Temperature Measurement Data Correction Method and System Based on Intelligent Temperature Sensor

Abstract

This invention relates to the field of industrial big data technology, specifically to a method and system for correcting temperature measurement data based on intelligent temperature sensors, solving the technical problem of low accuracy in existing technologies. The method includes: acquiring monitoring data from multiple temperature sensors distributed within a monitoring area; performing spatial thermodynamic characteristic analysis on the multiple temperature sensors based on the monitoring data to determine the covariance adjustment coefficient for each temperature sensor; the covariance adjustment coefficient characterizes the influence of thermal disturbances at the temperature sensor's location on the process noise covariance matrix; for each temperature sensor, adjusting the corresponding process noise covariance matrix according to the temperature sensor's covariance adjustment coefficient, and performing Kalman filtering on the temperature measurement data of the temperature sensor based on the adjusted process noise covariance matrix to obtain corrected temperature measurement data.

Description

Technical Field

[0001] This invention relates to the field of industrial big data technology, specifically to a method and system for correcting temperature measurement data based on intelligent temperature sensors. Background Technology

[0002] With the rapid development of high-density electronic equipment environments such as data centers, temperature monitoring has become a crucial link in ensuring stable equipment operation and optimizing energy efficiency. Inside data centers, complex, non-uniform, and rapidly changing thermal disturbances exist due to factors such as sudden changes in server load and air conditioning airflow organization, placing higher demands on the accuracy and real-time performance of temperature monitoring. Currently, data centers commonly employ distributed temperature sensor networks for real-time monitoring, using Kalman filtering algorithms to denoise and correct the raw temperature measurement data.

[0003] However, existing temperature monitoring systems typically employ fixed-parameter Kalman filtering, which uses a fixed noise covariance matrix to describe the uncertainty of the system model. This static parameter setting cannot adapt to the dynamic and anisotropic propagation characteristics of thermal disturbances within data centers. In areas of thermal disturbance, fixed parameters can lead to lag or overshoot in the filter estimation, failing to accurately track actual temperature changes. In stable areas, fixed parameters may cause excessive sensitivity to measurement noise, affecting the smoothness and reliability of temperature estimation. This results in potential lag or overshoot in the filter estimation, failing to accurately track actual temperature changes, thus impacting the overall accuracy of temperature estimation and reducing the performance of the thermal management system. Summary of the Invention

[0004] To address the problem of low accuracy in existing technologies, the present invention aims to provide a method and system for correcting temperature measurement data based on an intelligent temperature sensor. The specific technical solution adopted is as follows:

[0005] This application provides a method for correcting temperature measurement data based on a smart temperature sensor, including:

[0006] Acquire monitoring data from multiple temperature sensors distributed within the monitoring area; the monitoring data includes temperature measurement data, spatial location coordinates, and airflow direction vector of the corresponding temperature sensor.

[0007] Based on the monitoring data, spatial thermodynamic characteristic analysis is performed on the multiple temperature sensors to determine the covariance adjustment coefficient of each temperature sensor; the covariance adjustment coefficient is used to characterize the degree of influence of thermal disturbance at the location of the temperature sensor on the process noise covariance matrix;

[0008] For each temperature sensor, the process noise covariance matrix corresponding to the temperature sensor is adjusted according to the covariance adjustment coefficient of the temperature sensor, and the temperature measurement data of the temperature sensor is processed by Kalman filtering based on the adjusted process noise covariance matrix to obtain the corrected temperature measurement data.

[0009] This application provides a temperature measurement data correction system based on an intelligent temperature sensor, comprising:

[0010] The data acquisition unit is used to acquire monitoring data from multiple temperature sensors distributed within the monitoring area; the monitoring data includes temperature measurement data, spatial location coordinates, and airflow direction vector of the corresponding temperature sensor.

[0011] The feature analysis unit is used to perform spatial thermodynamic feature analysis on the multiple temperature sensors based on the monitoring data, and determine the covariance adjustment coefficient of each temperature sensor; the covariance adjustment coefficient is used to characterize the degree of influence of thermal disturbance at the location of the temperature sensor on the process noise covariance matrix;

[0012] The calibration unit is used to adjust the process noise covariance matrix corresponding to each temperature sensor according to the covariance adjustment coefficient of the temperature sensor, and perform Kalman filtering on the temperature measurement data of the temperature sensor based on the adjusted process noise covariance matrix to obtain calibrated temperature measurement data.

[0013] The present invention has the following beneficial effects:

[0014] In view of the technical problem of low accuracy in existing technologies, this application provides a method and system for temperature measurement data correction based on intelligent temperature sensors. By acquiring multi-dimensional monitoring data from multiple temperature sensors distributed within a monitoring area, and combining this with spatial thermodynamic feature analysis to dynamically generate a covariance adjustment coefficient, the process noise covariance matrix is ​​adjusted. Based on the adjusted process noise covariance matrix, the temperature measurement data from the temperature sensors is processed by Kalman filtering to obtain corrected temperature measurement data, achieving adaptive correction of temperature data. Compared to existing technologies that use fixed parameters for data filtering, this application can flexibly adjust the filtering strategy according to the real-time thermal disturbance state of the monitoring area. It can quickly respond to temperature changes in areas of strong disturbance and effectively suppress noise in stable areas, thereby comprehensively improving the accuracy and adaptability of temperature measurement data correction and providing more reliable temperature data support for thermal management in data centers. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a temperature measurement data correction method based on an intelligent temperature sensor, provided in one embodiment of the present invention.

