] S100. Cluster all building units of the target building according to the basic attributes and historical energy consumption corresponding to each building unit in the target area to obtain a cluster list A = (A1, A2,..., A i ,..., A n ), where i = 1, 2,..., n; among them, A i is the i-th cluster obtained by clustering, and n is the number of clusters obtained by clustering; the energy consumption characteristics of the building units within each cluster are the same.

[0023] In this embodiment, the target area can be a large office area or industrial area, and the historical energy consumption can be the energy consumption of each building unit every day or every hour within a past period of time.

[0024] Further, the basic attributes include the longitude and latitude coordinates, height, azimuth angle, structure type, and date information of the building unit.

[0025] In this embodiment, each building unit can be each floor. The longitude and latitude coordinates corresponding to each building are different, the height of each building unit is also different, the azimuth angle of each building unit can be understood as the offset angle from the due south, and the structure type is the building structure type of the building, such as: concrete structure, steel frame structure, etc.; the heat insulation conditions corresponding to different structure types are different, and the date information can clarify the specific season.

[0026] Further, step S100 may include the following steps:

[0027] S110, obtain each basic attribute corresponding to each building unit in the target area to obtain a set of basic attribute lists D = (D1, D2,..., D p ,..., D q ), p = 1, 2,..., q; where D p is the basic attribute list corresponding to the p-th building unit, and q is the number of building units in the target area; D p = (D p，1 , D p，2 ,..., D p，u ,..., D p，v ), u = 1, 2,..., v; D p，u is the u-th basic attribute corresponding to the p-th building unit, and v is the number of basic attributes corresponding to each building unit.

[0028] In this embodiment, the basic attributes of each building unit can be determined according to the building parameters of each building in the target area. It can be understood that the basic attributes corresponding to each building unit include static parameters and date data.

[0029] S120, obtain the historical energy consumption of each building unit in each preset historical time period in the target area to obtain a set of historical energy consumption lists E = (E1, E2,..., E p ,..., E q ); where E p is the historical energy consumption list corresponding to the p-th building unit; E p = (E p，1 , E p，2 ,..., E p，x ,..., E p，y ), x = 1, 2,..., y; E p，x is the historical energy consumption of the p-th building unit in the x-th preset historical time period.

[0030] In this embodiment, the preset historical time period can be set according to the actual situation. For example, the duration of the preset historical time period can be set to one day, 12 hours, 2 hours, etc.; the historical energy consumption of each building unit is recorded, and E can be obtained through the historical records; it can be understood that the historical energy consumption is the behavior data of the building unit.

[0031] S130. Concatenate D and E to obtain the heterogeneous data group corresponding to each building unit, and further obtain the heterogeneous data group list F = (F1, F2,..., F p ,..., F q ); where F p is the heterogeneous data group corresponding to the p-th building unit; F p = (D p , E p ).

[0032] In this embodiment, by concatenating the static data and behavior data of each building unit, multi-source heterogeneity of data is achieved; it should be noted that the parameters in the heterogeneous data group are not entirely in the form of numerical values, and non-numerical data needs to be converted into numerical form.

[0033] S140. Perform normalization processing on each heterogeneous data group in F to obtain the feature vector list XL = (XL1, XL2,..., XL p ,..., XL q ); where XL p is the feature vector corresponding to the p-th building unit.

[0034] In this embodiment, for non-numerical data, encoding can be used for numericalization. For example, for type data, type 1 can be encoded as 1, type 2 can be encoded as 2, etc.; it should be noted that after all the data in the heterogeneous data group are numericalized, those skilled in the art can use existing normalization methods to perform normalization processing on each heterogeneous data group in F according to actual needs, which will not be elaborated here.

[0035] S150. Cluster XL to obtain A.

[0036] In this embodiment, the K-means clustering algorithm can be used to cluster XL; it should be noted that those skilled in the art can use the existing K-means clustering algorithm to cluster XL according to actual needs, which will not be elaborated here; it can be understood that the energy consumption characteristics of the building units within each cluster are the same. For example, the energy consumption characteristics of all building units within a certain cluster are stable energy consumption and high energy consumption, and the energy consumption characteristics of all building units within a certain cluster are large energy consumption fluctuations and low energy consumption.

[0037] S200, Obtain the initial historical energy consumption labels corresponding to each cluster in A to obtain the initial historical energy consumption label list B=(B1, B2, …, B i , …, B n ); where B i is the initial historical energy consumption label corresponding to A i .

