A method for intelligent regulation and control of a machine room environment by fusing multi-source sensing information

Abstract

The application relates to a machine room environment intelligent regulation and control method fusing multi-source sensing information, comprising the following steps: collecting temperature and humidity data of different areas in a machine room through a temperature and humidity sensor, collecting activity track and stay duration data of personnel in the machine room through a personnel positioning device, and transmitting the collected machine room environment data and personnel activity data to a data fusion module; in the data fusion module, aligning the machine room environment data and the personnel activity data according to time sequences, eliminating error data and outliers, and obtaining a unified format data set; calculating a correlation coefficient matrix between characteristic parameters of the machine room through a correlation analysis computer, screening out characteristic parameters with higher correlation than a preset correlation threshold, and the characteristic parameters comprising temperature, humidity, personnel access frequency and stay duration parameters.

Description

Technical Field

[0001] The present application relates to the field of information technology, and in particular to a method for intelligently controlling a computer room environment by integrating multi-source sensor information. Background Art

[0002] Computer room environmental management faces a contradiction between temperature and humidity control and the entry and exit activities of personnel. Frequent entry and exit of personnel will cause fluctuations in the temperature and humidity of the computer room, and the fluctuations in temperature and humidity will affect the stability of equipment operation. Existing computer room environmental control systems often separate temperature and humidity control from personnel management, and each operates independently. When the temperature and humidity sensors detect an abnormality, the air conditioning system will adjust it, but it cannot predict the impact of personnel activities. Similarly, the access control module only controls personnel entry and exit according to preset rules, without considering the impact on the computer room environment. This fragmented management method leads to delayed adjustment of the air conditioning system, energy waste, and a lack of targeted constraints on personnel activities.

[0003] How to use data to analyze the correlation between temperature and humidity changes and personnel activities to achieve coordinated control of the air-conditioning system and the access control module is an urgent problem to be solved. Based on the collected data, it is necessary to find the correlation between temperature and humidity and personnel activities through steps such as data cleaning, feature extraction, and model training, and then guide the coordinated control of the air-conditioning and access control modules to achieve refined and intelligent personnel management.

[0004] This solution proposes the following solutions to the above shortcomings: temperature and humidity data of different areas in the computer room, as well as personnel activity trajectories and length of stay, are collected through temperature and humidity sensors and personnel positioning devices, and these data are transmitted to the data fusion module for time series alignment and data cleaning to obtain a data set in a unified format. Next, characteristic parameter correlation analysis is performed to screen out parameters closely related to environmental changes, analyze and establish the association pattern between the computer room environment and personnel activities, determine the target structure and threshold, and then deploy the association pattern to the air conditioning and access control modules to monitor and dynamically adjust the equipment operating parameters in real time to achieve environmental control. At the same time, the control effect is displayed through the data visualization module, and the control strategy is evaluated. The control process is optimized using feedforward operations to ensure that the environmental parameters return to the target range, thereby achieving intelligent and efficient management of the computer room environment. Summary of the Invention

[0005] In order to solve the problems existing in the above-mentioned prior art, the purpose of this application is to provide an intelligent control method for computer room environment by integrating multi-source sensor information.

[0006] The present application discloses a method for intelligently controlling a computer room environment by integrating multi-source sensor information, comprising:

[0007] S101. Using temperature and humidity sensors to collect temperature and humidity data from different areas of the computer room, using personnel positioning equipment to collect data on the movement paths and duration of stay of personnel in the computer room, and transmitting the collected computer room environment data and personnel activity data to a data fusion module.

[0008] S102: In the data fusion module of step S101, the computer room environment data and the personnel activity data are aligned in time series, and erroneous data and outliers are removed to obtain a unified format data set;

[0009] S103, calculating the correlation coefficient matrix between the characteristic parameters of the computer room in step S101 through correlation analysis, and screening out characteristic parameters whose correlation is higher than a preset correlation threshold, the characteristic parameters including temperature, humidity, frequency of personnel entry and exit, and length of stay;

[0010] S104: Analyze the correlation pattern between the computer room environment and personnel activities based on the feature parameters with correlations higher than a preset threshold screened out in step S103 to obtain a target structure and target threshold for the correlation pattern between the computer room environment and personnel activities, where the target threshold includes a target computer room environment threshold and a target personnel activity threshold;

[0011] S105. Based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the air conditioning equipment control module, and the computer room temperature and humidity data are obtained in real time. It is determined whether the real-time temperature and humidity data of the computer room exceed the target computer room environment threshold range. If the target computer room environment threshold range is exceeded, the preset air conditioning equipment adjustment mechanism is triggered to dynamically adjust the air supply temperature, humidity, and air volume of the air conditioning equipment to restore the temperature and humidity;

[0012] S106. Based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the access control module to obtain real-time personnel entry and exit frequency and length of stay data to determine whether the target personnel activity threshold is exceeded. If the target personnel activity threshold is exceeded, the access control authorization time period and number of times are dynamically adjusted to control personnel entry and exit activities;

[0013] S107: Based on the control adjustments of the air conditioning and access control in steps S105 and S106, the environmental data of the computer room after the control adjustments are detected, including the temperature and humidity in different areas of the computer room and the activity trajectories and duration of stay of personnel in the computer room. The environmental data is generated into a curve graph through a preset data visualization module to obtain the environmental change trend of the computer room after the control adjustments;

[0014] S108. Evaluate the current control strategy and method based on the environmental change trend after the computer room is regulated in step S107. By presetting a target environmental change trend curve, fit the environmental change trend of the computer room after regulation with the pre-set target environmental change trend for use in regulating the computer room environment.

[0015] Preferably, in step S101, the temperature and humidity distribution is calculated by using the inverse distance weighted interpolation method through evenly arranged temperature and humidity sensors and network point coordinates, and a fine distribution map is obtained by increasing the density of sampling points in the hot spot area. At the same time, the positioning base station is used to receive the signal and calculate the positioning tag position, the noise is eliminated by curve smoothing, the movement trajectory and residence time are calculated, the data is segmented by time window, the temperature and humidity peak values ​​and the correlation values ​​of the positioning tag activity status are extracted, and data association analysis is performed to identify anomalies and activity patterns. The sensors are arranged in an area of ​​400 square meters at a spacing of 4 meters to form a 6×6 grid with a sampling frequency of 1 second / time. The high temperature area adopts 1-meter small grid encrypted sampling. The positioning base stations are installed at a spacing of 8 meters, and the signal is sent every 0.1 second. Through time window data alignment and feature extraction, support and confidence threshold analysis, typical abnormal scenes are identified.

[0016] Preferably, in the step S102, the computer room environment data and the positioning signal data are fused through millisecond timestamps, the data points are supplemented within the 10-second sampling interval using a cubic spline curve, and the outliers are processed by the box plot method to obtain the environment data set after abnormal processing. At the same time, according to the computer room building outline and the coverage range of the positioning base station, the coordinate sequence of the positioning signal data is extracted and smoothed, the environment data set and the positioning trajectory data are aligned through a 10-second time window, a mapping relationship table is established, and normalization is performed to form a unified format data set, synchronize the sensor and base station time, and handle data discontinuity, outliers and positioning drift problems.

[0017] Preferably, in step S103, by analyzing the paired data of the temperature and humidity sensors in the computer room at the same grid point, applying median smoothing and time window statistics, the temperature and humidity time series and personnel activity parameters are obtained, the parameters are normalized and the Pearson correlation coefficient is calculated, the correlation coefficient matrix is ​​constructed, and the dimension reduction is performed through hierarchical clustering and Euclidean distance calculation, and temperature and humidity are selected as feature parameters.

[0018] Preferably, in the step S104, a random forest regressor is used to establish a mapping between temperature and humidity and the patterns of personnel activities through time series analysis of the measured data of the computer room environment and the recorded data of personnel activities for 30 consecutive days, and the model parameters are optimized through grid search. The data set is divided into training, verification and test sets, and the data is normalized. Based on the prediction deviation distribution, the temperature and humidity threshold intervals and the activity parameter dividing values ​​are divided, and a threshold rule table is constructed. The parameters are adjusted by verifying the data set, the optimal model is determined, and the prediction accuracy of the threshold rule is verified on the test set, and finally the temperature and humidity threshold intervals of the computer room and the personnel activity threshold intervals are obtained.