[0017] Figure 2 This is a system architecture diagram of a temperature measurement data correction system based on an intelligent temperature sensor, provided as an embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the temperature measurement data correction method and system based on the intelligent temperature sensor proposed in this invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] In all division and logarithmic operations covered in this application, a smoothing mechanism is employed to prevent computer program crashes or invalid values ​​from being generated due to a zero denominator or a zero input. Specifically, a positive correction factor is superimposed on the denominator term of the division operation or the argument term of the logarithmic function. For example, the value is This ensures the robustness and feasibility of the algorithm under extreme conditions.

[0021] The normalization function mentioned in this application Unless otherwise specified, all values ​​are normalized using maximum and minimum values. The maximum and minimum values ​​are preset empirical extreme values ​​derived from a large amount of historical experimental data. If the calculated result exceeds the [0,1] interval, it is restricted to the [0,1] range by a truncation function (i.e., if the result is less than 0, it is taken as 0, and if it is greater than 1, it is taken as 1) to eliminate the influence of outliers on the evaluation index.

[0022] In view of the technical problem of low accuracy in existing technologies, this application provides a method and system for temperature measurement data correction based on intelligent temperature sensors. By acquiring multi-dimensional monitoring data from multiple temperature sensors distributed within a monitoring area, and combining this with spatial thermodynamic feature analysis to dynamically generate a covariance adjustment coefficient, the process noise covariance matrix is ​​adjusted. Based on the adjusted process noise covariance matrix, the temperature measurement data from the temperature sensors is processed by Kalman filtering to obtain corrected temperature measurement data, achieving adaptive correction of temperature data. Compared to existing technologies that use fixed parameters for data filtering, this application can flexibly adjust the filtering strategy according to the real-time thermal disturbance state of the monitoring area. It can quickly respond to temperature changes in areas of strong disturbance and effectively suppress noise in stable areas, thereby comprehensively improving the accuracy and adaptability of temperature measurement data correction and providing more reliable temperature data support for thermal management in data centers.

[0023] The specific scheme of the temperature measurement data correction method and system based on the intelligent temperature sensor provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0024] Please see Figure 1 The diagram illustrates a flowchart of a temperature measurement data correction method based on a smart temperature sensor according to an embodiment of the present invention. The method includes the following steps:

[0025] Step 101: Obtain monitoring data from multiple temperature sensors distributed within the monitoring area.

[0026] The monitoring data includes temperature measurement data from the corresponding temperature sensor, spatial location coordinates, and airflow direction vector.

[0027] In some embodiments, the monitoring area is an area such as a data center that requires temperature monitoring. Temperature sensors are deployed in a distributed manner, positioned near the air inlets (cold aisle), outlets (hot aisle), and critical equipment of the server racks, forming a three-dimensional monitoring network to ensure comprehensive monitoring. The temperature measurement data is the raw temperature collected in real time by the temperature sensors, which can be represented, for example, by a temperature sequence arranged chronologically. Spatial location coordinates are used to characterize the three-dimensional position of the temperature sensors within the monitoring area, typically represented using a three-dimensional Cartesian coordinate system. The airflow direction vector characterizes the airflow direction at the sensor's location, usually determined based on the cold and hot aisle design of the data center; for example, the cold aisle direction is aligned with the aisle direction, while the hot aisle direction is opposite.

[0028] For example, each temperature sensor collects temperature data in real time at a preset sampling frequency (e.g., 1Hz) and transmits the timestamped raw data synchronously to the central processing unit via a wired or wireless communication network. At the same time, a time synchronization mechanism ensures the consistency of the data sequence of multiple temperature sensors, providing a high-quality data foundation for subsequent analysis.

[0029] Step 102: Based on the monitoring data, perform spatial thermodynamic characteristic analysis on multiple temperature sensors to determine the covariance adjustment coefficient of each temperature sensor.

[0030] The covariance adjustment coefficient is used to characterize the degree of influence of thermal disturbance at the location of the temperature sensor on the process noise covariance matrix.

[0031] Among them, spatial thermal characteristic analysis is based on information such as temperature distribution and airflow direction within the monitoring area to extract features that can reflect the dynamic changes in the local thermal environment, and then quantify the impact of thermal disturbance intensity on filter parameters.

[0032] The covariance adjustment coefficient is a key parameter connecting the thermal environment characteristics and the process noise covariance matrix. Its value is positively correlated with the intensity of thermal disturbance: the stronger the thermal disturbance, the larger the covariance adjustment coefficient, and the greater the adjustment range of the process noise covariance matrix, making the Kalman filter more confident in the measured values; the weaker the thermal disturbance, the closer the covariance adjustment coefficient is to the reference value, and the process noise covariance matrix remains relatively stable, making the Kalman filter more confident in the model predictions.

[0033] In some embodiments, spatial thermal feature analysis can combine multi-dimensional features such as temperature change trends and local thermal equilibrium states, and achieve feature quantification and fusion through a preset algorithm to finally obtain a covariance adjustment coefficient that can accurately reflect the influence of thermal disturbance, thus avoiding judgment bias caused by single feature analysis.

[0034] Step 103: For each temperature sensor, adjust the process noise covariance matrix corresponding to the temperature sensor according to the covariance adjustment coefficient of the temperature sensor, and perform Kalman filtering on the temperature measurement data of the temperature sensor based on the adjusted process noise covariance matrix to obtain the corrected temperature measurement data.

[0035] The process noise covariance matrix is ​​used to describe the uncertainty of the system model in the Kalman filter algorithm. This application dynamically adjusts the initially set baseline process noise covariance matrix using covariance adjustment coefficients, replacing the fixed matrix parameter settings of the traditional Kalman filter algorithm.

