Building data management system and method based on multi-modal data fusion

Abstract

The invention belongs to the technical field of building data management, and relates to a building data management system and method based on multi-modal data fusion, and the method comprises the steps: obtaining the real-time operation data of a target building multi-source sensing network, and carrying out the cross-domain data association analysis of the real-time operation data, so as to generate an energy consumption distribution state diagram; retrieving a collaborative abnormal region of light intensity and a medium through space coordinate mapping, generating an energy-saving early warning signal based on a topological position of the collaborative abnormal region, and dynamically adjusting energy regulation and control strategy parameters distributed to different functional regions by integrating the energy consumption distribution state diagram and the energy-saving early warning signal. And the comparison feedback data before and after the energy regulation and control operation is collected, the energy-saving effect performance corresponding to the energy regulation and control is analyzed, the energy regulation and control strategy parameters are corrected according to the energy-saving effect performance, a new round of cross-domain data association analysis is triggered until the preset energy-saving target is reached, and the building data management level and the energy-saving and consumption-reducing capacity are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of building data management, and specifically relates to a building data management system and method based on multimodal data fusion. Background Art

[0002] Building energy management is key to achieving energy conservation and sustainable development. Optimizing and fine-tuning energy consumption is particularly crucial in large public buildings and commercial complexes. Its core objective is to collect, analyze, and regulate diverse energy data within a building to improve utilization efficiency, reduce energy consumption, and achieve energy conservation and emission reduction goals. However, building energy systems involve complex interactions among multiple subsystems, including power, HVAC, lighting, and water supply and drainage. These subsystems have diverse data types and dispersed sources, posing significant challenges to data integration and governance.

[0003] Existing technologies for building energy management often use single data source monitoring and analysis as the core model: independent monitoring of various energy usage through equipment such as electricity meters, water meters, and gas meters, trend forecasting and energy-saving optimization based on historical data, and the application of discrete algorithm models in the data processing process, analyzing different data sources separately and then summarizing them for processing.

[0004] However, the above technical means have certain limitations when addressing complex building energy management needs, as shown in the following: Building energy management involves a wide variety of data that changes dynamically. A single data source monitoring and analysis method is difficult to fully reflect the overall operating status of the building. In addition, the existing data processing method is not adaptable enough when facing multimodal data, which makes it easy for data silos to appear and makes it impossible to effectively explore the potential correlations between data.

[0005] Existing building energy-saving optimization strategies are mostly based on static rules or empirical models, lacking the ability to quickly respond to real-time dynamic changes. This may lead to inaccurate energy regulation and even unnecessary energy waste, restricting the overall effectiveness of building energy management and making it difficult to fully realize the potential of multimodal data in energy-saving optimization. Summary of the Invention

[0006] In view of this, in order to solve the problems raised in the above background technology, a building data governance system and method based on multimodal data fusion is proposed.

[0007] The technical solution adopted by the present invention to solve its technical problems is: First, the present invention provides a building data governance system based on multimodal data fusion, including: a data acquisition module, a data analysis module, an energy control module and an energy-saving optimization module.

[0008] The data acquisition module is connected to the data analysis module, the data analysis module is connected to the energy regulation module, the energy regulation module is connected to the energy-saving optimization module, and the energy-saving optimization module is connected to the data analysis module.

[0009] The data acquisition module obtains the real-time operation data of the target building's multi-source perception network.

[0010] The data analysis module performs cross-domain data correlation analysis on the real-time operation data to generate an energy consumption distribution status diagram, retrieves the collaborative abnormal area of ​​light intensity and medium through spatial coordinate mapping, and generates an energy-saving warning signal based on the topological position of the collaborative abnormal area.

[0011] The energy control module dynamically adjusts the energy control strategy parameters distributed to different functional areas based on the energy consumption distribution state diagram and the energy-saving early warning signal.

[0012] The energy-saving optimization module collects comparative feedback data before and after energy control operations, analyzes the energy-saving effect corresponding to energy control, and accordingly modifies the energy control strategy parameters and triggers a new round of cross-domain data correlation analysis until the preset energy-saving target is achieved.

[0013] In a second aspect, the present invention provides a building data governance method based on multimodal data fusion, including: obtaining real-time operation data of the multi-source perception network of the target building.

