Intelligent building information management system and method based on big data analysis

Abstract

The invention discloses an intelligent building information management system and method based on big data analysis, and relates to the technical field of data analysis, and the method comprises the following steps: 1, collecting historical and real-time data of equipment, carrying out the division according to a fixed interval, cleaning abnormal values, carrying out the normalization processing, and generating a multi-dimensional time sequence table containing multiple pieces of information; step 2, collecting recent state data of the equipment to form a query fragment, constructing a historical sliding window feature vector, comparing by using a DTW algorithm, and counting predicted closing duration distribution; step 3, calculating a predicted closing time length, detecting power consumption when the equipment is not closed for the time length, and judging a use state or an idle state according to a standard deviation and a threshold value; and step 4, when the equipment is idle, sending a closing prompt, processing and recording according to different operations of the user, and feeding back a misjudgment condition to a worker. According to the invention, the conditions of energy waste caused by forgetting to turn off building equipment and insufficient adaptability of a curing scene mode in the prior art can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to an intelligent building information management system and method based on big data analysis. Background Technology

[0002] Intelligent building information management uses information technology to monitor and control data such as the operating status and energy consumption of equipment within a building, aiming to achieve efficient and energy-saving management goals. With the development of technologies such as the Internet of Things and big data, it plays an increasingly important role in improving management efficiency and reducing operating costs. The ability to collect, analyze, and predict equipment data has also become a key indicator for measuring management level.

[0003] However, existing intelligent building information management technologies have many shortcomings. On the one hand, some users neglect or forget, causing equipment to remain running for extended periods without being used, resulting in significant energy waste. On the other hand, most buildings manage equipment using fixed scenario modes, such as timed on / off switching or fixed operating modes, which are difficult to adapt to changes in equipment operating needs at different times and in different usage scenarios. This not only fails to achieve precise energy saving but may also affect the normal use of equipment and user experience. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent building information management system and method based on big data analysis to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent building information management method based on big data analysis, the method comprising the following steps: Step 1: Collect historical and real-time data from the equipment, divide the data into fixed intervals, clean outliers, normalize the data, and generate a multi-dimensional time series table containing multiple information. Step 2: Collect recent device status data to form query fragments, construct historical sliding window feature vectors, use the DTW algorithm for comparison, and statistically analyze the distribution of expected shutdown durations; Step 3: Calculate the expected shutdown time. If the device is not shut down within the expected time, detect the power consumption and determine whether it is in use or idle based on the standard deviation and threshold. Step 4: Send a shutdown reminder when all devices are idle, process and record the user's different operations, and report any misjudgments to the staff.

[0006] In step 1, the on / off time, power consumption data, and corresponding timestamp information of different devices in the historical data are obtained; the data is then divided according to the chronological order based on a fixed time interval. Deploy smart meters or sensors to acquire power consumption data and on / off status of different devices at the same frequency at fixed time intervals; The acquired power consumption data is cleaned to remove data that is interrupted due to abnormal data transmission and data that does not conform to normal logic; the removed data is then filled with the previous valid data. The power consumption data of similar devices are normalized to obtain a multi-dimensional time series table for each device, which includes device number, timestamp, switch status, and normalized power consumption data.

[0007] In step 2, for each device, the device's operating status and power consumption data are collected at a fixed frequency, and the switch status and power consumption data in the most recent N time segments are extracted, preprocessed, and combined to form a query segment Q; A time segment represents the time interval between two adjacent data collection timestamps; In the historical multidimensional time series table, a sliding window feature vector set is constructed for each device, with a window length of N time segments; for each consecutive N time segments in the historical data, a corresponding feature vector is formed; for the entire historical data, a series of feature vector sets are obtained. The DTW algorithm is used to calculate the similarity between the query fragment Q and each historical window; all historical windows are sorted according to the similarity value, and the K historical windows with the highest similarity are selected. For the selected K historical windows, the distribution of the closing time offset within the windows is statistically analyzed: For each historical window, the number of time segments from the start of the window to the actual shutdown of the device is calculated; the distribution of the expected shutdown duration P corresponding to the query segment Q is obtained. The market closure distribution P includes the number of time segments required for each historical window device to close and the corresponding probability.