[0036] For example, the adjusted process noise covariance matrix satisfies the following formula:

[0037]

[0038] in, For the first The adjusted process noise covariance matrix corresponding to each temperature sensor. For the first Covariance adjustment coefficient of each temperature sensor The reference process noise covariance matrix is ​​denoted as .

[0039] When the temperature sensor is located in an area with strong thermal disturbance. Increase As the temperature increases, the Kalman filter algorithm will place more trust in the temperature measurement data, thus quickly tracking the actual temperature changes; when the thermal disturbance is weak or the regional temperature is stable, Decrease As a result, the Kalman filter algorithm places more trust in model predictions and effectively suppresses measurement noise.

[0040] Thus, this application can be based on the adjusted The Kalman filter prediction and update steps are performed to denoise and correct the original temperature measurement data, and finally output accurate corrected temperature measurement data.

[0041] Based on the above technical solution, this application acquires multi-dimensional monitoring data from multiple temperature sensors distributed within a monitoring area, and dynamically generates a covariance adjustment coefficient by combining spatial thermal characteristic analysis. This adjusts the process noise covariance matrix. Based on the adjusted process noise covariance matrix, Kalman filtering is applied to the temperature measurement data from the temperature sensors to obtain calibrated temperature measurement data, achieving adaptive correction of the temperature data. Compared to existing technologies that use fixed parameters for data filtering, this application can flexibly adjust the filtering strategy according to the real-time thermal disturbance state of the monitoring area. It can quickly respond to temperature changes in areas of strong disturbance and effectively suppress noise in stable areas, thereby comprehensively improving the accuracy and adaptability of temperature measurement data correction and providing more reliable temperature data support for thermal management in data centers.

[0042] As a possible embodiment of this application, step 102 above can be implemented through the following steps:

[0043] Step 201: For each temperature sensor, perform spatial gradient analysis to determine the spatial gradient corresponding to the temperature sensor.

[0044] The spatial gradient is used to characterize the temperature change trend of the corresponding temperature sensor location in the airflow direction.

[0045] This application utilizes spatial gradient analysis combined with the temperature propagation characteristics of airflow direction to uncover the spatial distribution patterns of temperature within a monitoring area. Since airflow direction significantly impacts heat transfer within a data center, the temperature transfer efficiency along the airflow direction is far higher than against or perpendicular to the airflow direction over the same spatial distance. Therefore, the spatial gradient needs to highlight the differences in temperature changes along the airflow direction, providing a basis for determining the intensity and direction of thermal disturbance propagation.

[0046] Step 202: For each temperature sensor, perform a thermal balance deviation analysis to determine the corresponding thermal balance deviation degree of the temperature sensor.

[0047] Among them, thermal balance deviation is used to characterize the degree of deviation of the local thermal balance state at the location of the corresponding temperature sensor.

[0048] It is important to note that the stability of the local thermal environment depends on the balance between heat generation and heat removal. Thermal balance deviation analysis quantifies the combined effects of factors such as heat accumulation, active cooling, and upstream heat input to determine whether the current location is in thermal equilibrium. When heat generation exceeds heat removal, the thermal balance deviation exhibits a specific trend; conversely, it shows a different trend. This characteristic accurately reflects the intrinsic causes of local thermal disturbances. A larger thermal balance deviation indicates that heat generation may exceed heat removal, and the local thermal environment tends to be unstable; a smaller thermal balance deviation indicates that cooling is dominant, and the local thermal environment tends to be stable.

[0049] Step 203: For each temperature sensor, determine the covariance adjustment coefficient of the temperature sensor based on the spatial gradient value and thermal balance deviation of the temperature sensor.

[0050] Among them, the spatial gradient reflects the spatial propagation characteristics of thermal disturbance, and the thermal balance deviation reflects the energy balance characteristics of thermal disturbance. Both characterize the thermal disturbance state from different dimensions. This application can comprehensively cover different types of thermal disturbance scenarios (such as spatially propagating disturbances caused by strong convection and energy imbalance disturbances caused by local heat sources) by fusing these two features, thereby obtaining a covariance adjustment coefficient that is more consistent with the actual thermal environment.

[0051] In one possible implementation, this application can normalize the spatial gradient value and the thermal equilibrium deviation to obtain a normalized spatial gradient and a normalized thermal equilibrium deviation.

[0052] For example, this application can normalize the spatial gradient value and thermal balance deviation of the current measurement based on the spatial gradient value and thermal balance deviation obtained from historical measurements. Normalization can employ a standardized normalization method to eliminate differences in the two characteristic dimensions and numerical ranges, ensuring the fairness of the fusion process.

[0053] For example, the normalized spatial gradient satisfies the following formula:

[0054]

[0055] in, For the first The normalized spatial gradient corresponding to each temperature sensor For the first Spatial gradient values ​​corresponding to each temperature sensor For the first time in historical measurements The average spatial gradient value corresponding to each temperature sensor For the first time in historical measurements The standard deviation of the spatial gradient values ​​corresponding to each temperature sensor.

[0056] The normalized thermal equilibrium deviation satisfies the following formula:

[0057]

[0058] in, For the first Normalized thermal balance deviation corresponding to each temperature sensor For the first Thermal balance deviation corresponding to each temperature sensor For the first time in historical measurements The average thermal balance deviation corresponding to each temperature sensor For the first time in historical measurements The standard deviation of thermal balance deviation corresponding to each temperature sensor.

[0059] Subsequently, the normalized spatial gradient and the normalized thermal equilibrium deviation are weighted and fused to obtain the disturbance intensity index.

[0060] The disturbance intensity index is used to characterize the intensity of thermal disturbance at the location of the temperature sensor.

