CN116954342B - Method for monitoring operation of power supply of omnibearing efficient heat dissipation host based on Internet of things - Google Patents
Method for monitoring operation of power supply of omnibearing efficient heat dissipation host based on Internet of things Download PDFInfo
- Publication number
- CN116954342B CN116954342B CN202311198369.3A CN202311198369A CN116954342B CN 116954342 B CN116954342 B CN 116954342B CN 202311198369 A CN202311198369 A CN 202311198369A CN 116954342 B CN116954342 B CN 116954342B
- Authority
- CN
- China
- Prior art keywords
- data
- dimension data
- relevant
- dimension
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 93
- 238000012544 monitoring process Methods 0.000 title claims abstract description 32
- 230000017525 heat dissipation Effects 0.000 title claims abstract description 19
- 238000001514 detection method Methods 0.000 claims abstract description 6
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 4
- 230000000875 corresponding effect Effects 0.000 description 16
- 230000000694 effects Effects 0.000 description 6
- 230000003044 adaptive effect Effects 0.000 description 5
- 238000003064 k means clustering Methods 0.000 description 5
- 230000005856 abnormality Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000002596 correlated effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 208000033999 Device damage Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- KJONHKAYOJNZEC-UHFFFAOYSA-N nitrazepam Chemical compound C12=CC([N+](=O)[O-])=CC=C2NC(=O)CN=C1C1=CC=CC=C1 KJONHKAYOJNZEC-UHFFFAOYSA-N 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/28—Supervision thereof, e.g. detecting power-supply failure by out of limits supervision
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3058—Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3065—Monitoring arrangements determined by the means or processing involved in reporting the monitored data
- G06F11/3072—Monitoring arrangements determined by the means or processing involved in reporting the monitored data where the reporting involves data filtering, e.g. pattern matching, time or event triggered, adaptive or policy-based reporting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Probability & Statistics with Applications (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to an omnidirectional efficient heat dissipation host power supply operation monitoring method based on the Internet of things, which comprises the following steps: collecting multidimensional data when a host power supply operates; acquiring relevant dimension data of current dimension data and a non-uniformity degree value of the relevant dimension data; obtaining an influence weight value of the relevant dimension data according to the non-uniformity degree value of the relevant dimension data of the current dimension data; according to the change of the data point distribution characteristics in the continuous incremental clustering process, and combining the influence weight value of the related dimension data to acquire the self-adaptive membership weight value of the boundary data point; clustering is carried out according to the self-adaptive membership weight value of the boundary data point, and an accurate detection result is obtained. The invention ensures the accuracy of the result influencing the change of the boundary data point under the condition of continuously adding data and continuously updating data, so that the obtained monitoring model is more accurate, and the accuracy of the monitoring result of the running state of the host power supply is ensured.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to an omnidirectional efficient heat dissipation host power supply operation monitoring method based on the Internet of things.
Background
With the continuous development of technology, the internet of things technology is widely applied in various fields, one of which is application in electronic equipment management. Host power supplies are a core component of electronic devices and play a vital role in computers, servers and other high performance devices. However, overheating problems of the host power supply continue to exist and poor heat dissipation management can lead to reduced device performance, reduced lifetime, and even device damage. Therefore, it becomes important to study how to realize the operation monitoring method of the omnibearing efficient heat dissipation host power supply.
The monitoring data in the operation process of the host power supply is monitored in real time by arranging a plurality of sensors on the host power supply, the operation state of the host power supply is reflected in real time by analyzing and processing the data, potential fault signs of the host power supply can be identified, and quick early warning is carried out. But during the data analysis process, abnormal data during operation needs to be monitored. The rough K-Means is an improvement on the traditional K-Means method, and can improve defects when data points of cross boundaries among clusters are processed, but in the operation monitoring process of a host power supply, data of multiple dimensions are required to be collected, the data of each dimension is unevenly distributed, and the cross boundary data points among the clusters are continuously changed in the continuous updating process of the data, so that the membership weight value of the boundary data points in the updating process is unreasonably calculated, the clustering result is inaccurate, and the operation state monitoring result of the host power supply is affected.
Disclosure of Invention
The invention provides an omnidirectional efficient heat dissipation host power supply operation monitoring method based on the Internet of things, which aims to solve the existing problems.
The invention discloses an omnidirectional efficient heat dissipation host power supply operation monitoring method based on the Internet of things, which adopts the following technical scheme:
the embodiment of the invention provides an omnidirectional efficient heat dissipation host power supply operation monitoring method based on the Internet of things, which comprises the following steps of:
acquiring multi-dimensional data of a host power supply during operation, and taking any one dimensional data as current dimensional data;
acquiring relevant dimension data of current dimension data, and recording the relevant dimension data as relevant dimension data; acquiring data space distribution characteristics of related dimension data; acquiring dimension distribution influence weight values of related dimension data; acquiring direct influence values of the related dimension data under the condition that other related dimension data are not changed; according to the dimension distribution influence weight value of the related dimension data and the direct influence value of the related dimension data under the condition that other related dimension data are not changed, the related dimension data and the current dimension data change distribution characteristics; obtaining the non-uniformity degree value of the relevant dimension data according to the data space distribution characteristics of the relevant dimension data and the change distribution characteristics of the relevant dimension data and the current dimension data;
Obtaining an allowable threshold fluctuation value of the relevant dimension data according to the non-uniformity degree value of the relevant dimension data; similarity in data space distribution between data points of the current dimension data and data points of the relevant dimension data according to the allowable threshold fluctuation value of the relevant dimension data; obtaining an influence weight value of the relevant dimension data according to the similarity of the data points of the current dimension data and the data points of the relevant dimension data in the data space distribution;
according to the influence weight value of the related dimension data, the self-adaptive membership weight value of the boundary data point of the current dimension data to the cluster;
clustering is carried out on the clustered self-adaptive membership weight values according to the boundary data points of the current dimension data, and a detection result is obtained.