[0019] Preferably, in the step S105, a temperature and humidity sensor array and personnel activity monitoring are arranged in the computer room, 4-20 mA analog signals are collected at intervals of 30 seconds and converted into digital quantities, temperature and humidity data are recorded, and the changing trends of hot spots and personnel gathering areas are calculated. A fuzzy controller is used to generate air conditioning adjustment instructions, which are sent to the air conditioning controller through the 485 bus protocol for dynamic adjustment of temperature, humidity and air volume. The sensors are arranged in a 6×6 grid dot matrix, and the hot spot area is determined based on the surrounding temperature difference. The fuzzy controller adjusts the adjustment step according to the deviation level, and updates the control rules in real time to adapt to the temperature and humidity response characteristics, so that the environmental parameters can quickly return to the target threshold range. In actual operation, the controller gives priority to adjusting the air supply volume of the air conditioner according to the temperature increase and personnel activity in the cabinet area, while the humidity adjustment is performed through the proportion of the humidifier switching time, and the response characteristic curve is used to optimize the control parameters.

[0020] Preferably, in step S106, by analyzing the access control card reader and positioning base station data, combined with temperature and humidity monitoring, the entry and exit and stay of personnel in the computer room area are finely managed. The system counts personnel activities in a 30-minute cycle, calculates environmental load indicators, and sets access control rules accordingly, including entry and exit frequency and stay time limits. Personnel have different entry and exit permissions according to their authorization levels. The system evaluates card swiping applications in real time, locks and issues warnings for violations. During peak hours or when the environmental load is too high, the system automatically adjusts the access control rules to restrict personnel activities, and the authorization rules are combined with operating procedures.

[0021] Preferably, in step S107, the temperature and humidity sensors and positioning base stations arranged at 6×6 grid points are used to synchronously collect and record the temperature, humidity and personnel position data in the computer room at intervals of 5 minutes, calculate the 10-minute average to eliminate fluctuations, generate an environmental data sequence with a standard timestamp, use linear interpolation to supplement the data points, draw temperature and humidity change curves, and mark the time periods exceeding the threshold and the control adjustment moments with broken line graphs. At the same time, the computer room area is divided into 4m×4m grids, the length of time personnel stay is accumulated and a regional density distribution map is generated, the temperature and humidity average values ​​and arrow marks indicating the movement path are superimposed, and the sensor uses a two-wire current output.

[0022] Preferably, in step S108, the computer room system generates an actual environmental curve by analyzing the temperature and humidity data collected every 5 minutes, and compares it with the reference curve of a normal working day, marking the control monitoring points that exceed the threshold. Based on the load forecast, the system generates a control instruction 30 minutes in advance, controls the air supply parameters of the air conditioner through the gear, and records the changes in environmental parameters. If the rate of change exceeds the threshold, access control is performed and personnel entry and exit are adjusted. The data on Tuesday with no activity is used as a reference. When the temperature and humidity are abnormal, the system uses a hierarchical adjustment mechanism and access control. If the temperature drop rate is insufficient, personnel entry is restricted and the stay time is adjusted.

[0023] The advantage of the intelligent control method for computer room environment that integrates multi-source sensor information described in the present application is that the method collects computer room temperature and humidity and personnel activity data, performs data fusion processing and correlation analysis on the data, and analyzes the correlation pattern between the computer room environment and personnel activities. Based on the correlation pattern between the computer room environment and personnel activities, the present invention monitors the computer room environment and personnel activities in real time. When the preset threshold is exceeded, the air-conditioning equipment parameters and access control authorization are automatically adjusted to maintain an ideal computer room environment. At the same time, the present invention also presents the environmental change trend through data visualization, and performs fitting analysis with the preset target trend, and continuously optimizes the control strategy. This method realizes the coordinated management of the computer room environment and personnel activities, improves the intelligence and accuracy of computer room environment control, and effectively ensures the safe and stable operation of computer room equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flowchart of the steps of a method for intelligently controlling a computer room environment by integrating multi-source sensor information described in this application;

[0025] Figure 2 This is a flow chart of step S103 of the method for intelligently controlling a computer room environment by integrating multi-source sensor information described in this application;

[0026] Figure 3 This is a flow chart of step S108 of a method for intelligently controlling a computer room environment by integrating multi-source sensor information described in this application. DETAILED DESCRIPTION

[0027] To further understand the content of the present invention, the present invention is described in detail with reference to the accompanying drawings and examples. The present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0028] like Figure 1-Figure 3 As shown, the method for intelligently controlling a computer room environment by integrating multi-source sensor information described in this application includes the following steps:

[0029] S101. Using temperature and humidity sensors to collect temperature and humidity data from different areas of the computer room, using personnel positioning equipment to collect data on the movement paths and duration of stay of personnel in the computer room, and transmitting the collected computer room environment data and personnel activity data to a data fusion module.

[0030] S102: In the data fusion module of step S101, the computer room environment data and the personnel activity data are aligned in time series, and erroneous data and outliers are removed to obtain a unified format data set;

[0031] S103, calculating the correlation coefficient matrix between the characteristic parameters of the computer room in step S101 through correlation analysis, and screening out characteristic parameters whose correlation is higher than a preset correlation threshold, the characteristic parameters including temperature, humidity, frequency of personnel entry and exit, and length of stay;

[0032] S104: Analyze the correlation pattern between the computer room environment and personnel activities based on the feature parameters with correlations higher than a preset threshold screened out in step S103 to obtain a target structure and target threshold for the correlation pattern between the computer room environment and personnel activities, where the target threshold includes a target computer room environment threshold and a target personnel activity threshold;

[0033] S105. Based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the air conditioning equipment control module, and the computer room temperature and humidity data are obtained in real time. It is determined whether the real-time temperature and humidity data of the computer room exceed the target computer room environment threshold range. If the target computer room environment threshold range is exceeded, the preset air conditioning equipment adjustment mechanism is triggered to dynamically adjust the air supply temperature, humidity, and air volume of the air conditioning equipment to restore the temperature and humidity;

[0034] S106. Based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the access control module to obtain real-time personnel entry and exit frequency and length of stay data to determine whether the target personnel activity threshold is exceeded. If the target personnel activity threshold is exceeded, the access control authorization time period and number of times are dynamically adjusted to control personnel entry and exit activities;

[0035] S107: Based on the control adjustments of the air conditioning and access control in steps S105 and S106, the environmental data of the computer room after the control adjustments are detected, including the temperature and humidity in different areas of the computer room and the activity trajectories and duration of stay of personnel in the computer room. The environmental data is generated into a curve graph through a preset data visualization module to obtain the environmental change trend of the computer room after the control adjustments;

[0036] S108. Evaluate the current control strategy and method based on the environmental change trend after the computer room is regulated in step S107. By presetting a target environmental change trend curve, fit the environmental change trend of the computer room after regulation with the pre-set target environmental change trend for use in regulating the computer room environment.

[0037] like Figure 1As shown, in step S101, the temperature and humidity sampling data of the sensor grid points are obtained, and the temperature and humidity spatial distribution data are obtained based on the sampling data by using inverse distance weighted interpolation; the distance from the positioning tag to the base station is determined based on the signal strength received by the positioning base station, and the noise of the positioning tag position coordinates is eliminated by using a curve smoothing function to obtain the moving trajectory points; the temperature and humidity spatial distribution data and the moving trajectory points are segmented according to the time window, and the corresponding data pairs of the temperature and humidity peak points and the positioning tag residence positions are extracted from the segmented data segments; if the temperature and humidity change deviates from the regional average value by more than a preset threshold, the positioning tag activity association value is obtained from the corresponding data pair, and the temperature and humidity anomalies and the positioning tag activity patterns are obtained by using data association analysis.

[0038] Specifically, in step S101, based on the grid point coordinates of the temperature and humidity sensors evenly arranged in the equipment room and the distance between adjacent sensors, temperature and humidity sampling data is obtained from each sensor node, and the spatial distribution of the grid point data is calculated using inverse distance weighted interpolation. The temperature and humidity distribution map of the hotspot area is obtained by increasing the sampling point density in the fluctuation area;

[0039] By placing a positioning base station in the computer room to receive the signal transmitted by the positioning tag, the distance from the positioning tag to the base station is converted according to the signal strength attenuation curve, the positioning tag position coordinates are established in the computer room coordinate system, the position coordinate data is de-noised based on the curve smoothing function, and the positioning tag movement trajectory points and residence time are calculated from the smoothed coordinate sequence;

[0040] The temperature and humidity distribution data and the positioning tag trajectory data are segmented into 10-second time windows. The corresponding data pairs of the temperature and humidity peak points and the positioning tag residence positions are extracted within the segmented data segments. The correlation value between the temperature and humidity change and the positioning tag activity status is obtained from the corresponding data pairs.