[0061] First, this application needs to calculate the weights of the spatial gradient feature and the thermal equilibrium deviation feature, ensuring that the sum of the weights is 1. The weight calculation adopts a dynamic allocation mechanism based on instantaneous intensity and historical stability. For example, the weights of the spatial gradient feature satisfy the following formula:

[0062]

[0063] in, For the first The weights of the spatial gradient features corresponding to each temperature sensor. For the first Spatial gradient values ​​corresponding to each temperature sensor For the first time in historical measurements Standard deviation of spatial gradient values ​​corresponding to each temperature sensor For the first Thermal balance deviation corresponding to each temperature sensor For the first time in historical measurements The standard deviation of the thermal balance deviation corresponding to each temperature sensor This is a safety parameter used to correct fractions where the denominator is 0; its specific value can be determined based on... The value of the value determines the outcome, such as . For the first The time stability correction factor for the spatial gradient characteristics corresponding to each temperature sensor satisfies the following formula:

[0064]

[0065] in, For the first Temporal stability correction factor for the spatial gradient characteristics corresponding to each temperature sensor For the first Spatial gradient values ​​corresponding to each temperature sensor For the first time in historical measurements The average spatial gradient value corresponding to each temperature sensor A preset spatial gradient noise threshold (e.g., 0.1) is set to prevent the stability factor calculation from failing when the historical mean is too small.

[0066] The weights of the thermal equilibrium deviation characteristic satisfy the following formula:

[0067]

[0068] in, For the first The weights of the thermal balance deviation characteristics corresponding to each temperature sensor. For the first The weights of the spatial gradient features corresponding to each temperature sensor.

[0069] When the spatial gradient is significant and stable, the spatial gradient feature has a higher weight; when the thermal equilibrium deviation is significant and stable, the thermal equilibrium deviation feature has a higher weight; when both are significant but unstable, the weights tend to be balanced. Therefore, this application can distinguish feature dominance through ratio normalization, avoiding feature dilution caused by equal fusion. Furthermore, a time stability correction factor is used... By weakening features with drastic fluctuations (such as reducing the weight of gradient values ​​that change abruptly) and highlighting stable and reliable features (such as increasing the weight of continuous thermal equilibrium deviations), dynamic adjustment based on confidence level is achieved.

[0070] The disturbance intensity index satisfies the following formula:

[0071]

[0072] in, For the first The disturbance intensity index corresponding to each temperature sensor For the first The weights of the spatial gradient features corresponding to each temperature sensor. For the first Normalized spatial gradient corresponding to each temperature sensor For the first The weights of the thermal balance deviation characteristics corresponding to each temperature sensor. For the first Normalized thermal balance deviation corresponding to each temperature sensor. The larger the value, the stronger the thermal disturbance. This disturbance intensity index integrates spatial propagation characteristics and energy balance characteristics, and comprehensively and accurately quantifies the thermal disturbance state at the current location.

[0073] Thus, this application can perform nonlinear mapping on the reference adjustment coefficient based on the disturbance intensity index to obtain the covariance adjustment coefficient of the temperature sensor.

[0074] For example, the covariance adjustment factor satisfies the following formula:

[0075]

[0076] in, For the first Covariance adjustment coefficients for each temperature sensor Covariance adjustment factor (for example, The value is 1.0, corresponding to the adjustment coefficient benchmark value in the stable region. To adjust the amplitude control parameters (for example, The maximum increase of the control adjustment coefficient is 0.8. For nonlinearity control parameters (exemplary, The sensitivity of the effect of the control disturbance intensity on the adjustment coefficient is 1.5. This is the hyperbolic tangent function, used to nonlinearly map the calculation results to the range of -1 to 1.

[0077] when When it is positive and the absolute value is large (indicating strong thermal disturbance). Approaching 1, Approaching This increases the adjustment coefficient; when When the value is negative and the absolute value is large (indicating strong cooling or abnormal backflow). Approaching -1, Approaching 0, the adjustment coefficient is reduced; when When it approaches 0 (indicating weak thermal disturbance or a stable state). Approaching 0, Approaching By maintaining the baseline value of the adjustment coefficient, the nonlinear mapping ensures a nonlinear relationship between the change of the adjustment coefficient and the intensity of thermal disturbance.

[0078] Based on the above technical solution, this application performs spatial gradient analysis on each temperature sensor to determine its corresponding spatial gradient, and thermal balance deviation analysis to determine its corresponding thermal balance deviation. By capturing the spatial propagation law of thermal disturbance and the energy balance state through these two dimensions of features, the limitations of single-feature analysis are avoided. Thus, for each temperature sensor, determining the covariance adjustment coefficient based on its corresponding spatial gradient value and thermal balance deviation allows for a more comprehensive and accurate reflection of the impact of thermal disturbance on the process noise covariance matrix, further improving the reliability of the covariance adjustment coefficient and laying the foundation for subsequent accurate adjustment of the process noise covariance matrix.

[0079] As a possible embodiment of this application, step 201 above can be implemented through the following steps:

[0080] Step 301: For each temperature sensor, construct a set of neighboring sensors corresponding to the temperature sensor based on the spatial coordinates of multiple temperature sensors and the airflow direction vector.

[0081] Because the deployment of temperature sensors in data centers is non-uniform, and airflow direction has a significant impact on the heat transfer range (for example, a near-end sensor perpendicular to the airflow direction has a much smaller temperature impact on the target sensor than a far-end sensor in the direction of the airflow), this application can combine spatial location and airflow direction as dual factors to screen neighboring sensors, ensuring that the sensors in the neighborhood set have a significant heat transfer correlation with the target sensor.