Preferably, the acquiring the relevant dimension data of the current dimension data includes the following specific steps:
calculating pearson correlation coefficients of the current dimension data and other dimension data; if there is dimension data in other dimension dataThe pearson correlation coefficient with the current dimension data is greater than a preset threshold +.>Dimension data is->And similarly, acquiring all relevant dimension data of the current dimension data as relevant dimension data of the current dimension data.
Preferably, the acquiring the data space distribution characteristics of the related dimension data includes the following specific steps:
preset parametersFor->Related dimension data, in->Convex hull in data-level space of individual related dimension dataDetecting to obtain a convex hull range with the largest data in the space, obtaining a square area corresponding to the smallest external square of the convex hull range, and performing +_ on the square area>Dividing equally to obtain->The method comprises the steps of obtaining the ratio of the average distance of the data points in each block to the side length of the block, wherein the obtaining method of the average distance of the data points in the block comprises the following steps: the distances between all the data points in the block are obtained, and the distances are summed to obtain the average value of the distances as the average distance of the data points in the block;
then the firstData space distribution characteristics of the individual related dimension data itself>The calculation method of (1) is as follows:
；
in the method, in the process of the invention,indicate->Data space distribution characteristics of the relevant dimension data; />Indicate->The data plane of the data of the relevant dimension data is +.>The ratio between the average distance of the data points in each block and the side length of the block；/>Indicate->The average of the ratio between the average distance of data points in all tiles and the side length of the tile in the data plane space of the respective dimension data.
Preferably, the step of acquiring the dimension distribution influence weight value of the relevant dimension data includes the following specific steps:
for the firstRelated dimension data, except +.>The length of the period of time during which the other relevant dimension data of the relevant dimension data will not change, then +.>Dimension distribution influence weight value +.>Wherein->Representing the total time period length of the multi-dimensional data when the host power supply is in operation, < >>Indicating except->The length of the time period occupied when other dimension data of the relevant dimension data is unchanged.
Preferably, the acquiring the direct influence value of the relevant dimension data under the condition that other relevant dimension data are unchanged includes the following specific steps:
for the firstThe relevant dimension data is->Direct influence value of the respective relevant dimension data without change of the other relevant dimension data +.>The acquisition method comprises the following steps: calculating the data change curve and the +.f of the current dimension data in the time range of the other related dimension data unchanged>Pearson correlation coefficients between data change curves of the respective correlation dimension data are characterized by calculating inverse proportion function values of accumulated difference values between the pearson correlation coefficients >The accumulated difference value refers to the absolute value of the difference value of the pearson correlation coefficients of adjacent time periods calculated according to the sequence of the time periods, and then the accumulated values are accumulated in a plurality of sections.
Preferably, the step of influencing the weight value according to the dimension distribution of the relevant dimension data and directly influencing the value of the relevant dimension data under the condition that other relevant dimension data are not changed includes the following specific steps:
for the firstRelated dimension data, th->The method for calculating the change distribution characteristics of the relevant dimension data and the current dimension data comprises the following steps:
；
in the method, in the process of the invention,indicate->The change distribution characteristics of the relevant dimension data and the current dimension data; />Indicate->Direct influence values of the relevant dimension data under the condition that other relevant dimension data are unchanged; />Indicate->The dimension distribution of the relevant dimension data affects the weight value; />Mapping function representing pearson correlation coefficient, +.>Indicate->Pearson correlation coefficient values for the respective correlation dimension data and the current dimension data; />Indicate->Direct influence values of the relevant dimension data under the change of other relevant dimension data; />Indicating except->The total number of relevant dimension data of the current dimension data of the relevant dimension data.
Preferably, the step of obtaining the non-uniformity degree value of the relevant dimension data according to the data space distribution feature of the relevant dimension data and the variation distribution feature of the relevant dimension data and the current dimension data comprises the following specific steps:
for the firstRelated dimension data according to->Data space distribution characteristics of the individual related dimension data itself and +.>The variation distribution characteristics of the relevant dimension data and the current dimension data are obtained>The non-uniformity degree value of each relevant dimension data is calculated by the following expression:
；
in the method, in the process of the invention,indicate->A degree of non-uniformity value for each of the associated dimensional data; />Indicate->Data space distribution characteristics of the relevant dimension data; />Indicate->Individual related dimension data and current dimension numberAccording to the change distribution characteristics; />An exponential function representing a base of a natural constant is represented.
Preferably, the obtaining the allowable threshold fluctuation value of the relevant dimension data according to the non-uniformity degree value of the relevant dimension data includes the following specific steps:
for the firstRelated dimension data, will be->Degree of non-uniformity value of individual related dimension data +.>Performing linear normalization to obtain normalized +.>Degree of non-uniformity value of individual related dimension data +. >The method comprises the steps of carrying out a first treatment on the surface of the Then 1 minus normalized +.>Degree of non-uniformity value of individual related dimension data +.>And a preset threshold->Is the product of->The allowable threshold fluctuation value of each relevant dimension data.
Preferably, the similarity in data space distribution between the data points of the current dimension data and the data points of the relevant dimension data according to the allowable threshold fluctuation value of the relevant dimension data comprises the following specific steps:
for the firstRelated dimension data, the data point and the +.th of the current dimension data are obtained according to the allowable threshold fluctuation value of the related dimension data and the data point of the local range of the boundary data point of the current dimension data and the related dimension data>Similarity in data spatial distribution between data points of the respective related dimension data is calculated as:
；
in the method, in the process of the invention,data points representing current dimension data and +.>Similarity in data spatial distribution between data points of the individual related dimension data; />A number of data points representing a local range of boundary data points of the current dimension data;first +.>Data distribution characteristics within a neighborhood of data points; / >Indicate->The first +.>Data distribution characteristics within a neighborhood of data points; />Indicate->Allowable threshold fluctuation values of the respective related dimension data; />An exponential function representing a base of a natural constant is represented.