[0041] When the temperature and humidity changes deviate from the regional average by more than ±5 degrees Celsius or the relative humidity is 20%, the location and dwell time of the positioning tag during that period are extracted from the correlation value. Data association analysis based on a support threshold of 0.6 and a confidence threshold of 0.8 is used to identify patterns in temperature and humidity anomalies and positioning tag activity. Based on these patterns, historical data is compared to identify similar scenarios.

[0042] When deploying temperature and humidity sensors in the computer room, they are arranged in a grid with a spacing of 4 meters. For a single area of ​​400 square meters, 36 sensor nodes are deployed to form a 6×6 grid array. The sensor sampling frequency is 1 second, and data is collected. The temperature measurement accuracy is 0.5 degrees Celsius, and the relative humidity measurement accuracy is 2%.

[0043] For high-temperature areas, small grid sampling points are added to the standard grid with a small grid spacing of 1 meter to construct a detailed temperature and humidity distribution of the hotspot area;

[0044] When performing inverse distance weighting, the distance attenuation exponent is set to 2, and the closer the sampling point is, the greater the weight is, so as to achieve continuous and smooth interpolation of temperature and humidity distribution;

[0045] Positioning base stations were installed on the walls of the equipment room at intervals of 8 meters. They used 2.4 GHz signals, and the positioning tags transmitted signals every 0.1 seconds. The signal strength attenuation curve used a logarithmic attenuation model. At a distance of 1 meter, the signal strength was -40 dBm, with an attenuation factor of 3. When the base station received a signal strength of -70 dBm, the ranging error was less than 0.5 meters. The curve smoothing function used a five-point average. For dwell determination, a stay of more than 30 seconds within a 2-meter range was counted as a dwell.

[0046] When aligning the time window data, the sampling interval of temperature and humidity distribution data is 1 second, and the sampling interval of positioning trajectory data is 0.1 second. The data is segmented by 10-second time windows, and the maximum and minimum points of temperature and humidity data and the positioning trajectory residence points are extracted in each time window;

[0047] If the temperature change in the split window exceeds 5 degrees Celsius or the relative humidity change exceeds 20%, the activity trajectory features of the positioning tag in the window are extracted;

[0048] In the data association analysis, a support threshold of 0.6 indicates that similar temperature and humidity anomalies and human activity patterns appear in at least 60% of all time windows, and a confidence threshold of 0.8 indicates that when temperature and humidity anomalies occur, there is an 80% probability that they correspond to a specific human activity pattern;

[0049] Typical abnormal scenarios in the computer room include a sudden temperature drop of 5°C near the air conditioning outlet, a local temperature rise of 2°C due to personnel staying in front of equipment for more than 3 minutes, and a temperature fluctuation of 4°C due to maintenance work in the rear area of ​​a cabinet.

[0050] Similar scenes are judged based on the temperature change amplitude, humidity change amplitude, personnel location deviation less than 1 meter, and residence time deviation less than 30 seconds;

[0051] In actual applications, temperature and humidity grid monitoring covers all areas of the computer room, including the equipment area, aisle area, and air-conditioning area, with a focus on monitoring equipment-intensive areas; personnel positioning covers the passages inside and outside the computer room and the equipment maintenance passages, realizing the correlation monitoring of personnel activities and environmental changes. Temperature and humidity anomalies and personnel activity patterns include short-term temperature rises caused by maintenance work, sudden temperature changes caused by equipment switching on and off, and airflow disturbances caused by maintenance operations.

[0052] like Figure 1As shown, in step S102, according to the sampling timestamp of the computer room environment data and the sampling timestamp of the positioning signal data, a cubic spline curve is used to supplement the data points to obtain the environment positioning fusion data; for the environment positioning fusion data, the quartiles and interquartile range of the environment parameters are calculated through a box plot, the abnormal values ​​beyond the quartile range are truncated and the environment data set after abnormal processing is obtained by forward data filling; according to the positioning signal data, the excessive coordinate points are limited according to the adjacent coordinate spacing threshold and smoothed based on the distance threshold to obtain the positioning trajectory status mark; for the environment data set after abnormal processing and the positioning trajectory status mark, they are segmented and aligned through a time window and a mapping relationship table of environment parameters and trajectory status is established, and the environment parameter values ​​and trajectory status values ​​in the mapping relationship are normalized to obtain a unified format data set.

[0053] Specifically, in step S102, based on the sampling timestamps of the computer room environment data and the positioning signal data, a timestamp normalization format with a time accuracy of milliseconds is used to supplement data points within 10 seconds of the computer room environment data sampling interval using a cubic spline curve, and the original data are merged according to the timestamps to obtain environmental positioning fusion data under a unified time scale;

[0054] From the environmental positioning fusion data under a unified time scale, the quartiles and interquartile ranges of environmental parameters are calculated using box plots. Data that exceeds the upper and lower quartiles by 1.5 times the interquartile range are marked as outliers. Outliers are truncated according to the upper and lower quartiles, and missing points are filled with forward data to obtain the outlier-processed environmental dataset.

[0055] Based on the coordinate range of the computer room building outline and the coverage range of the positioning base station, a coordinate sequence is extracted from the positioning signal data. The out-of-limit coordinate points are limited according to the adjacent coordinate spacing threshold, and the coordinate points are smoothed based on the 4-meter distance threshold to obtain the corresponding status mark of the positioning trajectory. The boundary of the trajectory points that exceed the building outline or base station coverage is defined. The environmental dataset and the positioning trajectory data are segmented and aligned within a 10-second time window. A mapping relationship table between environmental parameters and trajectory status is established within the time window based on the status mark. The environmental parameter values ​​and trajectory status values ​​in the mapping relationship are normalized, and a unified format dataset of environmental parameters and positioning trajectory is obtained from the normalized data.

[0056] The time synchronization between the temperature and humidity sensors in the equipment room and the positioning base station adopts the network time protocol. The timestamp accuracy reaches the millisecond level. The temperature and humidity data are collected every 10 seconds, and the positioning signal is collected every 1 second. In the alignment process, cubic spline interpolation is used to supplement the temperature and humidity data points. The interpolation interval is 10 seconds, and 9 interpolation data points are generated within this interval. The timestamp uses a 13-bit length format and is accurate to milliseconds. For example, 163888888888 represents 20:21:28:888 milliseconds on December 7, 2021;

[0057] In environmental data outlier detection, the quartile Q1 of temperature data is 22 degrees Celsius, Q3 is 26 degrees Celsius, the interquartile range (IQR) is 4 degrees Celsius, and the outlier judgment range is 16-32 degrees Celsius. The Q1 of humidity data is 45%, Q3 is 55%, and the IQR is 10%. The outlier judgment range is 30%-70%. Data outside the judgment range is truncated according to the boundary value. For temperatures below 16 degrees Celsius, the value is 16 degrees Celsius, and for temperatures above 32 degrees Celsius, the value is 32 degrees Celsius.

[0058] When processing positioning signals, the computer room building outline adopts a rectangular coordinate system, 40 meters long and 25 meters wide. Four positioning base stations are installed at the four corners. The positioning tag transmits power of 4dBm, and the base station receives with a sensitivity of -90dBm. In open environments, the ranging accuracy is better than 1 meter. A distance between adjacent coordinate points exceeding 4 meters is considered an abnormal jump and is processed using distance limiting to maintain the continuity of the motion trajectory. For coordinate points outside the building outline, the nearest boundary point is used instead. Within a 10-second time window, the temperature and humidity data contains 10 data points, and the positioning data contains 10 coordinate points.

[0059] In the environmental parameter mapping, the temperature value range of 16-32 degrees Celsius is normalized to the interval of 0-1, and the humidity value range of 30%-70% is normalized to the interval of 0-1;

[0060] In the trajectory state value, the coordinate point projection on the x-axis is normalized to the range of 0-1, and the y-axis projection is normalized to the range of 0-1;

[0061] For boundary coordinate points, the x-axis takes a value of 0 or 1, and the y-axis takes a value of 0 or 1;

[0062] In actual application scenarios, temperature and humidity data alignment processing eliminates the data discontinuity problem caused by differences in sampling frequencies of different sensors. Environmental data outlier processing avoids data deviations caused by sensor failures, line interference, and other reasons. Positioning signal trajectory smoothing solves positioning drift caused by multipath effects and signal obstruction. Data normalization realizes unified measurement between different physical quantities, facilitating subsequent exploration of the correlation between environmental parameters and human activity patterns.