[0082] In one possible implementation, this application can, for each temperature sensor, use the temperature sensor as the center and, based on the distance between the temperature sensor and other temperature sensors among multiple temperature sensors and the angle of the airflow direction, select neighboring sensors from multiple temperature sensors to construct a set of neighboring sensors corresponding to the temperature sensor.

[0083] Among them, distance screening can set a distance threshold based on the spatial influence range of heat diffusion. For example, according to the rack layout and heat diffusion characteristics of the data center, the distance threshold can be set to 3 meters, that is, only other temperature sensors with an Euclidean distance of less than 3 meters from the temperature sensor are considered as candidate neighboring sensors.

[0084] Airflow direction angle screening can be based on setting an angle threshold according to airflow consistency. For example, the angle threshold can be set to 60 degrees, that is, when the angle between the airflow direction of the temperature sensor and the airflow direction of the candidate neighboring sensor is less than 60 degrees, the candidate neighboring sensor is included as a neighboring sensor in the set of neighboring sensors.

[0085] By using both distance and angle thresholds for filtering, it is possible to ensure that there is a significant thermal transfer correlation between the neighboring sensors in the neighborhood sensor set and the temperature sensor. This avoids sensor interference with negligible thermal effects due to excessive distance, and also eliminates sensor interference with extremely low thermal transfer efficiency due to large differences in airflow direction. This provides accurate correlated sensor data for subsequent spatial gradient calculation.

[0086] Step 302: Determine the spatial gradient corresponding to the temperature sensor based on the temperature measurement data, spatial location coordinates, and airflow direction vector of each neighboring sensor in the temperature sensor and the corresponding neighboring sensor set.

[0087] The neighborhood sensor set provides temperature distribution information around the temperature sensor. Therefore, the temperature measurement data can be weighted by combining the spatial position relationship (distance) and airflow direction relationship (angle) between each neighborhood sensor and the temperature sensor in the neighborhood sensor set. This highlights the influence of key sensors (close to the sensor and with consistent airflow direction) and weakens the interference of irrelevant sensors, ultimately obtaining a spatial gradient that can truly reflect the temperature change trend in the airflow direction.

[0088] In one possible implementation, this application can determine the spatial weighting factor corresponding to each neighboring sensor in the neighboring sensor set based on the distance between the temperature sensor and the neighboring sensor and the angle between the airflow direction.

[0089] For example, the spatial weighting factor satisfies the following formula:

[0090]

[0091] in, For the first In the neighborhood sensor set corresponding to the temperature sensor, the first... Spatial weighting factors corresponding to each neighboring sensor For the first The temperature sensor and its corresponding set of neighboring sensors. The distance between neighboring sensors This is the distance attenuation coefficient, determined based on the thermal diffusion characteristics of the data center; for example, it can be 1.5. For the first The temperature sensor and its corresponding set of neighboring sensors. The angle between the airflow directions of neighboring sensors. This is the natural exponential function, used to normalize the calculation result to a range between 0 and 1. It is a cosine function used to map the included angle to a range between 0 and 1.

[0092] The smaller, The closer the value is to 1, the greater its weight, reflecting the principle that the closer the distance, the stronger the thermal effect. The smaller the value, the more consistent the airflow direction. The closer the value is to 1, the greater the weight, reflecting the principle that heat transfer efficiency is higher when the airflow direction is consistent. This spatial weighting factor can quantify the influence of different neighboring sensors on the target sensor, providing a basis for subsequent weighted processing of temperature data.

[0093] Subsequently, based on the spatial weight factors of each neighboring sensor in the neighborhood sensor set and the airflow direction angle, the temperature measurement data of the temperature sensor and each neighboring sensor in the neighborhood sensor set are weighted to obtain the spatial gradient.

[0094] For example, the spatial gradient satisfies the following formula:

[0095]

[0096] in, For the first Spatial gradient values ​​corresponding to each temperature sensor For the first The set of neighboring sensors corresponding to each temperature sensor For the first In the neighborhood sensor set corresponding to the temperature sensor, the first... Spatial weighting factors corresponding to each neighboring sensor For the first In the neighborhood sensor set corresponding to the temperature sensor, the first... Each neighborhood sensor at time Temperature measurement value, For the first A temperature sensor at time Temperature measurement value, For the first The temperature sensor and its corresponding set of neighboring sensors. The angle between the airflow directions of neighboring sensors.

[0097] Characterizing the temperature difference between the neighboring sensor and the temperature sensor, by The temperature difference is projected onto the airflow direction, and through... Weighted calculations are performed to highlight the influence of key neighborhood sensors. The sign of the resulting spatial gradient value can characterize the propagation trend of thermal disturbance. A positive value indicates that the thermal disturbance tends to propagate outward from the temperature sensor, while a negative value indicates that there is a backflow or abnormal propagation trend. The magnitude of the spatial gradient value represents the propagation intensity of the thermal disturbance.

[0098] Based on the above technical solution, this application can construct a set of neighborhood sensors with heat transfer correlation for each temperature sensor according to the spatial position coordinates of multiple temperature sensors and the airflow direction vector. This provides high-quality input data for spatial gradient calculation, avoids interference from irrelevant sensors on the calculation results, and ensures that the spatial gradient can accurately characterize the temperature change trend in the airflow direction by performing spatial gradient calculation based on multi-dimensional data. This provides reliable feature support for the accurate calculation of the subsequent covariance adjustment coefficient.

[0099] As a possible embodiment of this application, step 202 above can be implemented through the following steps:

[0100] Step 401: For each temperature sensor, construct the upstream sensor set corresponding to the temperature sensor based on the spatial coordinates of multiple temperature sensors and the airflow direction vector.