Preferably, the adaptive membership weight value of the boundary data point of the current dimension data to the cluster according to the influence weight value of the related dimension data comprises the following specific steps:
for the current dimension dataThe boundary data point is the +.>The boundary data point is about>Adaptive membership weight value for individual clusters +.>Is calculated by the following steps:
；
in the method, in the process of the invention,represents the +.>The boundary data point is about>Self-adaptive membership weight values of the clusters; />The +.f representing the current dimension data now>Data points within the local range of the boundary data points to +.>Average Euclidean distance of the cluster centers; />Data points within the local range of all boundary data points representing the last current dimension data to +.>Average Euclidean distance of the cluster centers; />Total number of relevant dimension data representing current dimension data; / >Indicate->Influence weight values of the relevant dimension data; />Indicate->The +.>A number of data points within a corresponding time period range of the plurality of boundary data points; />Indicate->The +.>Of the data points within the corresponding time period of the boundary data points, the +.>The +.>Data distribution characteristics and +.>The average data distribution characteristic of the data points of the clusters is smaller than that of the firstThe number of data points of the allowable threshold fluctuation value of the relevant dimension data;
and similarly, obtaining the self-adaptive membership weight value of each boundary data point of the current dimension data for each cluster.
The technical scheme of the invention has the beneficial effects that: the invention provides a coarse K-Means clustering method of self-adaptive membership weight values. The method comprises the steps of processing a current dimension, acquiring a relevant dimension of the current dimension, acquiring non-uniformity of the relevant dimension, and further calculating an influence weight value of a data point of the current dimension; and combining the change in the local neighborhood of the boundary data point in the continuously updated clustering process to comprehensively acquire the self-adaptive membership weight value of the boundary data point so as to obtain an accurate clustering result. The defect that in the traditional coarse K-Means clustering process, single dimensional data change fails to reflect the operation state monitoring result of the host power supply is avoided, an accurate clustering result is obtained under the influence of multiple dimensions, the accuracy of the result influencing the change of boundary data points under the condition of continuously adding data and continuously updating data is ensured, the obtained monitoring model is more accurate, and the accuracy of the operation state monitoring result of the host power supply is ensured.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of steps of the method for monitoring the operation of the power supply of the omnidirectional efficient heat dissipation host based on the internet of things.
Detailed Description
In order to further explain the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the omnibearing efficient heat dissipation host power operation monitoring method based on the internet of things according to the invention by combining the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, 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 belongs.
The invention provides a specific scheme of an omnidirectional efficient heat dissipation host power supply operation monitoring method based on the Internet of things, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a method for monitoring operation of an omnidirectional efficient heat dissipation host power supply based on the internet of things according to an embodiment of the present invention is shown, and the method includes the following steps:
step S001: and collecting multi-dimensional data of the host power supply during operation.
It should be noted that, in the course of performing coarse K-Means clustering on the collected multidimensional data, the embodiment needs to collect multiple dimensional data and distributes the dimensional data unevenly, and the cross boundary data points between clusters are continuously changed in the continuous updating process of the data, so that the membership weight value of the boundary data points in the updating process is unreasonably calculated. In the analysis process, different dimension data can be affected by the influence of uneven distribution on a clustering result, and in some dimension data, abnormality can exist, but in single dimension data, whether abnormality exists can not be judged, but in other dimension data, data distribution has obvious difference, for example, when voltage data is analyzed, the change of the voltage data is normal, but larger abnormality can occur in temperature data, humidity data and fan data. In addition, as new data are continuously acquired along with the development of time, in the process of continuously updating and clustering (which can be regarded as an incremental clustering mode), the continuous data updating also has an effect on the change of boundary data points, according to the content, the embodiment processes arbitrary dimension data, acquires the related dimension data of the dimension data and acquires the non-uniformity of the related dimension data, and further calculates the influence weight value of the data points of the dimension data; and combining the change in the local neighborhood of the boundary data point in the continuously updated clustering process to comprehensively acquire the self-adaptive membership weight value of the boundary data point so as to obtain an accurate clustering result.
Specifically, a plurality of sensors are arranged to collect data in an operational state during testing of a host power supply, wherein the data includes multi-dimensional data such as voltage dimension data, power dimension data, temperature dimension data, humidity dimension data, fan dimension data, and the like. Any one dimension data is recorded as current dimension data, wherein the acquisition frequency of each dimension data is set to be the same, and the model of the sensor is not set, and the method can be determined according to the specific implementation situation of an implementer.
Step S002: and acquiring relevant dimension data of the current dimension data and a non-uniformity degree value of the relevant dimension data.
For example, when the current dimension data is temperature dimension data, since a temperature change does not change significantly in a short time (i.e., a temperature change has hysteresis) and a temperature change process is more careful in a monitoring process, when analyzing the temperature dimension data, it is necessary to consider dimension data related to the temperature dimension data, and to characterize a change in the temperature dimension data by a change in other dimension data, and therefore it is necessary to analyze dimension data related to the current dimension data. When the self-adaptive membership weight value of the boundary data points among clusters of the current dimension data is analyzed, the change of the current dimension data of the self-adaptive membership weight value needs to be considered, and meanwhile, the data distribution non-uniformity degree of the related dimension data also needs to be considered.
1. And acquiring relevant dimension data of the current dimension data.
Presetting a threshold valueWherein the present embodiment is +.>Examples are described, the present embodiment is not particularly limited, wherein +.>Depending on the particular implementation.