[0038] In this embodiment, since the energy consumption characteristics of all building units within each cluster are the same, therefore, the historical energy consumption of a building unit can be randomly selected from B i , and then the corresponding historical energy consumption time curve can be obtained. By the method in steps S410 - S440, determine the initial historical energy consumption labels corresponding to each cluster in A to obtain B

[0039] S300, Input the basic attributes and historical energy consumption of the target building unit into the energy consumption prediction model of the preset digital twin platform to obtain the target predicted energy consumption time curve S of the target building unit within the preset time period T; where the start time of T is the current time T now .

[0040] In this embodiment, the target building unit can be any building unit within the target area; the preset digital twin platform can display the energy consumption data of each building unit. At the same time, the digital twin platform is also preset with an energy consumption prediction model, which can predict the energy consumption within a future period according to the basic attributes and historical energy consumption of the building unit; the energy consumption prediction model can be a trained Long Short-Term Memory (LSTM) network. LSTM is a special Recurrent Neural Network (RNN) structure that can learn long-term dependence information. It performs particularly well in processing and predicting important events with long intervals and delays in time series, which makes LSTM very suitable for processing and predicting important events with long intervals and delays in time series data

[0041] S400, According to S, determine the target predicted energy consumption label BS corresponding to the target building unit within T

[0042] Further, step S400 may include the following steps:

[0043] S410, Divide T into several consecutive and equal-duration sub-time periods to obtain the sub-time period list ZT=(ZT1, ZT2, …, ZT a , …, ZT b ), a = 1, 2, …, b; where ZT a is the a-th sub-time period obtained by dividing T, and b is the number of sub-time periods obtained by dividing T

[0044] In this embodiment, the duration of the sub-time period can be set according to actual needs. For example, the duration of the sub-time period is 10 minutes or 5 minutes, etc.

[0045] S420. According to S, obtain the sub-predicted energy consumption corresponding to each sub-time period of the target building unit in T, so as to obtain a sub-predicted energy consumption list ZY = (ZY1, ZY2,..., ZY a , …, ZY b ); where ZY a is the sub-predicted energy consumption corresponding to the target building unit within ZT a .

[0046] In this embodiment, the X-axis of S is the time point and the Y-axis is the corresponding predicted energy consumption. Therefore, according to S, the sub-predicted energy consumption corresponding to each sub-time period of the target building unit in T can be obtained.

[0047] S430. According to ZY, determine the predicted average energy consumption QA = (1 / b) × ∑ b a=1 ZY a and the predicted energy consumption volatility QB = (1 / b) × ∑ b a=1 (ZY a - ((1 / b) × ∑ b a=1 ZY a )) 2 .

[0048] In this embodiment, the predicted average energy consumption QA can represent the predicted average energy consumption level of the target building unit within Y, and the predicted energy consumption volatility QB can represent the fluctuation of the predicted energy consumption of the target building unit within T.

[0049] S440. According to QA, QB and the preset energy consumption label mapping table, determine BS; where the preset energy consumption label mapping table includes several rows, and each row corresponds to a combination of an average energy consumption range and an energy consumption volatility range and an energy consumption label.

[0050] In this embodiment, the preset energy consumption label mapping table stores several combinations of average energy consumption ranges and energy consumption volatility ranges and the corresponding energy consumption labels; for example: the energy consumption label corresponding to the first combination is high energy consumption with stable energy consumption, and the energy consumption label corresponding to the second combination is high energy consumption with low energy consumption fluctuation, etc.; all combinations can be traversed. If both QA and QB fall within the average energy consumption range and the energy consumption volatility range in a combination, then the energy consumption label corresponding to this combination is determined as BS.

[0051] S500. If T now = T end, the target metered energy consumption time curve Y of the target building unit within T is obtained; where T end is the end time of T.

[0052] In this embodiment, when the current time is T end , the target metered energy consumption time curve Y of the target building unit within T can be obtained; Y is obtained by an energy consumption metering device, such as an electric energy meter, through metering the energy consumption of the target building unit.

[0053] S600. According to Y, the target metered energy consumption label BY corresponding to the target building unit within T is determined.

[0054] In this embodiment, after obtaining Y, BY can also be determined by the method in steps S410 - S440, which will not be elaborated here.

[0055] S700. If BS is the same as B' and BY is the same as B', it is determined that the metered energy consumption of the target building unit within T is normal; otherwise, go to S800.