[0101] Because heat transfer has a significant directionality, especially in the airflow organization environment of data centers, heat is usually transferred from upstream to downstream along the airflow direction. Therefore, this application can screen out upstream sensors that have a thermal input effect on the temperature sensor. The temperature changes of these upstream sensors will affect the local thermal equilibrium state of the target sensor through heat transfer.

[0102] In one possible implementation, this application can, for each temperature sensor, select upstream sensors from multiple temperature sensors based on the distance between the temperature sensor and other temperature sensors among multiple temperature sensors, and the angle between the airflow direction vector of the temperature sensor and the spatial vector pointing from other temperature sensors to the temperature sensor, and construct an upstream sensor set corresponding to the temperature sensor.

[0103] The angle between the airflow direction vector of the temperature sensor and the spatial vector pointing from other temperature sensors to the temperature sensor can be used to determine whether other temperature sensors are upstream of the temperature sensor in terms of heat transfer. For example, when the angle is less than 90 degrees, it indicates that the temperature sensor is upstream of the airflow direction. At the same time, combined with the distance threshold (refer to the distance threshold in the above embodiment), sensors that are close to the temperature sensor and are upstream are selected as upstream sensors.

[0104] By using both distance and directional angle as selection criteria, it can be ensured that the temperature sensor in the upstream sensor set is the source of heat input for that temperature sensor, avoiding interference from downstream sensors or irrelevant directional sensors on the thermal balance analysis, and providing accurate upstream heat input data for subsequent thermal balance deviation calculation.

[0105] Step 402: Determine the thermal balance deviation corresponding to the temperature sensor based on the temperature measurement data of the temperature sensor and the corresponding upstream sensor set.

[0106] Among them, the temperature data of the upstream sensor can reflect the intensity of external heat input. Combined with the temperature data of the temperature sensor itself and the preset thermodynamic parameters, it is possible to comprehensively analyze the combined effects of factors such as heat accumulation, active cooling, and upstream heat input, thereby accurately judging the degree of deviation of the local thermal equilibrium state.

[0107] In one possible implementation, this application can determine the thermal coupling weight factor corresponding to each upstream sensor in the upstream sensor set based on the temperature measurement data of the temperature sensor and the upstream sensor.

[0108] Among them, the thermal coupling weight factor is used to characterize the degree of contribution of the corresponding upstream sensor to the heat transfer of the temperature sensor.

[0109] For example, the thermal coupling weighting factor satisfies the following formula:

[0110]

[0111] in, For the first The upstream sensor set corresponding to the temperature sensor is the first one. Thermal coupling weight factor corresponding to each upstream sensor The basic thermal coupling coefficient can be obtained by fitting historical measurement data or step response tests, and characterizes the first... The temperature sensor and its corresponding upstream sensor set. Static heat transfer efficiency between upstream sensors For the first The upstream sensor set corresponding to the temperature sensor is the first one. An upstream sensor at time Temperature measurement value, For the first The temperature sensor and its corresponding upstream sensor set. The thermal transfer delay between upstream sensors can be determined based on the thermal transfer distance and the properties of the medium. For the first A temperature sensor at time Temperature measurement value, The reference temperature difference is used to normalize the temperature difference value and can be determined based on historical measurement data. This is the function for finding the maximum value.

[0112] Characterizing the first The temperature sensor and its corresponding upstream sensor set. The normalized temperature difference between the upstream sensors, based on Eliminate the effects of heat transfer delay. A positive heat transfer contribution is ensured only when the temperature measurement from the upstream sensor (after delay) is higher than the temperature measurement from this temperature sensor; in this case, the thermal coupling weighting factor is... The product of the normalized temperature difference; if the temperature measurement value of the upstream sensor is lower than or equal to the temperature measurement value of the upstream sensor, the thermal coupling weighting factor is 0, indicating that there is no positive thermal input.

[0113] The larger the value, the greater the contribution of the upstream sensor to heat transfer. It can dynamically quantify the real-time thermal input intensity of upstream sensors, providing accurate upstream thermal contribution data for thermal balance analysis.

[0114] Subsequently, the thermal balance deviation of the temperature sensor is determined based on the thermal coupling weight factor, temperature measurement data, and temperature measurement data of each upstream sensor in the upstream sensor set.

[0115] For example, the thermal equilibrium deviation satisfies the following formula:

[0116]

[0117] in, For the first Thermal balance deviation corresponding to each temperature sensor For the first The smoothed value of the temperature change rate corresponding to each temperature sensor is used to characterize the trend of heat accumulation. For the first The local cooling coefficient corresponding to each temperature sensor can be obtained by fitting historical data, serving as a quantitative indicator characterizing the active cooling effect. For the first A temperature sensor at time Temperature measurement value, For reference temperature, the air conditioner set temperature or the average temperature of the area is usually taken. For the first The set of upstream sensors corresponding to each temperature sensor For the first The upstream sensor set corresponding to the temperature sensor is the first one. Thermal coupling weight factor corresponding to each upstream sensor For the first The upstream sensor set corresponding to the temperature sensor is the first one. An upstream sensor at time Temperature measurement value, For the first The temperature sensor and its corresponding upstream sensor set. Thermal transfer delay between upstream sensors For the first A temperature sensor at time Temperature measurement value.

[0118] For example, smoothed value of temperature change rate Differences can be smoothed using a sliding window:

[0119]

[0120] in, For the first Smoothed values ​​of temperature change rate corresponding to each temperature sensor It is a mean function. Indicates the first A temperature sensor at time to The average of all temperature measurements, Indicates the first A temperature sensor at time to The average of all temperature measurements, The sampling interval can be determined by the sampling frequency of the temperature sensor.