Specifically, the pearson correlation coefficient of the current dimension data and other dimension data is calculated; if the pearson correlation coefficient of the dimension data and the current dimension data in the other dimension data is larger than the preset threshold valueAnd if the dimension data has correlation with the current dimension data, taking the dimension data as the correlated dimension data of the current dimension data, and similarly, acquiring all correlated dimension data of the current dimension data. And recording the relevant dimension data of the current dimension data as relevant dimension data.
So far, the relevant dimension data of the current dimension data is obtained.
2. And acquiring the non-uniformity degree value of the related dimension data.
It should be noted that, the degree of non-uniformity value of the relevant dimension data is related to the data spatial distribution of the relevant dimension data, but the degree of non-uniformity value of the relevant dimension data, which is expected in the embodiment, is the influence of the data change of the current dimension data, that is, the influence of not only the distribution non-uniformity, but also the distribution influence of the data change feature of the current dimension data, and if the distribution of other dimension data and the current dimension data have relevance, the distribution non-uniformity is smaller. The present embodiment thus performs weight influence calculation by combining the distribution of the data change characteristics with the current dimension data with reference to the data space distribution characteristics of the correlation dimension data itself.
(1) Acquisition of the firstData spatial distribution characteristics of each relevant dimension data.
Specifically, at the firstPerforming convex hull detection in the space of the data layer of the relevant dimension data to obtain the convex hull range with the maximum data in the space, obtaining a square area corresponding to the smallest external square of the convex hull range, and performing 25 equal division on the square area to obtain 25 areas, wherein the implementation is described by dividing the square area into 25 areas, and other embodiments are not limited specifically and depend on specific implementation conditions; the ratio of the average distance of the data points in each block to the side length of the block (the distances between all the data points in the block are obtained and summed to obtain the average value as the average distance of the data points) is recorded as the distance ratio, and a distance change curve is constructed according to the block sequence (from left to right and from top to bottom), wherein the abscissa of the curve is the sequence number of the block, the ordinate is the distance ratio corresponding to the block, then the>Data space distribution characteristics of the individual related dimension data itself>The calculation method of (1) is as follows:
；
in the method, in the process of the invention,indicate->Data space distribution characteristics of the relevant dimension data; / >Indicate->The data plane of the data of the relevant dimension data is +.>The ratio between the average distance of the data points in the individual tiles and the side length of the tile;indicate->The average of the ratio between the average distance of data points in all tiles and the side length of the tile in the data plane space of the respective dimension data.
To this end, obtain the firstData spatial distribution characteristics of each relevant dimension data.
(2) Acquisition of the firstThe relevant dimension data and the current dimension data change distribution characteristics.
It should be noted that, in the process of obtaining the influence of the data change distribution feature in this embodiment, the change and the first dimension of the current dimension data need to be consideredThe correlation of the changes in the relevant dimension data, but the changes in the current dimension data may be affected by the various dimension data.
Specifically, obtain the firstDimension distribution influence weight value +.>: in analysis->When the dimension distribution of the relevant dimension data affects, consider +.>The influence of the relevant dimension data is obtained, so that the time period in which other relevant dimension data cannot change is obtained, and the time period is analyzed, so +.>Wherein->Representing the total time period length of the multi-dimensional data when the host power supply is in operation, < > >Indicating except->The length of the time period occupied when other dimension data of the relevant dimension data is unchanged. If->The larger the value, the more ∈>The relevant dimension data has a larger influence on the current dimension data, so the more needs to be considered +.>The direct influence of the related dimension data on the current dimension data (namely, considering the change regularity of the two dimension data in the same time period) is only needed, otherwise +.>The smaller the value, the corresponding more need to consider the +.>The corrected effect of the relevant dimension data on the current dimension data (i.e., corrected by the correlation of the other dimension data to the current dimension data as a weight in this embodiment).
Acquisition of the firstDirect influence value of the respective relevant dimension data without change of the other relevant dimension data +.>: calculating the data change curve and the +.f of the current dimension data in the time range of the other related dimension data unchanged>Pearson correlation coefficients between data change curves of the respective correlation dimension data (there are plural pearson correlation coefficients due to having plural time period ranges), wherein ∈is characterized by calculating an inverse proportion function value of accumulated difference values between the pearson correlation coefficients (accumulated difference values refer to absolute values of differences of pearson correlation coefficients of adjacent time periods calculated in order of time periods, and accumulated by plural times) >Is of a size of (a) and (b).
Acquisition of the firstDirect influence value of the respective related dimension data under the change of the other related dimension data +.>The calculation process is the same as->The present embodiment will not be described in detail.
The current dimension data is recorded as the firstThe individual dimension data;
obtain and divide the firstThe +.>Pearson correlation coefficient value of each correlation dimension data and current dimension data +.>: mapping function by pearson correlation coefficient +.>And->Pearson correlation coefficient value of each correlation dimension data and current dimension data +.>The method comprises the steps of obtaining a target product,if->Number of related dimensionsAccording to the Pelson correlation coefficient value and the +.>The pearson correlation coefficient value of the relevant dimension data and the current dimension data is simultaneously more than or equal to 0 or simultaneously less than 0, then the +.>The data distribution effect of the individual related dimension data on the current dimension data is positive; otherwise, let->The data distribution impact of the individual related dimension data on the current dimension data is reversed.
To sum up, the firstVariation distribution characteristics of each related dimension data and current dimension data>The calculation method of (1) is as follows:
；
in the method, in the process of the invention,indicate->The change distribution characteristics of the relevant dimension data and the current dimension data; / >Indicate->Direct influence values of the relevant dimension data under the condition that other relevant dimension data are unchanged; />Indicate->The dimension distribution of the relevant dimension data affects the weight value; />Mapping function representing pearson correlation coefficient, +.>Indicate->Pearson correlation coefficient values for the respective correlation dimension data and the current dimension data; />Indicate->Direct influence values of the relevant dimension data under the change of other relevant dimension data; />Indicating except->The total number of relevant dimension data of the current dimension data of the relevant dimension data.