[0008] In step 3, the product of the number of time segments required for each historical window device to close in the market distribution P and the corresponding probability is calculated, and the sum of the K products is obtained by rounding up to obtain the predicted time segment required for device closure. When the time segment reaches the time segment required for the device to be predicted to shut down, and the device is in an active state, power consumption detection is triggered: Set a sliding window and calculate the standard deviation σ(t) of the device's operating power consumption within the set sliding window; Set a threshold τ. When σ(t)>τ, the device is considered to be in use, the idle timer is reset, and the timing restarts. When σ(t)≤τ and the duration of this low-fluctuation state exceeds the preset time threshold δ1, the device is considered to be in an idle state.

[0009] In step 4, when all devices are determined to be in an idle state, a shutdown reminder is sent to the user: If the user does not respond within the specified time limit δ2 or confirms the shutdown within the specified time limit δ2, the device will be shut down and the shutdown time will be recorded. If a user chooses not to turn off the device within the specified time limit δ2 or restarts the device within the specified time limit δ3, the device shutdown time is recorded, marked as a misjudgment, and reported back to the staff.

[0010] The intelligent building information management system based on big data analytics includes a data preprocessing module, a status analysis module, a status detection module, and an interactive feedback module. The data preprocessing module is used to collect historical and real-time data from the equipment, divide it into fixed intervals, clean outliers, normalize it, and generate a multi-dimensional time series table containing multiple information. The status analysis module is used to collect recent status data of the equipment to form query fragments, construct historical sliding window feature vectors, compare them using the DTW algorithm, and statistically analyze the distribution of expected shutdown duration. The status detection module is used to calculate the expected shutdown duration, detect power consumption when the equipment is not shut down within the expected duration, and determine whether it is in use or idle based on the standard deviation and threshold. The interactive feedback module is used to send shutdown reminders when all equipment is idle, process and record different user operations, and provide feedback on misjudgments to staff. The output of the data preprocessing module is connected to the input of the state analysis module; the output of the state analysis module is connected to the input of the state detection module; and the output of the state detection module is connected to the input of the interactive feedback module.

[0011] The data preprocessing module includes a data acquisition unit, a data cleaning unit, and a normalization unit. The data acquisition unit is used to collect historical and real-time data of the device through smart meters or sensors, integrate on / off time, operating power consumption and timestamp information, and divide the time sequence into time segments at fixed intervals; the data cleaning unit is used to remove outlier and jump values ​​and fill them with previous valid data; the normalization unit is used to standardize the power consumption data of similar devices and generate a multi-dimensional time sequence table containing device number, timestamp, on / off status and normalized power consumption. The output of the data acquisition unit is connected to the input of the data cleaning unit; the output of the data cleaning unit is connected to the input of the normalization unit; and the output of the normalization unit is connected to the input of the state analysis module.

[0012] The state analysis module includes a query generation unit, a feature extraction unit, and a similarity statistics unit; The query generation unit is used to extract the operating status and power consumption data of the device in the most recent N time segments to form a query segment Q; the feature extraction unit is used to construct a sliding window feature vector set in the historical data, each window containing device status information of N consecutive time segments; the similarity statistics unit is used to calculate the similarity between the query segment Q and the historical windows using the DTW algorithm, select the top K similar windows, and statistically analyze the closing time distribution and probability. The output of the query generation unit is connected to the input of the feature extraction unit; the output of the feature extraction unit is connected to the input of the similarity statistics unit; and the output of the similarity statistics unit is connected to the input of the state detection module.