[0014] A cross-domain data association analysis is performed on the real-time operation data to generate an energy consumption distribution state diagram, and the collaborative abnormal area of ​​light intensity and medium is retrieved through spatial coordinate mapping, and an energy-saving early warning signal is generated based on the topological position of the collaborative abnormal area.

[0015] The energy consumption distribution state diagram and the energy-saving early warning signal are comprehensively used to dynamically adjust the energy control strategy parameters distributed to different functional areas.

[0016] Collect comparative feedback data before and after energy regulation operations, analyze the energy-saving effect corresponding to energy regulation, and accordingly modify the energy regulation strategy parameters and trigger a new round of cross-domain data correlation analysis until the preset energy-saving target is achieved.

[0017] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: The present invention divides the target area into energy consumption intensity level areas through multimodal data fusion and locates the coordinated abnormal areas that trigger energy-saving early warning signals. It not only achieves rapid response and accurate tracing of energy anomalies, avoids data island phenomena, but also provides an accurate basis for differentiated energy-saving regulation.

[0018] The present invention formulates a targeted zoning energy control strategy, and accurately configures differentiated control instructions and control rates according to the energy consumption intensity levels and energy-saving warning characteristics of different areas. It not only fully releases the energy-saving potential of different areas, but also achieves a dynamic balance between energy supply and demand, effectively improves the scientific nature and effectiveness of control measures, and promotes the transformation of building energy management from extensive to refined.

[0019] The present invention analyzes the energy-saving effect performance corresponding to energy regulation, corrects the energy regulation strategy parameters and triggers a new round of cross-domain data correlation analysis, so as to achieve continuous advancement of energy-saving goals through this closed-loop optimization mechanism, and significantly improve the level of building data governance and energy-saving and consumption-reduction capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0021] Figure 1 This is a block diagram of the system module structure provided by the first embodiment of the present invention.

[0022] Figure 2 This is a logic flow chart of the data analysis module in the first embodiment of the present invention.

[0023] Figure 3 A flowchart of the method steps provided for the second embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] Example 1 like Figure 1 As shown, in the first embodiment of the present invention, a building data governance system based on multimodal data fusion is provided, including: a data acquisition module, a data analysis module, an energy control module and an energy-saving optimization module.

[0026] The data acquisition module is connected to the data analysis module, the data analysis module is connected to the energy regulation module, the energy regulation module is connected to the energy-saving optimization module, and the energy-saving optimization module is connected to the data analysis module.

[0027] The data acquisition module obtains real-time operating data of the target building's multi-source perception network.

[0028] In a preferred embodiment of the present invention, the real-time operation data acquisition of the target building multi-source perception network includes: based on the current sensors, temperature and humidity sensors, light sensors and fluid flow sensors deployed in the target building, collecting monitoring data of the detection range covered by various sensors at their fixed deployment positions.

[0029] The minimum detection coverage of each type of sensor is extracted, and the target building space is discretized into several spatial coordinate units based on it.

[0030] It should be noted that the detection coverage parameters of the aforementioned sensors can refer to the standard detection space scope clearly defined in the technical specifications provided by the sensor manufacturer. To ensure the standardization and uniformity of spatial division, the coverage of this standard detection space scope must be standardized and defined in a regular rectangular grid format, providing a precise and consistent reference unit for subsequent spatial discretization operations.

[0031] The monitoring data of each sensor is collected into the corresponding spatial coordinate unit according to the spatial mapping relationship, and the current fluctuation data, environmental change data, light intensity distribution data and medium flow data of each spatial coordinate unit are output in real time.

[0032] It should be noted that the above-mentioned specific examples of real-time operating data are explained as follows: current fluctuation data is used to directly reflect the operating load of electrical equipment such as air conditioning, lighting, and office equipment in the target building, and at least includes the real-time current effective value, instantaneous power peak, power factor, and current harmonic distortion rate, which are core indicators for determining electricity energy consumption. Environmental change data is associated with the HVAC operation intensity of the target building, including at least ambient temperature and humidity. When the ambient temperature and humidity deviate from the comfort range, the air conditioning load will increase significantly, forming a strong coupling relationship with electricity consumption. Light intensity distribution data directly corresponds to the lighting energy consumption of the target building, and at least includes light intensity, light uniformity and natural light penetration. When there is insufficient light, the artificial lighting load increases, while excessive lighting will cause energy waste. It is the key basis for determining the lighting energy consumption characteristics. Medium flow data involves parameters such as hot and cold water flow, pipe pressure, and fluid temperature, reflecting the operating status of water supply and drainage or air-conditioning water circulation in the target building. Abnormal flow fluctuations often mean inefficient equipment operation or leakage, which is a direct reflection of water energy consumption.