[0056] In this embodiment, under normal circumstances, the energy consumption characteristics of each building unit have a relatively stable energy consumption behavior pattern. Therefore, if BS is the same as B' and BY is the same as B', it indicates that the energy consumption behavior of the target building unit has not changed significantly compared with the historical energy consumption behavior, and it can be determined that the metered energy consumption of the target building unit within T is normal.

[0057] S800. The first metered energy consumption label of each building unit within the cluster corresponding to B' within T is obtained to determine whether the metered energy consumption of the target building unit within T is abnormal; where B' is the initial historical energy consumption label corresponding to the target building unit in B.

[0058] Further, step S800 may include the following steps:

[0059] S810. The first metered energy consumption label of each building unit within the cluster corresponding to B' within T is obtained to obtain the first metered energy consumption label list C = (C1, C2,..., C j ,..., C m ), j = 1, 2,..., m; where C j is the first metered energy consumption label of the jth building unit within the cluster corresponding to B' within T, and m is the number of building units within the cluster corresponding to B'.

[0060] In this embodiment, the first metered energy consumption label of each building unit within the cluster corresponding to B' within T can be obtained by the method in steps S410 - S440 above.

[0061] S811. Obtain the quantity NUM1 of the first metering energy consumption labels in C that are different from B'.

[0062] S812. If BS is the same as B', BY is different from B' and NUM1 / m < α1, determine that the metering energy consumption of the target building unit within T is abnormal; where α1 is the first preset weight.

[0063] In this embodiment, α1 is a relatively small value, which can be set to 0.01 to 0.03; if BS is the same as B', BY is different from B' and NUM1 / m < α1, it means that the predicted energy consumption behavior of the target building unit by the digital twin platform is different from the measured energy consumption behavior of the target building unit. Moreover, among the building units in the same cluster as the target building unit, the proportion of the quantity of the first metering energy consumption labels that are different from B' in the total quantity is also very small; this situation can be considered as an abnormality of the energy consumption metering device of the target building unit, which is a sudden small-probability event.

[0064] S813. If BS is the same as B', BY is different from B' and NUM1 / m > α2, determine that the metering energy consumption of the target building unit within T is normal; where α2 is the second preset weight; α1 < α2.

[0065] In this embodiment, α2 is a value larger than α1, which can be set to 0.4 to 0.6; if BS is the same as B', BY is different from B' and NUM1 / m > α2, it means that the predicted energy consumption behavior of the target building unit by the digital twin platform is different from the measured energy consumption behavior of the target building unit. Moreover, among the building units in the same cluster as the target building unit, the proportion of the quantity of the first metering energy consumption labels that are different from B' in the total quantity is relatively large; this situation can be considered as the energy consumption metering device of the target building unit being interfered by environmental factors, for example: there is electromagnetic interference in the target area, etc.

[0066] S814. If BS is the same as B', BY is different from B' and α1 ≤ NUM1 / m ≤ α2, determine whether the metering energy consumption of the target building unit within T is abnormal in a manual way.

[0067] In this embodiment, if BS is the same as B', BY is different from B' and α1 ≤ NUM1 / m ≤ α2, there are various factors causing this situation, and it is necessary to determine whether the metering energy consumption of the target building unit within T is abnormal in a manual way.

[0068] Further, after step S810, the method further includes the following steps

[0069] S820. If BS is different from B', BY is the same as B' and (m - NUM1) / m > α3, it is determined that the measured energy consumption of the target building unit in T is normal, and the predicted energy consumption of the target building unit in T is abnormal; where α3 is the third preset weight; α3 > α2.

[0070] S821. If BS is different from B', BY is the same as B' and (m - NUM1) / m ≤ α3, determine whether the measured energy consumption of the target building unit in T is abnormal manually.

[0071] In this embodiment, if BS is different from B' and BY is the same as B', it may be that the prediction model makes a wrong prediction; however, it may also be that when the external environmental factors change greatly, the metering device also malfunctions; α3 can be set to a relatively large value, for example: α3 can be set to 0.98 to 0.99; in the above case, if (m - NUM1) / m > α3, it means that the energy consumption behaviors of the building units in the same cluster of the target building unit have not changed greatly, so it can be determined that the predicted energy consumption of the target building unit in T is abnormal; if (m - NUM1) / m ≤ α3, it means that the energy consumption behaviors of a certain number of building units in the same cluster of the target building unit have not changed greatly, and there are various factors causing this situation, and it is necessary to determine whether the measured energy consumption of the target building unit in T is abnormal manually.