[0013] The state detection module includes a shutdown prediction unit, a power consumption analysis unit, and a state judgment unit. The shutdown prediction unit is used to calculate the predicted shutdown duration based on the shutdown time distribution, and triggers power consumption detection if the shutdown is not completed within the time limit; the power consumption analysis unit is used to calculate the standard deviation of the operating power consumption within a sliding window and compare it with a threshold to determine the fluctuation status; the status judgment unit is used to mark the device as in use / idle state according to the standard deviation result, and determines it as idle if the continuous low fluctuation exceeds the threshold. The output of the prediction unit is connected to the input of the power consumption analysis unit and the interactive feedback module; the output of the power consumption analysis unit is connected to the input of the state judgment unit; and the output of the state judgment unit is connected to the input of the interactive feedback module.

[0014] The interactive feedback module includes a reminder triggering unit, an operation processing unit, and a misjudgment feedback unit; The reminder triggering unit is used to send a shutdown reminder to the user when all devices are idle; the operation processing unit is used to process the user's response within the time limit and record the operation result; the misjudgment feedback unit is used to mark misjudgment events and provide feedback to the staff. The output of the reminder triggering unit is connected to the input of the operation processing unit; the output of the operation processing unit is connected to the input of the misjudgment feedback unit.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: By acquiring detailed equipment data, including on / off times, operating power consumption, and timestamps, and performing cleaning and normalization processes, this invention can more accurately reflect the equipment's operating status. Compared with traditional monitoring methods, the data is more comprehensive and accurate, providing a reliable foundation for subsequent analysis. This invention utilizes historical data and the DTW algorithm to calculate similarity and predict equipment shutdown duration, achieving intelligent prediction of equipment operating status. This helps in advance planning of equipment maintenance and management, improving the efficiency and scientific nature of building equipment management, and reducing errors and uncertainties in manual judgment. This invention judges and reminds users to turn off equipment based on its idle status, effectively preventing prolonged idling, reducing energy consumption, achieving energy-saving goals, extending equipment lifespan, and improving resource utilization efficiency. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the steps of the intelligent building information management method based on big data analysis according to the present invention; Figure 2 This is a flowchart illustrating the intelligent building information management system based on big data analysis according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example: Figures 1-2 As shown, this invention provides a technical solution: an intelligent building information management method based on big data analysis, which includes the following steps: Step 1: Collect historical and real-time data from the equipment, divide the data into fixed intervals, clean outliers, normalize the data, and generate a multi-dimensional time series table containing multiple information. Step 2: Collect recent device status data to form query fragments, construct historical sliding window feature vectors, use the DTW algorithm for comparison, and statistically analyze the distribution of expected shutdown durations; Step 3: Calculate the expected shutdown time. If the device is not shut down within the expected time, detect the power consumption and determine whether it is in use or idle based on the standard deviation and threshold. Step 4: Send a shutdown reminder when all devices are idle, process and record the user's different operations, and report any misjudgments to the staff.

[0019] In step 1, the on / off time, power consumption data, and corresponding timestamp information of different devices in the historical data are obtained; the data is then divided according to the chronological order based on a fixed time interval. Deploy smart meters or sensors to acquire power consumption data and on / off status of different devices at the same frequency at fixed time intervals; The acquired power consumption data is cleaned to remove data that is interrupted due to abnormal data transmission and data that does not conform to normal logic; the removed data is then filled with the previous valid data. The power consumption data of similar devices are normalized to obtain a multi-dimensional time series table for each device, which includes device number, timestamp, switch status, and normalized power consumption data.

[0020] In step 2, for each device, the device's operating status and power consumption data are collected at a fixed frequency, and the switch status and power consumption data in the most recent N time segments are extracted, preprocessed, and combined to form a query segment Q; A time segment represents the time interval between two adjacent data collection timestamps; In the historical multidimensional time series table, a sliding window feature vector set is constructed for each device, with a window length of N time segments; for each consecutive N time segments in the historical data, a corresponding feature vector is formed; for the entire historical data, a series of feature vector sets are obtained. The DTW algorithm is used to calculate the similarity between the query fragment Q and each historical window; all historical windows are sorted according to the similarity value, and the K historical windows with the highest similarity are selected. For the selected K historical windows, the distribution of the closing time offset within the windows is statistically analyzed: For each historical window, the number of time segments from the start of the window to the actual shutdown of the device is calculated; the distribution of the expected shutdown duration P corresponding to the query segment Q is obtained. The market closure distribution P includes the number of time segments required for each historical window device to close and the corresponding probability.