To this end, obtain the firstThe relevant dimension data and the current dimension data change distribution characteristics.
According to the firstData space distribution characteristics of the individual related dimension data itself and +.>The variation distribution characteristics of the relevant dimension data and the current dimension data are obtained>The non-uniformity degree value of each relevant dimension data is calculated by the following expression:
；
in the method, in the process of the invention,indicate->A degree of non-uniformity value for each of the associated dimensional data; />Indicate->Data space distribution characteristics of the relevant dimension data; />Indicate->The change distribution characteristics of the relevant dimension data and the current dimension data; />An exponential function based on a natural constant is represented.
Thus, the non-uniformity degree value of the relevant dimension data is obtained.
Step S003: and obtaining an influence weight value of the relevant dimension data according to the non-uniformity degree value of the relevant dimension data of the current dimension data.
It should be noted that, by analyzing the influence weight value of the relevant dimension data, the boundary data point obtained by performing coarse K-Means on the multidimensional data when the host power supply is running is obtained, and the adaptive membership weight value between each cluster is obtained, wherein the larger the influence weight value of the relevant dimension data is, the more the distribution characteristics of the corresponding relevant dimension data and the current dimension data are analyzed. Therefore, the step is combined with the non-uniformity degree value of the related dimension data, and the allowable error of the similarity of the data distribution of the local range of the related dimension data and the local range of the current dimension data is obtained in a self-adaptive mode, so that the influence weight value of the related dimension data is obtained.
1. An allowable threshold fluctuation value of the relevant dimension data is acquired.
It should be noted that, the allowable threshold fluctuation value of the relevant dimension data is calculated according to the non-uniformity degree value of the relevant dimension data, wherein the larger the non-uniformity degree value of the relevant dimension data is, the non-uniformity of the distribution of the relevant dimension data is indicated, and the larger the corresponding allowable threshold fluctuation is when the similarity of the local range is calculated.
Presetting a threshold valueWherein the present embodiment is +.>Examples are described, the present embodiment is not particularly limited, wherein +.>Depending on the particular implementation.
Specifically, will beDegree of non-uniformity value of individual related dimension data +.>Performing linear normalization to obtain normalized +.>Degree of non-uniformity value of individual related dimension data +.>The method comprises the steps of carrying out a first treatment on the surface of the Then->Admission of individual related dimension dataThe calculation expression of the allowable threshold fluctuation value is:
；
in the method, in the process of the invention,indicate->Allowable threshold fluctuation values of the respective related dimension data; />Represents normalized->A degree of non-uniformity value for each of the associated dimensional data; />Is a preset threshold.
To this end, an allowable threshold fluctuation value of the relevant dimension data is obtained.
2. Data points of a local range of boundary data points of the current dimension data and the related dimension data are acquired.
Specifically, carrying out rough K-Means on the current dimension data and the related dimension data to obtain boundary data points of the current dimension data and the related dimension data; and acquiring the moment of the boundary data point of the current dimension data, and taking the moment as the center to acquire the distribution of all the data points in the time period range in the data space as the data point distribution of the local range of the boundary data point of the current dimension data, wherein the time period range is the first 6 moments and the last 6 moments of the moment, so as to acquire the data point of the local range of the boundary data point of the current dimension data. Similarly, data points of a local range of boundary data points of the relevant dimension data are obtained.
To this end, data points of a local range of boundary data points of the current dimension data and the related dimension data are obtained.
According to the phaseData points of the current dimension data and the data points of the local range of boundary data points of the current dimension data and the related dimension data are obtained by the allowable threshold fluctuation value of the dimension data and the data points of the local range of the boundary data points of the current dimension data and the related dimension dataSimilarity in data spatial distribution between data points of the respective related dimension data is calculated as:
；
in the method, in the process of the invention,data points representing current dimension data and +.>Similarity in data spatial distribution between data points of the individual related dimension data; />A number of data points representing a local range of boundary data points of the current dimension data;first +.>Data distribution characteristics in the neighborhood of data points by +.>The average Euclidean distance between data points within a neighborhood of data points (neighborhood is +.>Data points at distance->Data points nearest>Data point composition->For the preset parameters, by +.>The average Euclidean distance between data points within a neighborhood of data points (neighborhood is +. >Data points at distance->Data points nearest>Data point composition->For the preset parameters, the embodiments are not specifically limited, wherein +.>Depending on the particular implementation); />Indicate->The first +.>Data distribution characteristics within a neighborhood of data points; />Indicate->Allowable threshold fluctuation values of the respective related dimension data; />An exponential function representing a base of a natural constant is represented.
The embodiment combines the data point of the current dimension data with the firstSimilarity of data distribution of local range in data space distribution among data points of the respective related dimension data as +.>Influence weight value of each relevant dimension data.
So far, the influence weight value of the relevant dimension data is obtained.
Step S004: and according to the change of the data point distribution characteristics in the continuous incremental clustering process, and combining the influence weight value of the related dimension data to acquire the self-adaptive membership weight value of the boundary data point.
Because new current dimension data is continuously acquired with the development of time, in the process of continuously updating and clustering (which can be regarded as a mode of incremental clustering), the continuous current dimension data updating also has an effect on the change of boundary data points of the current dimension data, so that the membership weight value needs to be adjusted in the process of continuously updating according to the incremental clustering when the adaptive membership weight value is acquired.