[0063] like Figure 1 and Figure 2As shown, in step S103, the paired sampling point data of the temperature and humidity sensors are obtained, and the sampling point data are processed by the median smoothing function to obtain the temperature and humidity time series; the timestamp data is obtained according to the computer room access control record, the entry and exit frequency sequence is statistically analyzed by the timestamp data, and the residence time sequence is obtained by accumulating the time intervals of discrete sampling points; the temperature and humidity time series, the entry and exit frequency sequence, and the residence time sequence are normalized, and the Pearson coefficient is calculated to obtain a correlation coefficient matrix; for the parameter pairs whose absolute values ​​of the correlation coefficients in the correlation coefficient matrix exceed the correlation coefficient threshold, a parameter group is obtained based on Euclidean distance clustering, and the parameter with the minimum correlation is selected from the parameter group as the feature parameter.

[0064] Specifically, in step S103, based on the spatial pairing relationship of the temperature and humidity sensors in the computer room at the same grid point, paired data points with a 5-minute sampling interval are extracted from the original temperature and humidity data set. The temperature and humidity data are processed using a median smoothing function to obtain a continuous and stable temperature and humidity time series. For the temperature mutation point, the average value of the data before and after is used instead of the mutation value.

[0065] The entry and exit timestamps of personnel were extracted from the computer room access control record data. The number of entries and exits was counted in 30-minute time windows to obtain an entry and exit frequency sequence. The position coordinates of personnel within each time window were calculated based on the positioning signal sampling points. The residence time sequence was obtained by accumulating the time intervals of discrete sampling points.

[0066] The temperature series, humidity series, entry and exit frequency series, and residence time series were normalized at a confidence level of 0.95. The Pearson correlation coefficient was used to calculate the correlation between each pair of series. The significance of the correlation coefficient was determined using the T test to obtain a 4×4 dimensional correlation coefficient matrix. For parameter pairs with an absolute correlation coefficient greater than 0.7 in the correlation coefficient matrix, the similarity between parameter groups was calculated based on the Euclidean distance. Hierarchical clustering was used to group similar parameters. From each parameter group, the parameter with the smallest correlation with the parameters outside the group was selected as the feature parameter to obtain the feature parameter combination after dimensionality reduction.

[0067]

[0068] , N represents the total number of parameter groups, d represents the dimension of each parameter group, x ik represents the k-th dimension value of the i-th parameter group, x jk represents the k-th dimension value of the j-th parameter group, σ j represents some standard deviation or coefficient of variation of the jth parameter group. This formula measures the similarity between parameter groups by calculating the Euclidean distance between all parameter groups and normalizing it.

[0069] The temperature and humidity sensors in the computer room are arranged at a 4-meter grid spacing, forming a 6×6 layout matrix. Both a temperature sensor and a humidity sensor are installed at each grid point. The sampling interval is once every 5 minutes. The temperature measurement range is 16-32 degrees Celsius, and the humidity measurement range is 30%-70%.

[0070] Median smoothing uses 5 data points to calculate the median. For data points with a sudden change in temperature exceeding 5 degrees Celsius or a sudden change in humidity exceeding 10%, the average of the two previous and next data points is used to replace the data point.

[0071] Two-way access control card readers are installed at the entrances and exits of the computer room. They record entry and exit timestamps accurate to the second and count the number of entries and exits within a 30-minute time window. For example, within the time window of 8:00-8:30, three people were recorded as entering and two people were recorded as leaving.

[0072] The positioning signal sampling interval is 1 second. The personnel's position coordinates are calculated based on the signal strength received by the base station. The time intervals between adjacent sampling points are accumulated to obtain the length of stay in the area. For example, the cumulative stay in the area in front of the cabinet within the time window of 9:00-9:30 is 800 seconds.

[0073] During normalization, the temperature data is subtracted from the minimum value of 16 degrees Celsius and divided by the range of 16 degrees Celsius. The humidity data is subtracted from the minimum value of 30% and divided by the range of 40%. The entry and exit frequency is divided by the maximum frequency of 10 times / 30 minutes. The residence time is divided by the time window of 1800 seconds.

[0074] The Pearson correlation coefficient was calculated using a confidence level of 0.95, a degree of freedom of 58, and a critical correlation coefficient of 0.254, a value greater than which was considered to be significantly correlated;

[0075] In the correlation coefficient matrix, the correlation coefficient between temperature and residence time is 0.82, the correlation coefficient between temperature and entry and exit frequency is 0.75, the correlation coefficient between humidity and temperature is 0.71, and the correlation coefficient between residence time and entry and exit frequency is 0.68;

[0076] Euclidean distance was used to calculate parameter similarity, with a distance threshold of 0.3. Temperature and duration of stay were grouped together, and temperature and entry / exit frequency were grouped together. Finally, temperature and humidity were selected as feature parameters, and duration of stay and entry / exit frequency could be derived from temperature.

[0077] In actual applications, the temperature of the computer room operating area shows a strong positive correlation with the length of time people stay there, and the temperature of the cabinet area also shows a positive correlation with the frequency of people entering and exiting. However, the correlation between humidity and other parameters is relatively weak.

[0078] Through parameter clustering and dimensionality reduction, the most representative temperature and humidity parameters were retained, which not only reflected the environmental status but also indirectly reflected the patterns of human activities. The correlation analysis results showed that there was a clear correlation between the computer room environmental parameters and human activities, and the correlation characteristics in different areas were different.

[0079] like Figure 1 As shown, in step S104, based on the time series of the measured data of the computer room environment and the recorded data of personnel activity, the temperature data, humidity data, personnel entry and exit frequency data, and residence time data are normalized according to the time series to obtain a normalized data set; a random forest regressor is established according to the normalized data set, and the number of decision trees and tree depth parameters of the random forest regressor are obtained through grid search; if the prediction deviation of the random forest regressor for the validation data set exceeds the standard deviation threshold, the temperature and humidity values ​​of the corresponding sample points are recorded, and the candidate environmental threshold values ​​are obtained according to the probability distribution of the temperature and humidity values; kernel density estimation is used to calculate the probability density curves of personnel entry and exit frequency and residence time, the activity parameter cutoff value is extracted from the inflection point of the probability density curve, and a threshold rule table is constructed according to the candidate environmental threshold values ​​and the activity parameter cutoff value.

[0080] Specifically, in step S104, based on the time series of the measured data of the computer room environment and the recorded data of personnel activities, the data of 30 consecutive days are divided into the first 20 days as the training data set, the middle 5 days as the validation data set, and the last 5 days as the test data set. The temperature data, humidity data, personnel entry and exit frequency data, and length of stay data are normalized respectively;

[0081] A random forest regressor was used to establish a mapping between temperature and humidity and human activity patterns. A grid search was used to determine the number and depth of decision trees. The root mean square error (RMSE) was calculated on a validation dataset. Parameters were adjusted through multiple rounds of iterations, and the optimal random forest parameter combination was determined from the minimum RMSE. The standard values ​​associated with temperature and humidity and human activity were predicted using random forest regression. Sample points in the validation dataset with prediction deviations exceeding twice the standard deviation were recorded. The temperature and humidity intervals were then divided based on the probability distribution of the recorded temperature and humidity values, and the interval boundaries were obtained as candidate environmental threshold values.

[0082] Kernel density estimation was used to calculate the probability density curve of personnel entry and exit frequency and length of stay. The activity parameter cutoff value was extracted from the curve inflection point. The temperature and humidity thresholds were combined with the activity parameter cutoff value to construct a threshold rule table. The prediction accuracy of the threshold rule was verified using a test data set, and the temperature and humidity threshold ranges and personnel activity threshold ranges of the computer room were obtained.

[0083] The data collected from the computer room environment is divided into two categories: normal working days and rest days. Each category contains 720 sampling points. The temperature range on working days is 20-28 degrees Celsius, and the temperature range on rest days is 18-25 degrees Celsius. The relative humidity on working days is 45%-65%, and the relative humidity on rest days is 40%-60%. Personnel activity records include entry and exit frequencies of 0-8 times / hour and stay durations of 0-45 minutes / hour.