[0033] like Figure 2 As shown, the data analysis module performs cross-domain data correlation analysis on the real-time operation data to generate an energy consumption distribution status diagram, and retrieves the collaborative abnormal area of ​​light intensity and medium through spatial coordinate mapping, and generates an energy-saving warning signal based on the topological position of the collaborative abnormal area.

[0034] In a preferred embodiment of the present invention, cross-domain data correlation analysis is performed on the real-time operation data to generate an energy consumption distribution status diagram, including: performing spatiotemporal consistency verification on current fluctuation data and environmental change data, identifying abnormal data points therein and applying sliding window mean replacement.

[0035] It should be noted that the above-mentioned spatiotemporal consistency verification process is: align the current fluctuation data sequence and the environmental change data sequence of the same spatial coordinate unit according to the acquisition timestamp, quantify the data mean and data standard deviation in the sliding time window, and if the absolute deviation of the value of a data point in the window relative to the mean exceeds a preset multiple of the standard deviation, the data point is marked as an abnormal data point.

[0036] According to the mapping relationship between the verified current fluctuation data and the environmental change data, the energy intensity level is divided into different levels for each spatial coordinate unit of the target building, and a connected domain search is performed on the units of the same level to form the initial energy intensity level area.

[0037] It should be noted that the process of dividing the energy consumption intensity level of each spatial coordinate unit of the above-mentioned target building is as follows: the correlation analysis of each parameter in the current fluctuation data and the environmental change data is carried out through the calculation of the Pearson correlation coefficient, and a Pearson correlation coefficient matrix containing all parameter combinations between the two data is constructed. Based on the actual influence of different parameter combinations on the energy consumption representation, a weight coefficient matrix consistent with the dimension of the Pearson correlation coefficient matrix is ​​constructed. The weight coefficient assignment process can be determined, for example, by the hierarchical analysis method or the historical data sensitivity analysis, and the Pearson correlation coefficient matrix is ​​multiplied by the weight coefficient matrix to obtain a fusion matrix, and the accumulated value of the fusion matrix elements is used as the comprehensive correlation intensity between the current fluctuation data and the environmental change data.

[0038] The comprehensive correlation strength and current fluctuation data of spatial coordinate units are integrated through the fuzzy logic method to map the energy consumption intensity level. The specific steps are as follows: the comprehensive correlation strength reflecting the rationality of the matching between energy consumption and the environment and the current fluctuation data reflecting the scale of power load are divided into several fuzzy categories according to their typical performance characteristics in actual scenarios. These fuzzy categories must completely cover the common value range of the parameters, and a reasonable overlapping interval must be retained between adjacent categories to effectively handle the ambiguity of the parameters when they are in a critical state. The corresponding fuzzy category is retrieved based on the actual parameter value of the spatial coordinate unit, forming a fuzzy category set consisting of the comprehensive correlation strength fuzzy category and the current fluctuation parameter fuzzy category.

[0039] Based on the fuzzy category set obtained above, the corresponding trigger rules are retrieved and matched in the preset fuzzy rule base. This preset fuzzy rule base is a condition-conclusion database built based on professional experience in energy consumption analysis and actual scenario characteristics. The core construction points include: i. Rule antecedent composition: It is formed by combining the fuzzy categories of the comprehensive correlation strength and the fuzzy categories of the current fluctuation parameters, and must fully cover all possible combinations of the two types of parameters. ii. Rule consequent determination: The consequent is the corresponding energy intensity level. The level classification must simultaneously reflect the dual dimensions of energy consumption scale and matching rationality. For example, when the comprehensive correlation strength is in a higher category and the current fluctuation parameter is in a lower category, the corresponding energy intensity level is lower. When the comprehensive correlation strength is in a lower category and the current fluctuation parameter is in a higher category, the corresponding energy intensity level is higher. iii. Rule set characteristic requirements: All parameter combinations have corresponding rules and there are no contradictory rule conclusions.

[0040] Based on the membership calculation results of the triggering rules, the energy consumption intensity levels and their memberships corresponding to each rule are comprehensively considered, and the energy consumption intensity quantitative indicators are calculated by the center of gravity method. Combined with the preset level mapping standard, the calculated energy consumption intensity quantitative indicators are mapped to specific energy consumption intensity levels. The mapping standard clearly defines the quantitative indicator range corresponding to each level.