[0021] In step 3, the product of the number of time segments required for each historical window device to close in the market distribution P and the corresponding probability is calculated, and the sum of the K products is obtained by rounding up to obtain the predicted time segment required for device closure. When the time segment reaches the time segment required for the device to be predicted to shut down, and the device is in an active state, power consumption detection is triggered: Set a sliding window and calculate the standard deviation σ(t) of the device's operating power consumption within the set sliding window; Set a threshold τ. When σ(t)>τ, the device is considered to be in use, the idle timer is reset, and the timing restarts. When σ(t)≤τ and the duration of this low-fluctuation state exceeds the preset time threshold δ1, the device is considered to be in an idle state.

[0022] In step 4, when all devices are determined to be in an idle state, a shutdown reminder is sent to the user: If the user does not respond within the specified time limit δ2 or confirms the shutdown within the specified time limit δ2, the device will be shut down and the device shutdown time will be recorded. If a user chooses not to turn off the device within the specified time limit δ2 or restarts the device within the specified time limit δ3, the device shutdown time is recorded, marked as a misjudgment, and reported back to the staff.

[0023] The intelligent building information management system based on big data analytics includes a data preprocessing module, a status analysis module, a status detection module, and an interactive feedback module. The data preprocessing module is used to collect historical and real-time data from the equipment, divide it into fixed intervals, clean outliers, normalize it, and generate a multi-dimensional time series table containing multiple information. The status analysis module is used to collect recent status data of the equipment to form query fragments, construct historical sliding window feature vectors, compare them using the DTW algorithm, and statistically analyze the distribution of expected shutdown duration. The status detection module is used to calculate the expected shutdown duration, detect power consumption when the equipment is not shut down within the expected duration, and determine whether it is in use or idle based on the standard deviation and threshold. The interactive feedback module is used to send shutdown reminders when all equipment is idle, process and record different user operations, and provide feedback on misjudgments to staff. The output of the data preprocessing module is connected to the input of the state analysis module; the output of the state analysis module is connected to the input of the state detection module; and the output of the state detection module is connected to the input of the interactive feedback module.

[0024] The data preprocessing module includes a data acquisition unit, a data cleaning unit, and a normalization unit. The data acquisition unit is used to collect historical and real-time data of the device through smart meters or sensors, integrate on / off time, operating power consumption and timestamp information, and divide the time sequence into time segments at fixed intervals; the data cleaning unit is used to remove outlier and jump values ​​and fill them with previous valid data; the normalization unit is used to standardize the power consumption data of similar devices and generate a multi-dimensional time sequence table containing device number, timestamp, on / off status and normalized power consumption. The output of the data acquisition unit is connected to the input of the data cleaning unit; the output of the data cleaning unit is connected to the input of the normalization unit; and the output of the normalization unit is connected to the input of the state analysis module.

[0025] The status analysis module includes a query generation unit, a feature extraction unit, and a similarity statistics unit. The query generation unit is used to extract the operating status and power consumption data of the device in the most recent N time segments to form a query segment Q; the feature extraction unit is used to construct a sliding window feature vector set in the historical data, each window containing device status information of N consecutive time segments; the similarity statistics unit is used to calculate the similarity between the query segment Q and the historical windows using the DTW algorithm, select the top K similar windows, and statistically analyze the closing time distribution and probability. The output of the query generation unit is connected to the input of the feature extraction unit; the output of the feature extraction unit is connected to the input of the similarity statistics unit; and the output of the similarity statistics unit is connected to the input of the state detection module.