[0084] Data normalization uses the maximum and minimum method to uniformly map temperature, humidity, and human activity data to the 0-1 range;

[0085] The random forest regressor setup includes 100 decision trees with a maximum tree depth of 8 layers, a feature random selection ratio of 0.8, and a sample random selection ratio of 0.7. The root mean square error calculated on the validation dataset shows a temperature prediction error of 0.8 degrees Celsius, a humidity prediction error of 3%, an entry and exit frequency prediction error of 1 time / hour, and a stay duration prediction error of 5 minutes / hour. After multiple rounds of parameter adjustment, the optimal tree depth was determined to be 6 layers.

[0086] The temperature and humidity threshold intervals are divided based on the distribution of prediction deviations. The sample points with temperature prediction deviations exceeding 1.6 degrees Celsius account for 5%, and the sample points with humidity prediction deviations exceeding 6% account for 8%.

[0087] According to the peak point of the deviation distribution curve, the temperature is divided into three intervals: 16-22 degrees Celsius, 22-28 degrees Celsius, and 28-32 degrees Celsius. The humidity is divided into three intervals: 30%-45%, 45%-60%, and 60%-70%. The kernel density curve of the personnel activity parameter shows a bimodal distribution. The inflection point of the entry and exit frequency is at 4 times / hour, corresponding to the normal inspection frequency, 8 times / hour corresponds to the equipment maintenance frequency, and the inflection point of the stay time is at 15 minutes / hour, corresponding to the inspection stop, and 30 minutes / hour corresponds to the maintenance operation stop. In the threshold rule table, the temperature range of 22-28 degrees Celsius and the humidity range of 45%-60% correspond to the normal inspection mode with an entry and exit frequency of 4 times / hour and a stay time of 15 minutes / hour.

[0088] In actual applications, the data set division takes into account the periodicity of the computer room's operating rules to ensure that the training data covers the complete working cycle. The temperature and humidity threshold ranges are set in combination with the computer room's air-conditioning control parameters, and the personnel activity threshold ranges are based on standard operating procedures. The threshold rule table reflects the standard operating mode under different environmental conditions, including daily inspections, equipment maintenance, and emergency response scenarios. Each scenario corresponds to a specific temperature and humidity range and personnel activity parameters.

[0089] like Figure 1As shown, in step S105, sampling data of the temperature and humidity sensor array is obtained, and the sampling data is recorded by a data collector according to the coordinates of the sensor layout position; the temperature and humidity change trends of the hot spot area and the personnel gathering area are calculated based on the sampling data, and an air conditioning adjustment instruction is generated from the change trend and the target threshold by a fuzzy controller; an adjustment sequence is generated by the air conditioning adjustment instruction, and the adjustment sequence sends a temperature adjustment instruction, a humidity adjustment instruction and an air volume adjustment instruction to the air conditioning controller through a bus protocol; a temperature and humidity response characteristic curve is obtained from the adjustment sequence by piecewise linear fitting, and the input and output membership function parameters of the fuzzy controller are adjusted online according to the slope of the response characteristic curve.

[0090] Specifically, in step S105, based on the temperature and humidity sensor array in the computer room and the personnel activity monitoring data, the data is sampled at 30-second intervals by a data collector to obtain a 4-20 mA analog signal, and a signal converter is used to convert the analog quantity into a standard digital quantity. The temperature and humidity sampling data are recorded according to the coordinates of the sensor layout position, and the temperature and humidity change trends of the hot spot area and the personnel gathering area are calculated from the sampling data. The environmental parameter deviation value is calculated from the temperature and humidity change trend and the target threshold value, and an air conditioning adjustment instruction is generated by a fuzzy controller. The supply air temperature adjustment step, the humidifier adjustment step, and the variable frequency fan speed adjustment step are calculated based on the temperature deviation, the humidity deviation, and the hot spot area position. An adjustment sequence is generated so that the amplitude of each adjustment does not exceed the set threshold. According to the adjustment sequence, the temperature adjustment instruction, the humidity adjustment instruction, and the air volume adjustment instruction are sent to the air conditioning controller through the 485 bus protocol. The temperature adjustment step is 0.5 degrees Celsius, the relative humidity adjustment step is 2%, and the fan speed adjustment step is 5% of the rated speed. It is judged from the environmental parameter sampling values ​​whether the parameters return to the target threshold range;

[0091] The temperature and humidity response characteristic curve is calculated by piecewise linear fitting. The temperature and humidity adjustment rates are obtained from the slope of the curve. The fuzzy control rules are adjusted online based on the response characteristic curve, and the input and output membership function parameters of the fuzzy controller are updated to achieve dynamic temperature and humidity adjustment.

[0092] The temperature and humidity sensors in the computer room are arranged in a 6×6 grid matrix. Each grid point is equipped with one temperature sensor and one humidity sensor. The sensors output a 4-20 mA standard current signal.

[0093] 4 mA corresponds to a temperature of 16 degrees Celsius or a relative humidity of 30%;

[0094] 20 mA corresponds to a temperature of 32 degrees Celsius or a relative humidity of 70%;

[0095] The sampling period is set to 30 seconds, during which temperature and humidity data are collected at 36 measurement points and the coordinates of the personnel's location are recorded. The hotspot area is determined when the average temperature of the surrounding eight measurement points is 3 degrees Celsius higher.

[0096] The fuzzy controller sets 7 temperature deviation levels, from -3 to +3 degrees Celsius, and 5 humidity deviation levels, from -10% to +10%. The adjustment step is determined according to the deviation level:

[0097] The temperature deviation is within 1 degree Celsius and the step size is 0.5 degrees Celsius;

[0098] The temperature deviation is 1-2 degrees Celsius and the step size is 1 degree Celsius;

[0099] If the temperature deviation is more than 2 degrees Celsius, the step size is 2 degrees Celsius;

[0100] Humidity regulation uses a similar classification:

[0101] Use 2% step size within 5% deviation;

[0102] 5%-10% with 5% step size;

[0103] The fan speed is adjusted within the range of 50%-100% of the rated speed, with a step size of 5%;

[0104] The air conditioning controller receives adjustment instructions through the 485 bus with a communication rate of 9600 bits / second and adopts the standard Modbus protocol. Each instruction contains the device address, function code, data length, and check code.

[0105] The temperature adjustment command is written into register 40001, the humidity adjustment command is written into register 40002, and the fan adjustment command is written into register 40003. After receiving the command, the controller responds and performs the adjustment action within 100 milliseconds.

[0106] The temperature and humidity response characteristic curve records the parameter changes within 300 seconds, recording a data point every 30 seconds. The 10 data points constitute the response curve. The slope of the curve is calculated by piecewise linear fitting. The typical value of the temperature response rate is 0.1 degrees Celsius / second, and the humidity response rate is 0.2% / second.

[0107] Based on the response characteristics, the fuzzy rules are updated. The temperature adjustment fast response range is 1-2 degrees Celsius, and the humidity adjustment fast response range is 3%-8%;

[0108] In actual operation, when it is detected that the temperature in the cabinet area has risen and there are people staying in this area, the controller will give priority to increasing the air supply volume of the air conditioner near this area. When the temperature rises rapidly, the controller will simultaneously reduce the air supply temperature and increase the air volume. The humidity adjustment is relatively slow, mainly adjusted by the proportion of the humidifier on and off time. The response characteristic curve reflects the dynamic process of air conditioning adjustment and is used to optimize the control parameters to avoid oscillation caused by too fast adjustment or environmental recovery affected by too slow adjustment.

[0109] like Figure 1 As shown, in step S106, the card swiping record collected by the access control card reader and the location coordinates collected by the positioning base station are received, and the area entry and exit records and the area residence time are obtained according to the card swiping record and the location coordinates according to a fixed time period; for the area entry and exit records and the area residence time, the regional temperature and humidity data are used to calculate the environmental load index, and if the environmental load index exceeds the preset threshold range, the regional access control rules are obtained according to the rule-based threshold judgment; the authorization level corresponding to the identity identification is obtained, if the authorization level is high-level authorization, entry and exit at any time period and unlimited stay are allowed, if the authorization level is low-level authorization, entry and exit during peak hours and continuous stay time are restricted; the identity identification is obtained by real-time card swiping application, the authorization level is queried according to the identity identification, and the authorization level and the area access control rules are used to determine whether the entry and exit restrictions are met, and if the entry and exit restrictions are not met, the access control actuator is triggered to lock.