[0041] The spatial gradient characteristics of the superimposed light intensity distribution data are used to perform sub-grid level correction on the boundary of the initial energy consumption intensity level area.

[0042] It should be noted that the boundary correction process of the above initial energy intensity level area is as follows: The light intensity distribution data is preprocessed, and the spatial illuminance data collected by the light intensity sensor is interpolated and mapped to the spatial coordinate grid system of the target building to generate a light intensity distribution matrix aligned with the initial energy consumption intensity level regional grid. On this basis, sub-grid level light intensity data is generated through spatial interpolation technology to achieve high-precision characterization of the light intensity details within the grid unit.

[0043] Based on the sub-grid level light intensity data, the Sobel operator or gradient operator is used to calculate the light intensity gradient amplitude of each sub-grid unit to quantify the spatial gradient characteristics of the light intensity distribution.

[0044] The area to be corrected is located through correlation analysis between the light intensity gradient characteristics and the initial energy consumption boundary: the boundary line of the initial energy consumption level area is extracted, the sub-grid units on both sides of the boundary line are simultaneously identified, and the light intensity gradient amplitude of the sub-grid units on both sides of the boundary line is calculated. If the gradient amplitude exceeds the light intensity gradient threshold set based on the architectural lighting design standard and there is a difference in the initial energy consumption level on both sides, the boundary segment is marked as an area to be corrected.

[0045] Perform fine-tuning on the marked boundary segments to be corrected: track the continuous distribution trajectory of high-gradient sub-grid units along the direction of the light intensity gradient, determine the actual physical boundary of the light intensity mutation and use it as a benchmark. If the physical boundary pointed by the light intensity gradient deviates from the initial energy consumption level boundary, correct the boundary line to the position of the light intensity mutation. If the energy consumption levels on both sides need to be further subdivided due to the difference in light intensity, add a new subdivision boundary in the sub-grid unit to achieve accurate division of the energy consumption level area.

[0046] After the correction is completed, the rationality of the results is verified through two aspects: first, check whether the corrected boundary is consistent with the physical characteristics of the light intensity distribution; second, compare the energy consumption level area before and after correction with the actual power consumption data. If there is a deviation, dynamically adjust the light intensity gradient threshold or correction amplitude, and finally output the energy consumption intensity level area corrected at the sub-grid level.

[0047] Output a three-dimensional energy consumption distribution state map that marks the topological boundaries of high energy consumption areas, medium energy consumption areas, and low energy consumption areas.

[0048] In a preferred embodiment of the present invention, the collaborative abnormal area search process includes: extracting fluctuation features and detecting events on the time series composed of the light intensity distribution data and the medium flow data of each spatial coordinate unit.

[0049] An event association rule is established, and the rule content is: when an illumination sudden change event is detected in the light intensity distribution data and a flow velocity sudden change event is detected synchronously in the medium flow data, and the time difference between the two is within the preset time window and occurs in the same spatial coordinate unit, the spatial coordinate unit is marked as a collaborative anomaly unit.

[0050] It should be noted that the specific detection process of the above-mentioned sudden illumination change event is: performing a first-order derivative operation of the time series on the light intensity distribution data to obtain the light intensity change rate. If the absolute value of the light intensity change rate at a certain spatial coordinate point is greater than the preset light intensity change rate threshold, it indicates that there is a sudden illumination change event.

[0051] The specific detection process of a sudden flow rate event is as follows: a second-order derivative operation of the time series is performed on the medium flow data to obtain the flow change acceleration. If the flow change acceleration of a certain spatial coordinate point exceeds the preset amplitude of the baseline value, it indicates that a medium sudden change event has occurred, where the baseline value refers to the running mean of the medium flow change acceleration calculated based on the historical normal operation data, and the preset amplitude can be exemplified as twice the standard deviation.

[0052] A connected domain search is performed on the collaborative anomaly units, and spatially continuous collaborative anomaly units are merged to form a collaborative anomaly region, and the coordinate set of its minimum circumscribed rectangle boundary is output.

[0053] The embodiment of the present invention divides the target area into energy consumption intensity level areas through multimodal data fusion and locates the coordinated abnormal area that triggers the energy-saving warning signal, which not only achieves rapid response and accurate tracing of energy anomalies, avoids data silos, but also provides an accurate basis for differentiated energy-saving regulation.