[0121] In other words, the deviation from thermal equilibrium consists of three parts, the first part being... Characterizing the heat accumulation trend, the influence of instantaneous noise on the rate of change estimation is suppressed through sliding window smoothing; Part Two Characterizing the active cooling effect, this part increases when the temperature measured by the temperature sensor is higher than the reference temperature, indicating increased cooling demand; Part Three The comprehensive impact of upstream heat input is characterized by a weighted summation using dynamic thermal coupling weighting factors, quantifying the total positive heat input from all upstream sensors to this temperature sensor. The combination of these three parts achieves a comprehensive quantification of three key factors: heat accumulation, active cooling, and upstream heat input.

[0122] The positive or negative sign reflects the deviation trend of the local thermal equilibrium state. This indicates that the heat generation trend may exceed the heat removal trend, and the local thermal environment tends to be unstable; A negative value indicates that the cooling effect is dominant and the local thermal environment tends to be stable; A value close to 0 indicates that the temperature is close to thermal equilibrium. The magnitude of the value can accurately quantify the degree of deviation of the energy balance in the local thermal environment, providing a reliable energy characteristic basis for the subsequent calculation of the covariance adjustment coefficient.

[0123] Based on the above technical solution, this application constructs an upstream sensor set to accurately locate the external heat input sources affecting the thermal balance state of the target sensor, avoiding interference from irrelevant sensor data on thermal balance analysis. Based on the temperature data of the upstream sensor and the temperature sensor, the thermal balance deviation is calculated, ensuring that the thermal balance deviation can accurately reflect the energy balance state of the local thermal environment, and providing a reliable energy characteristic basis for the comprehensive calculation of the covariance adjustment coefficient.

[0124] Please see Figure 2 The diagram illustrates a system architecture of a temperature data correction system 20 based on a smart temperature sensor, according to an embodiment of the present invention. The temperature data correction system 20 based on a smart temperature sensor includes:

[0125] The data acquisition unit 21 is used to acquire monitoring data from multiple temperature sensors distributed within the monitoring area; the monitoring data includes temperature measurement data, spatial coordinates, and airflow direction vectors of the corresponding temperature sensors.

[0126] The feature analysis unit 22 is used to perform spatial thermodynamic feature analysis on multiple temperature sensors based on monitoring data, and to determine the covariance adjustment coefficient of each temperature sensor. The covariance adjustment coefficient is used to characterize the degree of influence of thermal disturbance at the location of the temperature sensor on the process noise covariance matrix.

[0127] The calibration unit 23 is used to adjust the process noise covariance matrix corresponding to each temperature sensor according to the covariance adjustment coefficient of the temperature sensor, and perform Kalman filtering on the temperature measurement data of the temperature sensor based on the adjusted process noise covariance matrix to obtain the calibrated temperature measurement data.

[0128] It should be noted that the various embodiments of this application can be referenced or learned from each other. For example, the same or similar steps, method embodiments, system embodiments and device embodiments can be referenced from each other without limitation.