[0026] The state detection module includes a shutdown prediction unit, a power consumption analysis unit, and a state judgment unit. The shutdown prediction unit is used to calculate the predicted shutdown duration based on the shutdown time distribution, and triggers power consumption detection if the shutdown is not completed within the time limit; the power consumption analysis unit is used to calculate the standard deviation of the operating power consumption within a sliding window and compare it with a threshold to determine the fluctuation status; the status judgment unit is used to mark the device as in use / idle state according to the standard deviation result, and determines it as idle if the continuous low fluctuation exceeds the threshold. The output of the prediction unit is connected to the input of the power consumption analysis unit and the interactive feedback module; the output of the power consumption analysis unit is connected to the input of the state judgment unit; and the output of the state judgment unit is connected to the input of the interactive feedback module.

[0027] The interactive feedback module includes a reminder triggering unit, an operation processing unit, and a misjudgment feedback unit; The reminder triggering unit is used to send a shutdown reminder to the user when all devices are idle; the operation processing unit is used to process the user's response within the time limit and record the operation result; the misjudgment feedback unit is used to mark misjudgment events and provide feedback to the staff. The output of the reminder triggering unit is connected to the input of the operation processing unit; the output of the operation processing unit is connected to the input of the misjudgment feedback unit.

[0028] In this embodiment, a 15-story office building is equipped with various electrical devices such as air conditioners, lighting fixtures, and printers. A smart building information management method based on big data analysis is applied to these devices. The specific implementation steps are as follows: Data Acquisition and Processing: Smart meters are deployed in the distribution boxes on each floor of the office building. Status sensors are also installed at each device to continuously collect data on the on / off times, power consumption, and corresponding timestamps of devices such as air conditioners, lighting fixtures, and printers at fixed 15-minute intervals. The acquired historical and real-time data are divided chronologically, cleaned to remove data with transmission errors, interruptions, and illogical jumps, and filled with the previous valid data. Then, the power consumption data of similar devices is normalized to generate a multi-dimensional time-series table containing device number, timestamp, on / off status, and normalized power consumption data.

[0029] Shutdown Duration Prediction: Taking a printer in an office building as an example, its operating status and power consumption data are collected at a fixed frequency of 15 minutes. The on / off status and power consumption data for the most recent 5 time segments (within 75 minutes) are extracted and combined after preprocessing to form a query segment Q. In the historical multidimensional time series table, a sliding window feature vector set is constructed for this printer, with a window length set to 5 time segments. For each consecutive 5 time segments in the historical data, a corresponding feature vector is formed, resulting in a series of feature vector sets. The DTW algorithm is used to calculate the similarity between the query segment Q and each historical window. All historical windows are sorted according to their similarity values, and the 3 historical windows with the highest similarity are selected. The distribution of the shutdown time offset in these 3 historical windows is statistically analyzed, i.e., the number of time segments from the start of the window to the actual shutdown of the printer is calculated, obtaining the expected shutdown duration distribution P corresponding to the query segment Q.

[0030] Device Status Judgment: Based on the shutdown duration distribution P, calculate the product of the number of time segments required for device shutdown in each historical window and the corresponding probability, sum the three products, and round up to obtain the predicted shutdown time segment required for the printer. When the time segment reaches the predicted shutdown time segment and the printer is not shut down, power consumption detection is triggered. A sliding window with a length of 3 time segments (i.e., 45 minutes) is set, and the standard deviation σ(t) of the printer's power consumption within this sliding window is calculated. A threshold τ=5 is set (set according to actual conditions). When σ(t)>5, the printer is considered to be in use, the idle timer is reset, and the timing restarts; when σ(t)≤5 and this low-fluctuation state lasts for more than the preset time threshold δ1=30 minutes, the printer is considered to be in an idle state.

[0031] Processing and Feedback: When all equipment in the office building is determined to be idle, the system sends a shutdown reminder to the user. If the user does not take any action within the specified time limit δ2=15 minutes, or confirms shutdown within 15 minutes, the system automatically shuts down the equipment and records the shutdown time. If the user chooses not to shut down within 15 minutes, or restarts the equipment within the specified time limit δ3=30 minutes, the system records the shutdown time, marks it as a misjudgment, and provides feedback to staff for subsequent optimization and adjustment of system parameters.