[0054] The energy control module dynamically adjusts the energy control strategy parameters distributed to different functional areas based on the energy consumption distribution state diagram and the energy-saving early warning signal.

[0055] In a preferred embodiment of the present invention, the dynamic adjustment process of the energy control strategy parameters of different functional areas includes: allocating load reduction instructions to the high energy consumption areas marked in the energy consumption distribution status diagram, and the instructions include equipment load reduction priority ranking and load reduction amplitude threshold.

[0056] It should be noted that the aforementioned equipment load reduction priority ranking refers to the hierarchical arrangement of equipment requiring load reduction in accordance with load reduction instructions in high-energy consumption zones. This ranking logic aims to maximize energy savings while minimizing functional impact. Specifically, load reduction is prioritized for equipment with a high energy consumption contribution and minimal impact on the building's core functions after load reduction, ensuring maximum energy savings for the same reduction amount while avoiding impact on critical functions. Prior to system development, this logic can be pre-calibrated based on the characteristics of different functional areas, taking into account the historical energy consumption contribution, functional importance, load reduction sensitivity, and operational stability of equipment, to facilitate its use during execution.

[0057] The load reduction threshold refers to the maximum load ratio or absolute upper limit allowed for a single device when performing load reduction operations on devices in high-energy consumption areas. The minimum load ratio allowed by the device's technical specifications can be used as a reference.

[0058] Dynamic balance control instructions are allocated to the medium energy consumption area, and the adaptive adjustment mode is matched according to the time-varying characteristics of the functional area to which it belongs.

[0059] It should be noted that the above-mentioned adaptive adjustment mode matching process is based on the current usage period of the functional area, the flow density and the environmental requirements. For example, the office area needs to maintain a high lighting brightness (300-500 lux) and air conditioning comfort level (24-26°C) during the working hours (9:00-18:00). During the non-working hours, the lighting brightness is automatically reduced (≤100 lux) and the air conditioning set temperature is increased (26-28°C) to reduce energy consumption.

[0060] Assign load maintenance instructions to low-energy consumption areas, lock the power supply to core equipment and shut down non-essential loads.

[0061] In a preferred embodiment of the present invention, the dynamic adjustment process of the energy control strategy parameters of different functional areas also includes: parsing the coordinates of the coordinated abnormal area in the energy-saving warning signal, and associating the energy consumption intensity level labels of the corresponding areas in the energy consumption distribution status diagram.

[0062] If the associated area is a high energy consumption area, the load reduction rate will be increased to a preset multiple of the benchmark to trigger a rapid response load reduction and switch to the backup energy supply mode.

[0063] If it is in the medium energy consumption zone, elastic rate correction is performed for the current adaptive adjustment mode.

[0064] If it is a low energy consumption area, the system will start step-by-step load recovery and maintain the power supply to the core equipment.

[0065] The embodiment of the present invention specifically formulates a zoning energy control strategy, and accurately configures differentiated control instructions and control rates based on the energy consumption intensity levels and energy-saving warning characteristics of different areas. This not only fully releases the energy-saving potential of different areas, but also achieves a dynamic balance between energy supply and demand, effectively improves the scientific nature and effectiveness of control measures, and promotes building energy management from extensive to refined.

[0066] The energy-saving optimization module collects comparison feedback data before and after the energy control operation, analyzes the energy-saving effect performance corresponding to the energy control, and accordingly modifies the energy control strategy parameters and triggers a new round of cross-domain data correlation analysis until the preset energy-saving target is achieved.

[0067] In a preferred embodiment of the present invention, the comparison feedback data collection process before and after the energy control operation includes: synchronously obtaining the current parameter set and environmental parameter set of each spatial coordinate unit in the steady-state period before control and the intermittent period after control, the current parameter set including the effective value of current, harmonic distortion rate and power factor, and the environmental parameter set including temperature, humidity and personnel distribution density.

[0068] By performing normalization processing on the current parameter set, the original current characteristics and the current reconstruction characteristics are obtained.

[0069] By performing thermal comfort index calculation on the environmental parameter set, the environmental reconstruction characteristics and the original characteristics of the environment are obtained.