[0110] Specifically, in step S106, based on the 1-second interval card swipe records collected by the access control card reader and the location coordinates collected by the positioning base station, the entry and exit records and location data of each identity identifier are counted according to a fixed time period of 30 minutes. The entry time, exit time, and identification number are extracted from the card swipe records, and the length of time spent in the area is obtained from the positioning data.

[0111] Calculate environmental load indicators based on regional temperature and humidity data. For areas exceeding the temperature and humidity threshold range, use rule-based threshold judgment to calculate the upper limit of personnel entry and exit frequency and the upper limit of stay time in the area. Generate regional access control rules based on the calculation results, including the hourly entry and exit limit and the single stay time limit.

[0112] There are 4 authorization levels based on personnel identification:

[0113] Advanced authorization allows entry and exit at any time and unlimited stay;

[0114] Intermediate authorization restricts entry and exit during non-working hours and continuous stays of more than 4 hours;

[0115] Low-level authorization restricts entry and exit during peak hours and continuous stays of more than 2 hours;

[0116] Temporary authorization only allows single entry and exit during working hours and stays of less than 1 hour;

[0117] The authorization level is obtained based on the real-time card swipe application and identity identification. The access control rule table is searched according to the time period and regional environmental status of the card swipe moment to determine whether the entry and exit frequency limit and the length of stay limit are met. For card swipe applications that exceed the limit, the access control actuator is locked, and the authorization is restricted through flashing red lights and beeps.

[0118] The computer room access control system uses a standard Wiegand protocol card reader with a reading distance of 5 cm, a response time of 0.1 seconds, and a card number length of 32 bits;

[0119] The positioning base station uses 2.4G active radio frequency positioning, with a signal coverage radius of 20 meters and a positioning accuracy of 0.5 meters;

[0120] During the peak period of 8:00-9:00, a typical area recorded 12 entries and 10 departures, with an average stay of 25 minutes per person. Peak area temperature rose by 2.5 degrees Celsius and relative humidity by 8%.

[0121] The environmental load index is calculated based on the temperature and humidity change rate. When the temperature change rate exceeds 1 degree Celsius / hour or the humidity change rate exceeds 5% / hour, the environmental load is judged to be too high;

[0122] In overloaded areas, the upper limit of entry and exit frequency is set at 6 times / hour, and the upper limit of stay time is set at 30 minutes / time;

[0123] Dynamic adjustment of regional access control rules. For example, if a temperature change rate of 1.2 degrees Celsius per hour is detected between 9:00 and 10:00, the upper limit of entry and exit frequency will be automatically reduced to 4 times per hour.

[0124] Identity authorization levels correspond to different job types:

[0125] Level 0 authorization corresponds to operation and maintenance management personnel, who have 24-hour access rights;

[0126] Level 1 authorization corresponds to patrol personnel, and entry and exit are restricted from 23:00 to 7:00;

[0127] Level 2 authorization corresponds to maintenance personnel, with access restricted from 8:00-10:00 and 14:00-16:00;

[0128] Level 3 authorization corresponds to temporary personnel, and entry and exit are only allowed from 9:00-11:00 and 15:00-17:00;

[0129] When the temperature in a high-temperature area exceeds 28 degrees Celsius, the authorized stay time for Level 2 and below will be automatically reduced by 50%;

[0130] The access control is implemented by magnetic lock, with a normally closed torque of 300 N and an unlocking time of 3 seconds;

[0131] When the card swipe application is rejected, the red light on the card reader flashes 3 times and the buzzer sounds 3 times;

[0132] Records show that between 9:00 and 10:00 a.m. on a weekday, a Level 2 authorized person swiped their card 100 times, of which 15 were rejected for exceeding the frequency of entry and exit limit and 8 were rejected for exceeding the length of stay limit;

[0133] Authorization rule records show that the regional temperature rose from 24 degrees Celsius to 26.5 degrees Celsius during that period, causing the system to automatically tighten entry and exit restrictions;

[0134] In actual applications, access control is linked to environmental monitoring. When the temperature in the cabinet area rises rapidly, the system automatically limits the number of people staying in the area. Authorization rules are combined with operating procedures to ensure that only necessary operation and maintenance activities are allowed during peak environmental load periods. Access control data records detailed entry and exit information, including card swiping time, location, authorization level, and access results, providing a basis for subsequent optimization of authorization rules.

[0135] like Figure 1 As shown, in step S107, the temperature and humidity data and personnel location data sent by the acquisition device are received, and the acquisition device records data at fixed time intervals after synchronizing the clock through the network time protocol; the average value of a fixed time period is calculated based on the temperature and humidity data to obtain an environmental data sequence, and the environmental data sequence includes a standard timestamp; temperature and humidity comparison data are extracted from the environmental data sequence, and the change in the number of people in the area is calculated based on the personnel location data, and a continuous change curve is obtained by linear interpolation; a line graph is generated for the continuous change curve, and if the temperature and humidity values ​​exceed the target threshold, the corresponding data segment is marked; a regional density distribution map is obtained by grid division, and the regional density distribution map maps color depth according to the length of time the personnel stay, and the regional temperature and humidity average values ​​are superimposed and displayed on the regional density distribution map.

[0136] Specifically, in step S107, based on the temperature and humidity sensor array and positioning base station arranged at a 6×6 grid point in the equipment room, the clocks of all collection devices are synchronized through the network time protocol, and temperature and humidity data and personnel location data are recorded at 5-minute intervals. The temperature and humidity data are averaged over a fixed time period of 10 minutes to eliminate data fluctuations and generate an environmental data sequence with a standard timestamp.

[0137] Extract temperature and humidity comparison data before and after air conditioning control from the environmental data series. Calculate the change in the number of people in the area based on the location data before and after access control. Generate associated data pairs based on the change in temperature and humidity and the change in the number of people. Use linear interpolation to supplement the data points within a 5-minute interval to obtain a continuous change curve.

[0138] Draw temperature and humidity curves in a line chart. The left axis shows temperature values ​​from 16 to 32 degrees Celsius, and the right axis shows humidity values ​​from 30% to 70%. The time axis displays the data for the last four hours. Data segments that exceed the target threshold are marked in red, and time stamps are added to control adjustment moments.

[0139] The computer room area is divided into 4m x 4m grids. The time people spend in each grid is accumulated and the color depth is mapped according to the hourly dwell time to generate a regional density distribution map. The average temperature and humidity values ​​for the area are superimposed on the map, and arrows are generated from the sequence of people's locations to indicate their movement paths.

[0140] The temperature and humidity sensor in the equipment room adopts a two-wire standard current output. The temperature range of 16-32 degrees Celsius corresponds to 4-20 mA, and the humidity range of 30%-70% corresponds to 4-20 mA.

[0141] In a 6×6 grid, each grid point is equipped with a temperature sensor and a humidity sensor, with adjacent grid points spaced 4 meters apart. All sensors are synchronized via a network time server, achieving clock accuracy better than 1 millisecond. Within a 5-minute sampling interval, 36 temperature and 36 humidity data points are collected, and a 10-minute average calculation eliminates the impact of air conditioning air supply fluctuations.

[0142] During the air conditioning process, records show that at 9:00 a.m., the average temperature in the cabinet area was 26.5 degrees Celsius, the relative humidity was 58%, and there were four people in the area at the same time. After the air conditioning was adjusted to cool down, the average temperature dropped to 24.2 degrees Celsius and the relative humidity dropped to 52% at 9:30 a.m., and the number of people in the area dropped to two. Linear interpolation of 5-minute interval data points yielded a temperature drop rate of 0.15 degrees Celsius / minute and a humidity drop rate of 0.4% / minute.

[0143] The line chart uses a dual Y-axis structure. The temperature curve corresponds to the left Y-axis with a scale interval of 2 degrees Celsius, and the humidity curve corresponds to the right Y-axis with a scale interval of 5%. The time axis displays the data of the last 4 hours with a scale interval of 15 minutes. When the temperature exceeds 28 degrees Celsius or the humidity exceeds 65%, the curve segment is marked in red. Vertical marking lines are added at 9:00 and 9:30 to mark the time when the air conditioning is adjusted. The computer room is divided into 36 areas according to a 4m x 4m grid, and the cumulative length of personnel stay in each area is calculated. The color depth mapping rule is that the stay time of 0-15 minutes is light blue, 15-30 minutes is blue, 30-45 minutes is dark blue, and more than 45 minutes is dark purple. The arrow mark indicates the direction of personnel movement. The arrow length is proportional to the movement speed. The typical inspection speed corresponds to an arrow length of 2 meters.