It should be further noted that, in the process of incremental clustering after the current dimension data is newly added, the boundary data point corresponding to the current dimension data changes, and in this embodiment, when the membership weight value of the boundary data point of the updated current dimension data is obtained, the change from the boundary data point of the current dimension data updated last time to the boundary data point of the current dimension data needs to be considered.
Specifically, the first dimension of the current dimension dataThe boundary data point is about>Self-adaptive membership weight value of individual clustersThe calculation method of (1) is as follows:
；
in the method, in the process of the invention,represents the +.>The boundary data point is about>Self-adaptive membership weight values of the clusters; />The +.f representing the current dimension data now>Data points within the local range of the boundary data points to +.>Average Euclidean distance of the cluster centers; />Data points within the local range of all boundary data points representing the last current dimension data to +.>Average Euclidean distance of the cluster centers; />Total number of relevant dimension data representing current dimension data; />Indicate->Influence weight values of the relevant dimension data; />Indicate->The +. >A number of data points within a corresponding time period range of the plurality of boundary data points; />Indicate->The +.>Of the data points within the corresponding time range of the boundary data points, satisfy +.>Is (wherein +.>Indicate->The +.>Data distribution characteristics of data points within a time range corresponding to each boundary data point, < ->Indicate->An average data distribution characteristic of the individual clustered data points; />Indicate->Allowable threshold fluctuation values for the individual related dimension data).
Wherein,weight value representing change course of incremental clustering, if the value is larger, the average Euclidean distance is increased after updating, and the weight value corresponds to the +.>The boundary data point is about>The smaller the adaptive membership weight of the clusters; />The influence degree of the related dimension data is expressed, and if the influence weight value of certain related dimension data is larger, the corresponding distribution characteristics of the data points in the same time period range of the related dimension data are needed to be considered; />The larger the distribution of boundary data points indicating the relevant dimension data is, the more approaching +.>Clustering, boundary data points of the corresponding related dimension data are corresponding to +. >The greater the membership weight of the individual clusters. />
And similarly, obtaining the self-adaptive membership weight value of each boundary data point of the current dimension data for each cluster.
So far, the self-adaptive membership weight value of the boundary data point of the current dimension data to the cluster is obtained.
Step S005: and carrying out coarse K-Means clustering according to the self-adaptive membership weight value of the boundary data point to obtain an accurate detection result.
And (3) according to the self-adaptive membership weight value of each boundary data point of the current dimension data to each cluster, re-calculating the mean center of the weight value and the boundary data point in the original iterative updating process, wherein the process is a known technology and is not repeated in the embodiment, so that an accurate coarse K-Means clustering result is obtained. Similar operations result in clustering results of other dimensions. The clustering result is stored as a monitoring model, multidimensional data acquired in the operation monitoring process of the host power supply is added into the monitoring model, and the classification is carried out by calculating Euclidean distances from newly added dimensional data points to the centers of all clusters in the monitoring model. The embodiment with the largest data point number in the cluster in the monitoring model is considered as a normal cluster type, and the corresponding classification result of the new data is accurately monitored.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (6)
1. The method for monitoring the operation of the power supply of the omnibearing efficient heat dissipation host computer based on the Internet of things is characterized by comprising the following steps of:
acquiring multi-dimensional data of a host power supply during operation, and taking any one dimensional data as current dimensional data;
acquiring relevant dimension data of current dimension data, and recording the relevant dimension data as relevant dimension data; acquiring data space distribution characteristics of related dimension data; acquiring dimension distribution influence weight values of related dimension data; acquiring direct influence values of the related dimension data under the condition that other related dimension data are not changed; according to the dimension distribution influence weight value of the related dimension data and the direct influence value of the related dimension data under the condition that other related dimension data are not changed, the related dimension data and the current dimension data change distribution characteristics; obtaining the non-uniformity degree value of the relevant dimension data according to the data space distribution characteristics of the relevant dimension data and the change distribution characteristics of the relevant dimension data and the current dimension data;
Obtaining an allowable threshold fluctuation value of the relevant dimension data according to the non-uniformity degree value of the relevant dimension data; according to the allowable threshold fluctuation value of the relevant dimension data, the similarity between the data point of the current dimension data and the data point of the relevant dimension data in the data space distribution is used as an influence weight value of the relevant dimension data;
according to the influence weight value of the related dimension data, the self-adaptive membership weight value of the boundary data point of the current dimension data to the cluster;
clustering the clustered self-adaptive membership weight values according to boundary data points of the current dimension data to obtain a detection result;
the method for acquiring the data space distribution characteristics of the related dimension data comprises the following specific steps:
preset parametersFor->Related dimension data, in->Performing convex hull detection in the space of the data layer of each related dimension data to obtain the convex hull range with the maximum data in the space, obtaining the square area corresponding to the smallest circumscribed square of the convex hull range, and performing +_on the square area>Dividing equally to obtain->Each block is acquired againThe ratio between the average distance of the data points and the side length of the block is obtained by the following steps: the distances between all the data points in the block are obtained, and the distances are summed to obtain the average value of the distances as the average distance of the data points in the block;
Then the firstData space distribution characteristics of the individual related dimension data itself>The calculation method of (1) is as follows:
；
in the method, in the process of the invention,indicate->Data space distribution characteristics of the relevant dimension data; />Indicate->The data plane of the data of the relevant dimension data is +.>The ratio between the average distance of the data points in the individual tiles and the side length of the tile;indicate->All regions in the data plane of the data of each relevant dimensionA mean value of the ratio between the average distance of the data points in the block and the side length of the block;