[0070] It should be noted that the above-mentioned thermal comfort index is usually calculated using the internationally accepted PMV-PPD model. Specifically, the temperature of the environmental parameter set is used as the dry-bulb temperature, the humidity is converted into a correction coefficient for latent heat dissipation by calculating the difference between the saturated water vapor pressure and the actual water vapor pressure, and the population distribution density is associated with the human metabolic rate and activity status. The input is input into the PMV-PPD model to obtain the output thermal comfort index. The PMV-PPD model is an existing technology and will not be elaborated here.

[0071] Based on the comparative analysis of the original features and the reconstructed features, the current optimization rate matrix and the environmental comfort change rate matrix are generated with the spatial coordinate unit as the row index and the time slice as the column index. The two together constitute the comparison feedback data.

[0072] In a preferred embodiment of the present invention, the analysis of the energy-saving effect performance corresponding to energy regulation includes: assigning fusion weights to each element of the matrix according to the element discreteness of the current optimization rate matrix and the environmental comfort change rate matrix, thereby performing linear weighted fusion processing on the current optimization rate matrix and the environmental comfort change rate matrix to obtain the energy-saving effect index of each spatial coordinate unit, extracting the energy-saving effect index of all spatial coordinate units in the same energy consumption intensity level area, and taking the minimum value as the energy-saving efficiency of the level area, so as to analyze the energy-saving effect performance after energy regulation in different energy consumption areas in the energy consumption distribution status diagram.

[0073] In a preferred embodiment of the present invention, the energy control strategy parameter modification process includes: The energy-saving efficiency of different energy consumption areas in the energy consumption distribution status diagram is compared with the benchmark energy-saving efficiency threshold pre-calibrated for the corresponding energy consumption intensity level, and the energy-saving attributes of each energy consumption intensity level area are determined based on the energy-saving efficiency standard amplitude. The energy-saving attributes include one of oversaturation, moderate and inefficiency.

[0074] It should be noted that the above-mentioned determination of the energy-saving attributes of each energy consumption intensity level area based on the energy-saving efficiency standard amplitude specifically refers to: when the energy-saving efficiency of an energy consumption area is less than or equal to the first preset percentage of the benchmark energy-saving efficiency threshold pre-calibrated for its corresponding energy consumption intensity level, its energy-saving attribute is determined to be a low-efficiency attribute.

[0075] When the energy-saving efficiency of the energy consumption zone is greater than or equal to a second preset percentage of a baseline energy-saving efficiency threshold pre-calibrated for the corresponding energy consumption intensity level, the energy-saving attribute is determined to be an oversaturation attribute.

[0076] The rest are considered moderate attributes, where the first preset percentage < 1 < the second preset percentage.

[0077] According to different energy-saving attributes and energy consumption zone types, the key parameters in the energy control strategy are adjusted differently, and the load control accuracy is collaboratively optimized based on the energy-saving attributes maintenance cycle.

[0078] It should be noted that the above-mentioned differentiated adjustments specifically include: (A) Correction of high energy consumption areas: When the energy-saving efficiency is lower than the threshold, the following measures are taken: equipment load shedding priority is increased, the upper limit of load shedding is increased, and the backup energy switching delay is shortened.

[0079] When the energy-saving efficiency exceeds the threshold, the load shedding priority is lowered, the upper limit of the load shedding amplitude is reduced, and the power supply is switched back to the main energy source.

[0080] (B) Correction in the medium energy consumption zone: When the energy-saving efficiency is lower than the threshold: reduce the deviation tolerance of dynamic balance control and increase the adjustment frequency.

[0081] When the energy-saving efficiency exceeds the threshold: expand the deviation tolerance of dynamic balance control and reduce the adjustment frequency.

[0082] (C) Low-energy zone correction: When the energy-saving efficiency is lower than the threshold: improve the power locking accuracy of core equipment and expand the scope of shutting down non-essential loads.

[0083] When the energy-saving efficiency exceeds the threshold: the step-by-step load recovery rate is accelerated and the power lock threshold is dynamically relaxed.

[0084] (D) Global coordinated correction: For areas that remain inefficient for two consecutive control cycles: improve the execution accuracy level of load control instructions.

[0085] For areas that are oversaturated for three consecutive control cycles: extend the control time interval.

[0086] The embodiment of the present invention analyzes the energy-saving effect performance corresponding to energy regulation, corrects the energy regulation strategy parameters and triggers a new round of cross-domain data correlation analysis, so as to achieve continuous advancement of energy-saving goals through this closed-loop optimization mechanism, and significantly improve the level of building data governance and energy-saving and consumption-reduction capabilities.