[0129] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0130] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A temperature measurement data correction method based on an intelligent temperature sensor, characterized in that, include: Acquire monitoring data from multiple temperature sensors distributed within the monitoring area; the monitoring data includes temperature measurement data, spatial location coordinates, and airflow direction vector of the corresponding temperature sensor. Based on the monitoring data, spatial thermodynamic characteristic analysis is performed on the multiple temperature sensors to determine the covariance adjustment coefficient of each temperature sensor; The covariance adjustment coefficient is used to characterize the degree of influence of thermal disturbance at the location of the temperature sensor on the process noise covariance matrix; For each temperature sensor, the process noise covariance matrix corresponding to the temperature sensor is adjusted according to the covariance adjustment coefficient of the temperature sensor, and the temperature measurement data of the temperature sensor is processed by Kalman filtering based on the adjusted process noise covariance matrix to obtain the corrected temperature measurement data. Based on the monitoring data, spatial thermodynamic characteristic analysis is performed on the multiple temperature sensors to determine the covariance adjustment coefficient for each temperature sensor, including: For each temperature sensor, spatial gradient analysis is performed to determine the spatial gradient corresponding to the temperature sensor; the spatial gradient is used to characterize the temperature change trend of the corresponding temperature sensor location in the airflow direction. For each temperature sensor, a thermal balance deviation analysis is performed to determine the corresponding thermal balance deviation degree; the thermal balance deviation degree is used to characterize the degree of deviation of the local thermal balance state at the location of the corresponding temperature sensor. For each temperature sensor, the covariance adjustment coefficient of the temperature sensor is determined based on the spatial gradient value and thermal balance deviation corresponding to the temperature sensor; Specifically, for each temperature sensor, a thermal balance deviation analysis is performed to determine the corresponding thermal balance deviation degree, including: For each temperature sensor, an upstream sensor set corresponding to the temperature sensor is constructed based on the spatial position coordinates of the multiple temperature sensors and the airflow direction vector; Determining the thermal balance deviation corresponding to the temperature sensor based on the temperature measurement data of the temperature sensor and each upstream sensor in the corresponding upstream sensor set includes: For each upstream sensor in the upstream sensor set, a thermal coupling weighting factor corresponding to the upstream sensor is determined based on the temperature measurement data of the temperature sensor and the upstream sensor; the thermal coupling weighting factor is used to characterize the degree of heat transfer contribution of the corresponding upstream sensor to the temperature sensor. The thermal balance deviation of the temperature sensor is determined based on the thermal coupling weight factor, temperature measurement data, and temperature measurement data of each upstream sensor in the upstream sensor set. 2.The smart temperature sensor based temperature measurement data correction method of claim 1, wherein, For each temperature sensor, spatial gradient analysis is performed to determine the spatial gradient corresponding to the temperature sensor, including: For each temperature sensor, a neighborhood sensor set corresponding to the temperature sensor is constructed based on the spatial position coordinates of the multiple temperature sensors and the airflow direction vector; The spatial gradient corresponding to the temperature sensor is determined based on the temperature measurement data, spatial coordinates, and airflow direction vector of each neighboring sensor in the temperature sensor and its corresponding neighboring sensor set. 3.The smart temperature sensor based temperature measurement data correction method of claim 2, wherein, For each temperature sensor, a neighborhood sensor set corresponding to the temperature sensor is constructed based on the spatial coordinates of the multiple temperature sensors and the airflow direction vector, including: For each temperature sensor, taking the temperature sensor as the center, based on the distance between the temperature sensor and other temperature sensors among the plurality of temperature sensors and the angle of airflow direction, neighboring sensors are selected from the plurality of temperature sensors to construct a set of neighboring sensors corresponding to the temperature sensor. 4.The intelligent temperature sensor based temperature measurement data correction method of claim 2, wherein, The spatial gradient corresponding to the temperature sensor is determined based on the temperature measurement data, spatial coordinates, and airflow direction vector of the temperature sensor and its corresponding neighboring sensor set, including: For each neighboring sensor in the set of neighboring sensors, a spatial weighting factor corresponding to the neighboring sensor is determined based on the distance between the temperature sensor and the neighboring sensor and the angle between the airflow direction. Based on the spatial weight factors of each neighboring sensor in the neighborhood sensor set and the airflow direction angle, the temperature measurement data of the temperature sensor and each neighboring sensor in the neighborhood sensor set are weighted to obtain the spatial gradient. 5.The smart temperature sensor based temperature measurement data correction method of claim 1, wherein, For each temperature sensor, an upstream sensor set corresponding to the temperature sensor is constructed based on the spatial coordinates of the multiple temperature sensors and the airflow direction vector, including: For each temperature sensor, based on the distance between the temperature sensor and other temperature sensors among the plurality of temperature sensors, and the angle between the airflow direction vector of the temperature sensor and the spatial vector pointing from other temperature sensors to the temperature sensor, upstream sensors are selected from the plurality of temperature sensors to construct the upstream sensor set corresponding to the temperature sensor. 6.The smart temperature sensor based temperature measurement data correction method of claim 1, wherein, For each temperature sensor, the covariance adjustment coefficient of the temperature sensor is determined based on the spatial gradient value and thermal equilibrium deviation corresponding to the temperature sensor, including: The spatial gradient value and the thermal equilibrium deviation are normalized to obtain the normalized spatial gradient and the normalized thermal equilibrium deviation. The normalized spatial gradient and the normalized thermal equilibrium deviation are weighted and fused to obtain a disturbance intensity index; the disturbance intensity index is used to characterize the thermal disturbance intensity at the location of the temperature sensor. The covariance adjustment coefficient of the temperature sensor is obtained by performing a nonlinear mapping on the benchmark adjustment coefficient based on the disturbance intensity index.

7. A temperature measurement data correction system based on intelligent temperature sensors, characterized by, include: The data acquisition unit is used to acquire monitoring data from multiple temperature sensors distributed within the monitoring area; the monitoring data includes temperature measurement data, spatial location coordinates, and airflow direction vector of the corresponding temperature sensor. The feature analysis unit is used to perform spatial thermodynamic feature analysis on the multiple temperature sensors based on the monitoring data, and determine the covariance adjustment coefficient of each temperature sensor. The covariance adjustment coefficient is used to characterize the degree of influence of thermal disturbance at the location of the temperature sensor on the process noise covariance matrix; The calibration unit is used to adjust the process noise covariance matrix corresponding to each temperature sensor according to the covariance adjustment coefficient of the temperature sensor, and perform Kalman filtering on the temperature measurement data of the temperature sensor based on the adjusted process noise covariance matrix to obtain calibrated temperature measurement data. Based on the monitoring data, spatial thermodynamic characteristic analysis is performed on the multiple temperature sensors to determine the covariance adjustment coefficient for each temperature sensor, including: For each temperature sensor, spatial gradient analysis is performed to determine the spatial gradient corresponding to the temperature sensor; the spatial gradient is used to characterize the temperature change trend of the corresponding temperature sensor location in the airflow direction. For each temperature sensor, a thermal balance deviation analysis is performed to determine the corresponding thermal balance deviation degree; the thermal balance deviation degree is used to characterize the degree of deviation of the local thermal balance state at the location of the corresponding temperature sensor. For each temperature sensor, the covariance adjustment coefficient of the temperature sensor is determined based on the spatial gradient value and thermal balance deviation corresponding to the temperature sensor; Specifically, for each temperature sensor, a thermal balance deviation analysis is performed to determine the corresponding thermal balance deviation degree, including: For each temperature sensor, an upstream sensor set corresponding to the temperature sensor is constructed based on the spatial position coordinates of the multiple temperature sensors and the airflow direction vector; Determining the thermal balance deviation corresponding to the temperature sensor based on the temperature measurement data of the temperature sensor and each upstream sensor in the corresponding upstream sensor set includes: For each upstream sensor in the upstream sensor set, a thermal coupling weighting factor corresponding to the upstream sensor is determined based on the temperature measurement data of the temperature sensor and the upstream sensor; the thermal coupling weighting factor is used to characterize the degree of heat transfer contribution of the corresponding upstream sensor to the temperature sensor. The thermal balance deviation of the temperature sensor is determined based on the thermal coupling weight factor, temperature measurement data, and temperature measurement data of each upstream sensor in the upstream sensor set.

Images