the method for acquiring the direct influence value of the related dimension data under the condition that other related dimension data are not changed comprises the following specific steps:
for the firstThe relevant dimension data is->Direct influence value of the respective relevant dimension data without change of the other relevant dimension data +.>The acquisition method comprises the following steps: calculating the data change curve and the +.f of the current dimension data in the time range of the other related dimension data unchanged>Pearson correlation coefficients between data change curves of the respective correlation dimension data are characterized by calculating inverse proportion function values of accumulated difference values between the pearson correlation coefficients>The accumulated difference value refers to the absolute value of the difference value of the pearson correlation coefficients of adjacent time periods according to the sequence of the time periods, and then the absolute values are accumulated in a plurality of sections;
The method comprises the following specific steps of:
for the firstRelated dimension data, th->The method for calculating the change distribution characteristics of the relevant dimension data and the current dimension data comprises the following steps:
；
in the method, in the process of the invention,indicate->The change distribution characteristics of the relevant dimension data and the current dimension data; />Indicate->Direct influence values of the relevant dimension data under the condition that other relevant dimension data are unchanged; />Indicate->The dimension distribution of the relevant dimension data affects the weight value; />Mapping function representing pearson correlation coefficient, +.>Indicate->Pearson correlation coefficient values for the respective correlation dimension data and the current dimension data; />Indicate->Direct influence values of the relevant dimension data under the change of other relevant dimension data; />Indicating except->The total number of relevant dimension data of the current dimension data of the relevant dimension data;
the self-adaptive membership weight value of the boundary data points of the current dimension data to the clusters according to the influence weight value of the related dimension data comprises the following specific steps:
For the current dimension dataThe boundary data point is the +.>The boundary data point is relative to the firstAdaptive membership weight value for individual clusters +.>Is calculated by the following steps:
；
in the method, in the process of the invention,represents the +.>The boundary data point is about>Self-adaptive membership weight values of the clusters; />The +.f representing the current dimension data now>Data points within the local range of the boundary data point to the firstAverage Euclidean distance of the cluster centers; />Data points within the local range of all boundary data points representing the last current dimension data to +.>Average Euclidean distance of the cluster centers; />Total number of relevant dimension data representing current dimension data; />Indicate->Influence weight values of the relevant dimension data; />Indicate->The +.>A number of data points within a corresponding time period range of the plurality of boundary data points; />Indicate->The +.>Of the data points within the corresponding time period of the boundary data points, the +.>The +.>Data distribution characteristics and +.>The average data distribution characteristic of the data points of the clusters is less than +. >The number of data points of the allowable threshold fluctuation value of the relevant dimension data;
and similarly, obtaining the self-adaptive membership weight value of each boundary data point of the current dimension data for each cluster.
2. The method for monitoring the operation of the power supply of the omnidirectional efficient heat dissipation host based on the internet of things according to claim 1, wherein the step of obtaining the relevant dimension data of the current dimension data comprises the following specific steps:
calculating pearson correlation coefficients of the current dimension data and other dimension data; if there is dimension data in other dimension dataThe pearson correlation coefficient with the current dimension data is greater than a preset threshold +.>Dimension data is->And similarly, acquiring all relevant dimension data of the current dimension data as relevant dimension data of the current dimension data.
3. The method for monitoring the operation of the power supply of the omnidirectional efficient heat dissipation host based on the internet of things according to claim 1, wherein the step of obtaining the dimension distribution influence weight value of the related dimension data comprises the following specific steps:
for the firstRelated dimension data, except +.>The length of the period of time during which the other relevant dimension data of the relevant dimension data will not change, then +. >Dimension distribution influence weight value +.>Wherein->Representing the total time period length of the multi-dimensional data when the host power supply is in operation, < >>Indicating except->The length of the time period occupied when other dimension data of the relevant dimension data is unchanged.
4. The method for monitoring the operation of the power supply of the omnidirectional efficient heat dissipation host based on the internet of things according to claim 1, wherein the method for obtaining the non-uniformity degree value of the relevant dimension data according to the data space distribution characteristic of the relevant dimension data and the variation distribution characteristic of the relevant dimension data and the current dimension data comprises the following specific steps:
for the firstRelated dimension data according to->Data space distribution characteristics of the individual related dimension data itself and +.>The variation distribution characteristics of the relevant dimension data and the current dimension data are obtained>The non-uniformity degree value of each relevant dimension data is calculated by the following expression:
；
in the method, in the process of the invention,indicate->A degree of non-uniformity value for each of the associated dimensional data; />Indicate->Data space distribution characteristics of the relevant dimension data; />Indicate->The change distribution characteristics of the relevant dimension data and the current dimension data; />An exponential function based on a natural constant is represented.
5. The method for monitoring the operation of the power supply of the omnidirectional efficient heat dissipation host based on the internet of things according to claim 1, wherein the method for obtaining the allowable threshold fluctuation value of the relevant dimension data according to the non-uniformity degree value of the relevant dimension data comprises the following specific steps:
for the firstRelated dimension data, will be->Degree of non-uniformity value of individual related dimension data +.>Performing linear normalization to obtain normalized +.>Degree of non-uniformity value of individual related dimension data +.>The method comprises the steps of carrying out a first treatment on the surface of the Then 1 minus normalized +.>Degree of non-uniformity value of individual related dimension data +.>And a preset threshold->Is the product of->The allowable threshold fluctuation value of each relevant dimension data.