[0087] Example 2 like Figure 3 As shown, a second embodiment of the present invention provides a building data governance method based on multimodal data fusion, including: obtaining real-time operation data of the multi-source perception network of the target building.

[0088] A cross-domain data association analysis is performed on the real-time operation data to generate an energy consumption distribution state diagram, and the collaborative abnormal area of ​​light intensity and medium is retrieved through spatial coordinate mapping, and an energy-saving early warning signal is generated based on the topological position of the collaborative abnormal area.

[0089] The energy consumption distribution state diagram and the energy-saving early warning signal are comprehensively used to dynamically adjust the energy control strategy parameters distributed to different functional areas.

[0090] Collect comparative feedback data before and after energy regulation operations, analyze the energy-saving effect corresponding to energy regulation, and accordingly modify the energy regulation strategy parameters and trigger a new round of cross-domain data correlation analysis until the preset energy-saving target is achieved.

[0091] The construction data governance method based on multimodal data fusion provided in the embodiment of the present invention has the same implementation principle and technical effects as those in the aforementioned system embodiment. For the sake of brief description, for matters not mentioned in the method embodiment, please refer to the corresponding content in the aforementioned system embodiment.

[0092] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. A building data management system based on multimodal data fusion, characterized by: include: Data acquisition module, which obtains real-time operation data of the target building's multi-source perception network; a data analysis module that performs cross-domain data correlation analysis on the real-time operation data to generate an energy consumption distribution state diagram, retrieves abnormal areas of light intensity and medium coordination through spatial coordinate mapping, and generates energy-saving warning signals based on the topological locations of the abnormal areas; An energy control module, which dynamically adjusts energy control strategy parameters distributed to different functional areas based on the energy consumption distribution state diagram and the energy-saving warning signal; The energy-saving optimization module collects comparative feedback data before and after energy control operations, analyzes the energy-saving effect corresponding to energy control, and accordingly modifies the energy control strategy parameters and triggers a new round of cross-domain data correlation analysis until the preset energy-saving target is achieved.

2. The building data management system based on multimodal data fusion according to claim 1 is characterized in that: The real-time operation data acquisition of the target building multi-source perception network includes: Based on the current sensors, temperature and humidity sensors, light sensors, and fluid flow sensors deployed in the target building, monitoring data is collected from the detection range covered by each type of sensor at its fixed deployment location; Extract the minimum detection coverage of each type of sensor, and discretize the target building space into several spatial coordinate units based on it; The monitoring data of each sensor is collected into the corresponding spatial coordinate unit according to the spatial mapping relationship, and the current fluctuation data, environmental change data, light intensity distribution data and medium flow data of each spatial coordinate unit are output in real time.

3. The building data management system based on multimodal data fusion according to claim 2 is characterized in that: Performing cross-domain data correlation analysis on the real-time operation data to generate an energy consumption distribution state diagram, including: Perform spatiotemporal consistency checks on current fluctuation data and environmental change data, identify abnormal data points, and apply sliding window mean replacement; Based on the mapping relationship between the verified current fluctuation data and the environmental change data, the energy intensity level of each spatial coordinate unit of the target building is divided into different levels. The connected domain search is performed on the units of the same level to form the initial energy intensity level area. Performing sub-grid level correction on the boundary of the initial energy consumption intensity level region by superimposing the spatial gradient characteristics of the light intensity distribution data; Output a three-dimensional energy consumption distribution state map that marks the topological boundaries of high energy consumption areas, medium energy consumption areas, and low energy consumption areas.

4. The building data management system based on multimodal data fusion according to claim 2 is characterized in that: The collaborative abnormal area retrieval process includes: Perform fluctuation feature extraction and event detection on the time series composed of light intensity distribution data and medium flow data of each spatial coordinate unit; Establish an event association rule, the content of which is: when an illumination sudden change event is detected in the light intensity distribution data and a flow velocity sudden change event is simultaneously detected in the medium flow data, and the time difference between the two is within a preset time window and occurs in the same spatial coordinate unit, the spatial coordinate unit is marked as a collaborative abnormal unit; A connected domain search is performed on the collaborative anomaly units, and spatially continuous collaborative anomaly units are merged to form a collaborative anomaly region, and the coordinate set of its minimum circumscribed rectangle boundary is output.