[0072] Further, after step S700 and before step S800, the method further includes the following steps:

[0073] S710. If BS is different from B' and BY is different from B', determine whether the measured energy consumption of the target building unit in T is abnormal manually.

[0074] In this embodiment, this situation may be caused by abnormalities in both the metering device and the prediction model, and it is necessary to determine whether the measured energy consumption of the target building unit in T is abnormal manually.

[0075] In this embodiment, building units with the same energy consumption characteristics are clustered into a cluster according to the basic attributes and historical energy consumption of each building unit, and each cluster corresponds to an initial historical energy consumption label; the energy consumption prediction model of the digital twin platform is used to predict the target predicted energy consumption time curve S of the target building unit in T, and then the target predicted energy consumption label BS is obtained; at the same time, the target measured energy consumption time curve Y of the target building unit in T is obtained to obtain the target measured energy consumption label BY, and then it is determined whether the measured energy consumption of the target building unit in T is normal according to BS and BY; thus, when the measured energy consumption is abnormal, this abnormality is determined through the prediction of the digital twin platform, avoiding the use of inaccurate energy consumption data by the digital twin platform and ensuring the normalcy of the analysis and processing results.

[0076] In addition, although the steps of the methods in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution, etc.

[0077] An embodiment of the present invention further provides a multi-source heterogeneous data management system based on digital twin. The system includes: a processor and a storage medium; wherein, at least one instruction or at least one program segment is stored in the storage medium, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the multi-source heterogeneous data management method based on digital twin as described in the above embodiment.

[0078] Although some specific embodiments of the present invention have been described in detail by way of examples, those skilled in the art should understand that the above examples are for illustrative purposes only and not for limiting the scope of the present invention. Those skilled in the art should also understand that various modifications can be made to the embodiments without departing from the scope and spirit of the present invention.

Claims

1. A multi-source heterogeneous data management method based on digital twins, characterized in that: The method comprises the following steps: S100, clustering all building units of the target building according to the basic attributes and historical energy consumption of each building unit in the target area to obtain a cluster list A = (A1, A2, ..., A i , …, A n ), i=1, 2,...,n; among them, A i is the i-th cluster obtained by clustering, n is the number of clusters obtained by clustering; the energy consumption characteristics of the building units in each cluster are the same; S200, obtaining the initial historical energy consumption label corresponding to each cluster in A, so as to obtain the initial historical energy consumption label list B corresponding to A = (B1, B2, ..., B i , …, B n ); where B i A i The corresponding initial historical energy consumption label; S300, input the basic attributes and historical energy consumption corresponding to the target building unit into the preset energy consumption prediction model of the digital twin platform to obtain the target predicted energy consumption time curve S of the target building unit within the preset time period T; wherein the start time of T is the current time T now ; S400, determining the target predicted energy consumption label BS corresponding to the target building unit in T according to S; S500, if the current time reaches T end , then the target metering energy consumption time curve Y of the target building unit within T is obtained; where T end is the end time of T; S600, determining the target metering energy consumption label BY corresponding to the target building unit in T according to Y; S700, if BS is the same as B' and BY is the same as B', it is determined that the energy consumption of the target building unit is measured normally within T; otherwise, enter S800; S800, obtaining the first metering energy consumption tag of each building unit in the cluster corresponding to B' in T, so as to determine whether the metering energy consumption of the target building unit in T is abnormal; wherein B' is the initial historical energy consumption tag corresponding to the target building unit in B; Step S800 includes the following steps: S810, obtaining the first metering energy consumption label of each building unit in the cluster corresponding to B' in T, so as to obtain a first metering energy consumption label list C corresponding to B' = (C1, C2, ..., C j , …, C m ), j = 1, 2, …, m; where C j is the first metering energy consumption label of the jth building unit in the cluster corresponding to B' in T, and m is the number of building units in the cluster corresponding to B'; S811, obtaining the number NUM1 of first energy consumption metering tags in C that are different from B'; S812, if BS is the same as B', BY is different from B' and NUM1 / m<α1, it is determined that the metering energy consumption of the target building unit in T is abnormal, and the abnormal reason is that the energy consumption metering device of the target building unit itself is abnormal; wherein α1 is the first preset weight; S813, if BS is the same as B', BY is different from B' and NUM1 / m>α2, it is determined that the metered energy consumption of the target building unit in T is normal and the energy consumption metering device of the target building unit is interfered by environmental factors; wherein α2 is the second preset weight; α1<α2; S814, if BS is the same as B', BY is different from B' and α1≤NUM1 / m≤α2, then determine whether the metered energy consumption of the target building unit in T is abnormal by manual means; After step S810, the method further includes the following steps: S820, if BS is different from B', BY is the same as B' and (m-NUM1) / m>α3, it is determined that the metered energy consumption of the target building unit in T is normal and the predicted energy consumption of the target building unit in T is abnormal; wherein α3 is a third preset weight; α3>α2; the value range of α3 is 0.98 to 0.99; S821, if BS is different from B', BY is the same as B' and (m-NUM1) / m≤α3, determine whether the metered energy consumption of the target building unit in T is abnormal by manual means.