[0032] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A smart building information management method based on big data analysis, characterized by: The method includes the following steps: Step 1: Collect historical and real-time data from the equipment, divide the data into fixed intervals, clean outliers, normalize the data, and generate a multi-dimensional time series table containing multiple information. Step 2: Collect recent device status data to form query fragments, construct historical sliding window feature vectors, use the DTW algorithm for comparison, and statistically analyze the distribution of expected shutdown durations; Step 3: Calculate the expected shutdown time. If the device is not shut down within the expected time, detect the power consumption and determine whether it is in use or idle based on the standard deviation and threshold. Step 4: Send a shutdown reminder when all devices are idle, process and record the user's different operations, and report any misjudgments to the staff.

2. The intelligent building information management method based on big data analysis according to claim 1, characterized in that: In step 1, the on / off time, power consumption data, and corresponding timestamp information of different devices in the historical data are obtained; the data is then divided according to the chronological order based on a fixed time interval. Deploy smart meters or sensors to acquire power consumption data and on / off status of different devices at the same frequency at fixed time intervals; The acquired power consumption data is cleaned to remove data that is interrupted due to abnormal data transmission and data that does not conform to normal logic. Use the previous valid data to fill in the removed data; The power consumption data of similar devices are normalized to obtain a multi-dimensional time series table for each device, which includes device number, timestamp, switch status, and normalized power consumption data.

3. The intelligent building information management method based on big data analysis according to claim 2, characterized in that: In step 2, for each device, the device's operating status and power consumption data are collected at a fixed frequency, and the switch status and power consumption data in the most recent N time segments are extracted, preprocessed, and combined to form a query segment Q; A time segment represents the time interval between two adjacent data collection timestamps; In the historical multidimensional time series table, a sliding window feature vector set is constructed for each device, with a window length of N time segments; for each consecutive N time segments in the historical data, a corresponding feature vector is formed; for the entire historical data, a series of feature vector sets are obtained. The DTW algorithm is used to calculate the similarity between the query fragment Q and each historical window; all historical windows are sorted according to the similarity value, and the K historical windows with the highest similarity are selected. For the selected K historical windows, the distribution of the closing time offset within the windows is statistically analyzed: For each historical window, the number of time segments from the start of the window to the actual shutdown of the device is calculated; the distribution of the expected shutdown duration P corresponding to the query segment Q is obtained. The market closure distribution P includes the number of time segments required for each historical window device to close and the corresponding probability.

4. The intelligent building information management method based on big data analysis according to claim 3, characterized in that: In step 3, the product of the number of time segments required for each historical window device to close in the market distribution P and the corresponding probability is calculated, and the sum of the K products is obtained by rounding up to obtain the predicted time segment required for device closure. When the time segment reaches the time segment required for the device to be predicted to shut down, and the device is in an active state, power consumption detection is triggered: Set a sliding window and calculate the standard deviation σ(t) of the device's operating power consumption within the set sliding window; Set a threshold τ. When σ(t)>τ, the device is considered to be in use, the idle timer is reset, and the timing restarts. When σ(t)≤τ and the duration of this low-fluctuation state exceeds the preset time threshold δ1, the device is considered to be in an idle state.

5. The intelligent building information management method based on big data analysis according to claim 4, characterized in that: In step 4, when all devices are determined to be in an idle state, a shutdown reminder is sent to the user: If the user does not respond within the specified time limit δ2 or confirms the shutdown within the specified time limit δ2, the device will be shut down and the device shutdown time will be recorded. If a user chooses not to turn off the device within the specified time limit δ2 or restarts the device within the specified time limit δ3, the device shutdown time is recorded, marked as a misjudgment, and reported back to the staff.