[0144] In actual applications, temperature and humidity data show that the temperature in the cabinet area changes dramatically, which is positively correlated with the length of time people stay. Through the personnel density distribution map, it is found that high-temperature areas often correspond to hot spots where people stay. The temperature and humidity trend curve reflects the effect of air conditioning regulation. Combined with personnel activity data, it intuitively shows the impact of personnel activities on the environment. The area division and arrow markings reveal the patterns of personnel activities, providing a basis for optimizing air conditioning control and personnel management.

[0145] like Figure 1 and Figure 3 As shown, in step S108, an environmental parameter data sequence collected by the temperature and humidity sensor is received, an actual environmental curve is generated according to the environmental parameter data sequence, and a temperature deviation value and a humidity deviation value are calculated by comparing the actual environmental curve with the target reference curve; for the monitoring period in which the temperature deviation value and the humidity deviation value exceed the preset deviation threshold, a load predictor is used to perform a feedforward operation, and the control target value and the advance control time are determined according to the feedforward operation result; an air conditioning supply parameter control instruction sequence is generated according to the control target value, the supply air temperature and the supply air humidity are adjusted by the control instruction sequence, and the environmental parameter change data during the gear adjustment process is collected; the temperature change rate and the humidity change rate are extracted from the environmental parameter change data, and if the temperature change rate and the humidity change rate exceed the preset threshold, access control is performed on the corresponding area until the environmental parameters return to the target reference curve range.

[0146] Specifically, in step S108, the actual environment curve is generated based on the data sequence collected every 5 minutes by the temperature and humidity sensors in the computer room. The temperature and humidity curve of a normal working day is selected from the operating data before the regulation as the target reference curve. The temperature deviation value and humidity deviation value are calculated after aligning the sampling points in time. The period where the deviation exceeds 1 degree Celsius for temperature or 5% for humidity is marked as a regulation monitoring point.

[0147] Based on the load forecaster, the environmental parameters of the monitoring period are fed forward, and the control target value is calculated based on the temperature and humidity change trend. The advance control time is set 30 minutes ahead from the control monitoring point, and the control instruction sequence is generated according to the temperature and humidity status at the advance control time.

[0148] The air supply parameters of the air conditioner are controlled according to the control target values ​​in the control instruction sequence:

[0149] The temperature high-level adjustment step is 2 degrees Celsius;

[0150] The temperature adjustment step is 1 degree Celsius in the middle position;

[0151] The temperature low gear adjustment step is 0.5 degrees Celsius;

[0152] The humidity high level adjustment step is 10%;

[0153] The humidity mid-range adjustment step is 5%;

[0154] Humidity low gear adjustment step 2%;

[0155] Record the environmental parameter change curve in each control cycle;

[0156] The air conditioning controller collects environmental response data within each control cycle, extracts the temperature and humidity change rates from the response data, implements access control for areas where the change rate exceeds the preset threshold, and adjusts the entry and exit restrictions until the environmental parameters return to the target curve range;

[0157] The data from the previous Tuesday, a normal working day, was used as a reference for temperature and humidity monitoring in the computer room. During that day, there was no equipment overhaul or maintenance activity, and the temperature remained stable at 24°C and 55% relative humidity. The current environmental curve shows that between 9:00 and 10:00, the temperature rose to 26°C and the relative humidity to 63%. These deviations exceeded the preset range, and the system marked this period as a control monitoring point. Load forecast calculations showed that between 8:30 and 9:00, the power of cabinets in area A increased by 20%, the personnel stay time increased by 15 minutes, the temperature rose by 0.1°C / minute, and the humidity rose by 0.3% / minute.

[0158] Based on these parameters, the system generates control instructions at 8:30 a.m., with the goal of reducing the temperature by 2 degrees Celsius and the humidity by 8% within 30 minutes;

[0159] The air conditioning gear control adopts a 3-level adjustment mechanism:

[0160] When the temperature deviation is greater than 2 degrees Celsius, the high-speed adjustment is activated and the supply air temperature is reduced by 2 degrees Celsius at one time;

[0161] When the temperature deviation is between 1-2 degrees Celsius, use the middle gear adjustment and reduce the supply air temperature by 1 degree Celsius twice;

[0162] When the temperature deviation is less than 1 degree Celsius, the low-speed adjustment is adopted, and the supply air temperature is reduced by 0.5 degrees Celsius in 4 steps, and the humidity is adjusted in a similar graded manner;

[0163] During the control period from 8:30 to 9:00, records showed that the temperature dropped from 26 degrees Celsius to 24.5 degrees Celsius, at a rate of 0.05 degrees Celsius per minute, and the humidity dropped from 63% to 58%, at a rate of 0.17% per minute. Because the temperature drop rate was lower than the expected 0.067 degrees Celsius per minute, the system issued an access control instruction, restricting entry to the area, adjusting the original limit of 4 people to 2 people, and shortening the stay time from 30 minutes to 15 minutes.

[0164] In actual operation, different areas exhibit different environmental characteristics. The cabinet area is sensitive to human activities, the temperature rises quickly, and a larger adjustment step is required. The temperature in the aisle area is relatively stable and suitable for small step adjustment. Access control is given priority to control high-density areas to reduce heat load by restricting human activities. Air conditioning is given priority to control the air supply area to quickly cool down by increasing the air supply volume. Appropriate control measures are selected according to the characteristics of each area to achieve precise control. Historical data shows that using this multi-level linkage control method, the regression time of environmental parameter deviation is shortened by an average of 50%.

[0165] In the description of this application, the technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

Claims

1. A method for intelligently controlling a computer room environment by integrating multi-source sensor information, characterized in that: The following steps are involved: S101. Using temperature and humidity sensors to collect temperature and humidity data from different areas of the computer room, using personnel positioning equipment to collect data on the movement paths and duration of stay of personnel in the computer room, and transmitting the collected computer room environment data and personnel activity data to a data fusion module. Specifically, the following steps are included: Obtaining temperature and humidity sampling data of sensor grid points, and obtaining temperature and humidity spatial distribution data using inverse distance weighted interpolation based on the sampling data; The distance between the positioning tag and the base station is determined based on the signal strength received by the positioning base station, and the curve smoothing function is used to eliminate the noise of the positioning tag position coordinates to obtain the moving trajectory point; Segmenting the temperature and humidity spatial distribution data and the movement trajectory points according to time windows, and extracting corresponding data pairs of temperature and humidity peak points and positioning tag residence positions from the segmented data segments; If the temperature and humidity variation deviates from the regional average by more than a preset threshold, the positioning tag activity association value is obtained from the corresponding data pair, and the temperature and humidity anomaly and the positioning tag activity pattern are obtained by data association analysis; S102: In the data fusion module of step S101, the computer room environment data and the personnel activity data are aligned in time series, and erroneous data and outliers are removed to obtain a unified format data set; specifically, the following steps are further included: According to the sampling timestamps of the computer room environment data and the positioning signal data, the cubic spline curve is used to supplement the data points to obtain the environmental positioning fusion data; For the environmental positioning fusion data, the quartiles and interquartile ranges of the environmental parameters are calculated through a box plot, and outliers beyond the quartile range are truncated and filled with forward data to obtain an outlier-processed environmental dataset; According to the positioning signal data, the out-of-limit coordinate points are limited according to the adjacent coordinate spacing threshold and smoothed based on the distance threshold to obtain the positioning trajectory status mark; For the environmental data set after abnormal processing and the positioning trajectory status mark, split and align them by time window and establish a mapping relationship table between environmental parameters and trajectory status, and normalize the environmental parameter values ​​and trajectory status values ​​in the mapping relationship to obtain a unified format data set; S103, calculating the correlation coefficient matrix between the characteristic parameters of the computer room in step S101 through correlation analysis, and screening out characteristic parameters whose correlation is higher than a preset correlation threshold, the characteristic parameters including temperature, humidity, frequency of personnel entry and exit, and length of stay; S104: Analyze the correlation pattern between the computer room environment and personnel activities based on the feature parameters with correlations higher than a preset threshold screened out in step S103 to obtain a target structure and target threshold for the correlation pattern between the computer room environment and personnel activities, where the target threshold includes a target computer room environment threshold and a target personnel activity threshold; S105. Based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the air conditioning equipment control module, and the computer room temperature and humidity data are obtained in real time. It is determined whether the real-time temperature and humidity data of the computer room exceed the target computer room environment threshold range. If the target computer room environment threshold range is exceeded, the preset air conditioning equipment adjustment mechanism is triggered to dynamically adjust the air supply temperature, humidity, and air volume of the air conditioning equipment to restore the temperature and humidity; S106. Based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the access control module to obtain real-time personnel entry and exit frequency and length of stay data to determine whether the target personnel activity threshold is exceeded. If the target personnel activity threshold is exceeded, the access control authorization time period and number of times are dynamically adjusted to control personnel entry and exit activities; S107: Based on the control adjustments of the air conditioning and access control in steps S105 and S106, the environmental data of the computer room after the control adjustments are detected, including the temperature and humidity in different areas of the computer room and the activity trajectories and duration of stay of personnel in the computer room. The environmental data is generated into a curve graph through a preset data visualization module to obtain the environmental change trend of the computer room after the control adjustments; S108. Evaluate the current control strategy and method based on the environmental change trend after the computer room is regulated in step S107. By presetting a target environmental change trend curve, fit the environmental change trend of the computer room after regulation with the pre-set target environmental change trend for use in regulating the computer room environment.

2. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 1, characterized in that: In step S103, the correlation coefficient matrix between the characteristic parameters of the computer room in step S101 is calculated by correlation analysis, and characteristic parameters with correlations higher than a preset correlation threshold are screened out. The characteristic parameters include temperature, humidity, frequency of personnel entry and exit, and length of stay parameters, including: Acquire paired sampling point data of temperature and humidity sensors, and process the sampling point data using a median smoothing function to obtain a temperature and humidity time series; Obtaining timestamp data based on the computer room access control records, using the timestamp data to calculate the entry and exit frequency sequence, and obtaining the residence time sequence by accumulating the time intervals of discrete sampling points; Normalizing the temperature and humidity time series, the entry and exit frequency series, and the residence time series, and calculating the correlation coefficient matrix using the Pearson coefficient; For parameter pairs whose absolute values ​​of correlation coefficients in the correlation coefficient matrix exceed a correlation coefficient threshold, a parameter group is obtained based on Euclidean distance clustering, and a parameter with minimum correlation is selected from the parameter group as a feature parameter.

3. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 1, characterized in that: In step S104, based on the feature parameters with correlations higher than a preset threshold selected in step S103, the computer room environment and personnel activity association pattern is analyzed to obtain a target structure and a target threshold for the computer room environment and personnel activity association pattern. The target threshold includes a target computer room environment threshold and a target personnel activity threshold, including: Based on the time series of the measured data of the computer room environment and the recorded data of personnel activities, the temperature data, humidity data, personnel entry and exit frequency data, and residence time data are normalized according to the time series to obtain a normalized data set; Establishing a random forest regressor based on the normalized data set, and obtaining the number of decision trees and tree depth parameters of the random forest regressor through grid search; If the prediction deviation of the random forest regressor for the validation dataset exceeds the standard deviation threshold, the temperature and humidity values ​​of the corresponding sample points are recorded, and a candidate environmental threshold value is obtained based on the probability distribution of the temperature and humidity values; Kernel density estimation is used to calculate the probability density curve of personnel entry and exit frequency and length of stay, and the activity parameter cutoff value is extracted from the inflection point of the probability density curve. A threshold rule table is constructed based on the environmental threshold candidate value and the activity parameter cutoff value.

4. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 1, characterized in that: In step S105, according to the target threshold in step S104, the computer room environment and personnel activity association mode is deployed to the air conditioning equipment control module, the computer room temperature and humidity data is obtained in real time, and it is determined whether the real-time temperature and humidity data of the computer room exceeds the target computer room environment threshold range. If it exceeds the target computer room environment threshold range, the preset air conditioning equipment adjustment mechanism is triggered to dynamically adjust the air supply temperature, humidity and air volume of the air conditioning equipment for temperature and humidity recovery, including: Acquire sampling data of the temperature and humidity sensor array, wherein the sampling data is recorded by a data collector according to the sensor arrangement position coordinates; Calculating the temperature and humidity change trends of hot spots and gathering areas based on the sampled data, and generating air conditioning adjustment instructions based on the change trends and target thresholds using a fuzzy controller; Generate an adjustment sequence through the air conditioning adjustment instruction, and send the temperature adjustment instruction, humidity adjustment instruction and air volume adjustment instruction to the air conditioning controller through the bus protocol; A temperature and humidity response characteristic curve is obtained from the adjustment sequence by piecewise linear fitting, and input and output membership function parameters of the fuzzy controller are adjusted online according to the slope of the response characteristic curve.

5. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 1, characterized in that: In step S106, based on the target threshold in step S104, the computer room environment and personnel activity association model is deployed to the access control module, and the personnel entry and exit frequency and residence time data are obtained in real time to determine whether the target personnel activity threshold is exceeded. If the target personnel activity threshold is exceeded, the access control authorization time period and number of times are dynamically adjusted to control personnel entry and exit activities, including: Receive the card swiping record collected by the access control card reader and the location coordinates collected by the positioning base station, and obtain the area entry and exit records and the area residence time according to the card swiping record and the location coordinates according to a fixed time period; Based on the entry and exit records and the length of stay in the area, the regional temperature and humidity data are used to calculate the environmental load index. If the environmental load index exceeds the preset threshold range, the regional access control rules are determined based on the rule-based threshold; Obtain the authorization level corresponding to the identity identifier. If the authorization level is high, entry and exit at any time and unlimited stay are allowed. If the authorization level is low, entry and exit during peak hours and continuous stay are restricted. An identity identifier is obtained by applying for a real-time card swiping application, and the authorization level is queried based on the identity identifier. The authorization level and the regional access control rules are used to determine whether the entry and exit restrictions are met. If the entry and exit restrictions are not met, the access control actuator is triggered to lock.

6. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 1, characterized in that: In step S107, based on the control adjustment of the air conditioning and access control in steps S105 and S106, the environmental data of the computer room after the control adjustment is detected, including the temperature and humidity in different areas of the computer room and the activity trajectory and length of stay of personnel in the computer room. The environmental data is formed into a curve graph through a preset data visualization module to obtain the environmental change trend of the computer room after the control, including: Receive temperature and humidity data and personnel location data sent by a collection device, which synchronizes its clock through the network time protocol and records the data at fixed time intervals; Performing fixed time period mean calculation on the temperature and humidity data to obtain an environmental data sequence, wherein the environmental data sequence includes a standard timestamp; Extracting temperature and humidity comparison data from the environmental data sequence, calculating the change in the number of people in the area based on the personnel position data, and obtaining a continuous change curve using linear interpolation; Generate a line graph for the continuous change curve, and mark the corresponding data segment if the temperature and humidity values ​​exceed the target threshold; A regional density distribution map is obtained through grid division. The regional density distribution map maps color depth according to the length of time people stay, and the average temperature and humidity values ​​of the region are superimposed on the regional density distribution map.

7. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 1, characterized in that: In step S108, the current control strategy and method are evaluated based on the environmental change trend of the computer room after control in step S107, and the environmental change trend of the computer room after control is fitted with the preset target environmental change trend by pre-setting a target environmental change trend curve for use in the control of the computer room environment, including: Receive the environmental parameter data sequence collected by the temperature and humidity sensor, generate an actual environmental curve according to the environmental parameter data sequence, and calculate the temperature deviation value and the humidity deviation value by comparing the actual environmental curve with the target reference curve.

8. The method for intelligently controlling a computer room environment by integrating multi-source sensor information according to claim 7, characterized in that: During the monitoring period when the temperature deviation value and the humidity deviation value exceed the preset deviation threshold, a load forecaster is used to perform a feedforward operation, and a control target value and an advance control time are determined according to the feedforward operation result; generating an air conditioning air supply parameter control instruction sequence according to the control target value, adjusting the supply air temperature and supply air humidity by the control instruction sequence, and collecting environmental parameter change data during the gear adjustment process; The temperature change rate and humidity change rate are extracted from the environmental parameter change data. If the temperature change rate and humidity change rate exceed the preset threshold, access control is performed on the corresponding area until the environmental parameters return to the target reference curve range.

Images