2. The multi-source heterogeneous data management method based on digital twins according to claim 1 is characterized in that: After step S700 and before step S800, the method further includes the following steps: S710: If BS is different from B' and BY is different from B', determine whether the metered energy consumption of the target building unit in T is abnormal by manual means.

3. The multi-source heterogeneous data management method based on digital twins according to claim 1 is characterized in that: Step S100 includes the following steps: S110, obtaining each basic attribute corresponding to each building unit in the target area to obtain a basic attribute list set D = (D1, D2, ..., D p , …, D q ), p = 1, 2, …, q; where D p is the basic attribute list corresponding to the pth building unit, q is the number of building units in the target area; D p =(D p，1 , D p，2 , …, D p，u , …, D p，v ), u=1, 2, …, v; D p，u is the uth basic attribute corresponding to the pth building unit, and v is the number of basic attributes corresponding to each building unit; S120, obtaining the historical energy consumption of each building unit in the target area in each preset historical time period to obtain a historical energy consumption list set E = (E1, E2, ..., E p ,…,E q ); where E p is the historical energy consumption list corresponding to the p-th building unit; E p =(E p，1 , E p，2 ,…,E p，x ,…,E p，y ), x=1, 2, …, y; E p，x is the historical energy consumption of the p-th building unit in the x-th preset historical time period; S130, concatenate D and E to obtain a heterogeneous data set corresponding to each building unit, and then obtain a heterogeneous data set list F = (F1, F2, ..., F p , …, F q ); where F p is the heterogeneous data set corresponding to the p-th building unit; F p =(D p , E p ); S140, normalize each heterogeneous data set in F to obtain a feature vector list XL corresponding to F = (XL1, XL2, ..., XL p ,…,XL q ); Among them, XL p is the eigenvector corresponding to the pth building unit; S150, clustering XL to obtain A.

4. The multi-source heterogeneous data management method based on digital twins according to claim 1 is characterized in that: BS is obtained through the following steps: S410, dividing T into a number of continuous sub-time periods of equal duration, so as to obtain a sub-time period list ZT corresponding to T = (ZT1, ZT2, ..., ZT a , …, ZT b ), a=1, 2, …, b; where ZT a is the ath sub-time period obtained by dividing T, and b is the number of sub-time periods obtained by dividing T; S420, according to S, obtain the corresponding sub-predicted energy consumption of the target building unit in each sub-time period in T, so as to obtain a sub-predicted energy consumption list ZY=(ZY1, ZY2, ..., ZY a ,…,ZY b ); Among them, ZY a For the target building unit in ZT a The corresponding sub-prediction energy consumption; S430, based on ZY, determine the predicted average energy consumption QA= (1 / b) × ∑ b a=1 ZY a And the predicted energy consumption fluctuation rate QB = (1 / b) × ∑ b a=1 (ZY a -((1 / b)×∑ b a=1 ZY a )) 2 ; S440, determining BS according to QA, QB and a preset energy consumption label mapping table; wherein the preset energy consumption label mapping table includes a plurality of rows, each row corresponding to a combination of an average energy consumption range and an energy consumption fluctuation rate range and an energy consumption label.

5. The multi-source heterogeneous data management method based on digital twins according to claim 1 is characterized in that: Basic attributes include the building unit's latitude and longitude coordinates, height, azimuth angle, structure type, and date information.

6. The multi-source heterogeneous data management method based on digital twins according to claim 1 is characterized in that: The building unit is a floor of a corresponding building.

7. A multi-source heterogeneous data management system based on digital twins, characterized in that: The system includes: a processor and a storage medium; wherein, the storage medium stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the multi-source heterogeneous data management method based on digital twins as described in any one of claims 1-6.

Images