6. A smart building information management system based on big data analysis, applied to the smart building information management method based on big data analysis as described in any one of claims 1-5, characterized in that: The system includes a data preprocessing module, a status analysis module, a status detection module, and an interactive feedback module; The data preprocessing module is used to collect historical and real-time data from the equipment, divide it into fixed intervals, clean outliers, normalize it, and generate a multi-dimensional time series table containing multiple information. The status analysis module is used to collect recent status data of the device to form query fragments, construct historical sliding window feature vectors, compare them using the DTW algorithm, and statistically analyze the distribution of expected shutdown duration. The status detection module is used to calculate the expected shutdown duration, detect power consumption when the device is not shut down within the expected duration, and determine the usage or idle status based on the standard deviation and threshold. The interactive feedback module is used to send a shutdown reminder when all devices are idle, process and record different user operations, and provide feedback on misjudgments to staff. The output of the data preprocessing module is connected to the input of the state analysis module; the output of the state analysis module is connected to the input of the state detection module; and the output of the state detection module is connected to the input of the interactive feedback module.

7. The intelligent building information management system based on big data analysis according to claim 6, characterized in that: The data preprocessing module includes a data acquisition unit, a data cleaning unit, and a normalization unit. The data acquisition unit is used to collect historical and real-time data of the device through smart meters or sensors, integrate on / off time, operating power consumption and timestamp information, and divide the time sequence into time segments at fixed intervals; the data cleaning unit is used to remove outlier and jump values ​​and fill them with previous valid data; the normalization unit is used to standardize the power consumption data of similar devices and generate a multi-dimensional time sequence table containing device number, timestamp, on / off status and normalized power consumption. The output of the data acquisition unit is connected to the input of the data cleaning unit; the output of the data cleaning unit is connected to the input of the normalization unit; and the output of the normalization unit is connected to the input of the state analysis module.

8. The intelligent building information management system based on big data analysis according to claim 7, characterized in that: The status analysis module includes a query generation unit, a feature extraction unit, and a similarity statistics unit. The query generation unit is used to extract the operating status and power consumption data of the device in the most recent N time segments to form a query segment Q; the feature extraction unit is used to construct a sliding window feature vector set in the historical data, each window containing device status information of N consecutive time segments; the similarity statistics unit is used to calculate the similarity between the query segment Q and the historical windows using the DTW algorithm, select the top K similar windows, and statistically analyze the closing time distribution and probability. The output of the query generation unit is connected to the input of the feature extraction unit; the output of the feature extraction unit is connected to the input of the similarity statistics unit; and the output of the similarity statistics unit is connected to the input of the state detection module.

9. The intelligent building information management system based on big data analysis according to claim 8, characterized in that: The state detection module includes a shutdown prediction unit, a power consumption analysis unit, and a state judgment unit. The shutdown prediction unit is used to calculate and predict the shutdown duration based on the shutdown time distribution, and triggers power consumption detection if the shutdown fails to occur within the time limit. The power consumption analysis unit is used to calculate the standard deviation of operating power consumption within a sliding window and compare it with a threshold to determine the fluctuation state; the state judgment unit is used to mark the device as in use / idle state based on the standard deviation result, and if the continuous low fluctuation exceeds the threshold, it is determined to be idle. The output of the prediction unit is connected to the input of the power consumption analysis unit and the interactive feedback module; the output of the power consumption analysis unit is connected to the input of the state judgment unit; and the output of the state judgment unit is connected to the input of the interactive feedback module.

10. The intelligent building information management system based on big data analysis according to claim 9, characterized in that: The interactive feedback module includes a reminder triggering unit, an operation processing unit, and a misjudgment feedback unit; The reminder triggering unit is used to send a shutdown reminder to the user when all devices are idle; the operation processing unit is used to process the user's response within the time limit and record the operation result. The misjudgment feedback unit is used to mark misjudgment events and provide feedback to staff; The output of the reminder triggering unit is connected to the input of the operation processing unit; the output of the operation processing unit is connected to the input of the misjudgment feedback unit.

Images