5. The building data management system based on multimodal data fusion according to claim 3 is characterized in that: The dynamic adjustment process of the energy control strategy parameters of different functional areas includes: Allocate load reduction instructions for high energy consumption areas marked in the energy consumption distribution diagram, including equipment load reduction priority and load reduction thresholds; Dynamic balance control instructions are allocated to medium energy consumption areas, and adaptive adjustment modes are matched according to the time-varying characteristics of the functional areas to which they belong; Assign load maintenance instructions to low-energy consumption areas, lock the power supply to core equipment and shut down non-essential loads.

6. The building data management system based on multimodal data fusion according to claim 5 is characterized in that: The dynamic adjustment process of the energy control strategy parameters of different functional areas also includes: Analyze the coordinates of the coordinated abnormal area in the energy-saving warning signal and associate the energy consumption intensity level label of the corresponding area in the energy consumption distribution status diagram; If the associated area is a high energy consumption area, the load reduction rate will be increased to a preset multiple of the baseline to trigger a rapid response load reduction and switch to the backup energy supply mode; If it is in the medium energy consumption zone, elastic rate correction is performed for the current adaptive adjustment mode; If it is a low energy consumption area, the system will start step-by-step load recovery and maintain the power supply to the core equipment.

7. The building data management system based on multimodal data fusion according to claim 3 is characterized in that: The comparison feedback data collection process before and after the energy regulation operation includes: Synchronously acquiring a current parameter set and an environmental parameter set of each spatial coordinate unit during a steady-state period before regulation and an intermittent period after regulation, wherein the current parameter set includes an effective current value, a harmonic distortion rate, and a power factor, and the environmental parameter set includes temperature, humidity, and a population distribution density; By performing normalization processing on the current parameter set, the original current characteristics and the current reconstruction characteristics are obtained; By performing thermal comfort index calculation on the environmental parameter set, the environmental reconstruction characteristics and the original characteristics of the environment are obtained; Based on the comparative analysis of the original features and the reconstructed features, the current optimization rate matrix and the environmental comfort change rate matrix are generated with the spatial coordinate unit as the row index and the time slice as the column index. The two together constitute the comparison feedback data.

8. The building data management system based on multimodal data fusion according to claim 7 is characterized in that: The energy-saving effect corresponding to the analysis of energy regulation includes: According to the element discreteness of the current optimization rate matrix and the environmental comfort change rate matrix, a fusion weight is assigned to each element of the matrix, and a linear weighted fusion process is performed on the current optimization rate matrix and the environmental comfort change rate matrix to obtain the energy-saving effect index of each spatial coordinate unit, and the energy-saving effect index of all spatial coordinate units in the same energy consumption intensity level area is extracted, and the minimum value is used as the energy-saving efficiency of the level area, so as to analyze the energy-saving effect performance after energy regulation in different energy consumption areas in the energy consumption distribution status diagram.

9. The building data management system based on multimodal data fusion according to claim 8 is characterized in that: The energy control strategy parameter correction process includes: Comparing the energy-saving efficiency of different energy consumption areas in the energy consumption distribution state diagram with the pre-calibrated benchmark energy-saving efficiency thresholds of the corresponding energy consumption intensity levels, and determining the energy-saving attributes of each energy consumption intensity level area based on the energy-saving efficiency standard amplitude, wherein the energy-saving attribute includes one of oversaturation, moderate and low efficiency; According to different energy-saving attributes and energy consumption zone types, the key parameters in the energy control strategy are adjusted differently, and the load control accuracy is collaboratively optimized based on the energy-saving attributes maintenance cycle.

10. A construction data management method based on multimodal data fusion, characterized in that: include: Obtain real-time operating data of the target building's multi-source perception network; Performing cross-domain data correlation analysis on the real-time operation data to generate an energy consumption distribution state diagram, retrieving abnormal areas of light intensity and medium coordination through spatial coordinate mapping, and generating energy-saving early warning signals based on the topological locations of the abnormal areas of coordination; Dynamically adjust the energy control strategy parameters distributed to different functional areas based on the energy consumption distribution state diagram and the energy-saving early warning signal; Collect comparative feedback data before and after energy regulation operations, analyze the energy-saving effect corresponding to energy regulation, and accordingly modify the energy regulation strategy parameters and trigger a new round of cross-domain data correlation analysis until the preset energy-saving target is achieved.

Images