6. The method for monitoring the operation of the power supply of the omnidirectional efficient heat dissipation host computer based on the internet of things according to claim 1, wherein the similarity between the data point of the current dimension data and the data point of the relevant dimension data in terms of data space distribution is used as the influence weight value of the relevant dimension data according to the allowable threshold fluctuation value of the relevant dimension data, and the method comprises the following specific steps:
for the firstRelated dimension data, the data point and the +.th of the current dimension data are obtained according to the allowable threshold fluctuation value of the related dimension data and the data point of the local range of the boundary data point of the current dimension data and the related dimension data >Similarity in data spatial distribution between data points of the respective related dimension data is taken as +.>The influence weight value of each relevant dimension data is calculated as follows:
；
in the method, in the process of the invention,indicate->Influence weight values of the relevant dimension data; />A number of data points representing a local range of boundary data points of the current dimension data; />First +.>Data distribution characteristics within a neighborhood of data points; />Indicate->The first +.>Data distribution characteristics within a neighborhood of data points; />Indicate->Allowable threshold fluctuation values of the respective related dimension data; />An exponential function based on a natural constant is represented.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311198369.3A CN116954342B (en) | 2023-09-18 | 2023-09-18 | Method for monitoring operation of power supply of omnibearing efficient heat dissipation host based on Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311198369.3A CN116954342B (en) | 2023-09-18 | 2023-09-18 | Method for monitoring operation of power supply of omnibearing efficient heat dissipation host based on Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116954342A CN116954342A (en) | 2023-10-27 |
| CN116954342B true CN116954342B (en) | 2023-12-15 |
Family
ID=88458667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311198369.3A Active CN116954342B (en) | 2023-09-18 | 2023-09-18 | Method for monitoring operation of power supply of omnibearing efficient heat dissipation host based on Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116954342B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117240602B (en) * | 2023-11-09 | 2024-01-19 | 北京中海通科技有限公司 | Identity authentication platform safety protection method |
| CN117289778B (en) * | 2023-11-27 | 2024-03-26 | 惠州市鑫晖源科技有限公司 | Real-time monitoring method for health state of industrial control host power supply |
| CN117421620B (en) * | 2023-12-18 | 2024-02-27 | 北京云摩科技股份有限公司 | Interaction method of tension state data |
| CN117435874B (en) * | 2023-12-21 | 2024-03-12 | 河北雄安睿天科技有限公司 | Abnormal data detection method and system for water supply and drainage equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022110557A1 (en) * | 2020-11-25 | 2022-06-02 | 国网湖南省电力有限公司 | Method and device for diagnosing user-transformer relationship anomaly in transformer area |
| CN115953604A (en) * | 2023-03-13 | 2023-04-11 | 泰安市金土地测绘整理有限公司 | Real estate geographic information mapping data acquisition method |
| CN116320042A (en) * | 2023-05-16 | 2023-06-23 | 陕西思极科技有限公司 | Internet of things terminal monitoring control system for edge calculation |
| CN116340801A (en) * | 2023-05-31 | 2023-06-27 | 烟台市福山区动物疫病预防控制中心 | Intelligent monitoring method for abnormal environmental data of livestock breeding |
| CN116431975A (en) * | 2023-06-12 | 2023-07-14 | 陕西巨人商务信息咨询有限公司 | Environment monitoring method and system for data center |
| CN116610482A (en) * | 2023-07-18 | 2023-08-18 | 山东理工大学 | Intelligent monitoring method for operation state of electrical equipment |
| CN116702080A (en) * | 2023-08-04 | 2023-09-05 | 山东荣信集团有限公司 | Gas system methanol-to-liquid process on-line monitoring system based on multidimensional sensor |
-
2023
- 2023-09-18 CN CN202311198369.3A patent/CN116954342B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022110557A1 (en) * | 2020-11-25 | 2022-06-02 | 国网湖南省电力有限公司 | Method and device for diagnosing user-transformer relationship anomaly in transformer area |
| CN115953604A (en) * | 2023-03-13 | 2023-04-11 | 泰安市金土地测绘整理有限公司 | Real estate geographic information mapping data acquisition method |
| CN116320042A (en) * | 2023-05-16 | 2023-06-23 | 陕西思极科技有限公司 | Internet of things terminal monitoring control system for edge calculation |
| CN116340801A (en) * | 2023-05-31 | 2023-06-27 | 烟台市福山区动物疫病预防控制中心 | Intelligent monitoring method for abnormal environmental data of livestock breeding |
| CN116431975A (en) * | 2023-06-12 | 2023-07-14 | 陕西巨人商务信息咨询有限公司 | Environment monitoring method and system for data center |
| CN116610482A (en) * | 2023-07-18 | 2023-08-18 | 山东理工大学 | Intelligent monitoring method for operation state of electrical equipment |
| CN116702080A (en) * | 2023-08-04 | 2023-09-05 | 山东荣信集团有限公司 | Gas system methanol-to-liquid process on-line monitoring system based on multidimensional sensor |
Non-Patent Citations (1)
| Title |
|---|
| 基于安全监测***的大坝安全多层次模糊综合评判方法;郑付刚;游强强;;河海大学学报(自然科学版)(第04期);第407-413页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116954342A (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116954342B (en) | Method for monitoring operation of power supply of omnibearing efficient heat dissipation host based on Internet of things | |
| CN107682319B (en) | Enhanced angle anomaly factor-based data flow anomaly detection and multi-verification method | |
| CN108805213B (en) | Power load curve double-layer spectral clustering method considering wavelet entropy dimensionality reduction | |
| CN108446599B (en) | Hyperspectral image band rapid selection method of p-value statistical modeling independence | |
| CN117113235B (en) | Cloud computing data center energy consumption optimization method and system | |
| CN116609720B (en) | Data-driven-based intelligent error compensation method and system for desk-top multimeter | |
| CN113255236B (en) | Non-invasive load self-adaptive identification method based on twin network | |
| CN110795690A (en) | Wind power plant operation abnormal data detection method | |
| CN113159106B (en) | Load curve clustering method, medium and equipment based on morphological trend characteristics | |
| CN117454255B (en) | Intelligent building energy consumption data optimized storage method | |
| CN117493921B (en) | Artificial intelligence energy-saving management method and system based on big data | |
| WO2023044770A1 (en) | Dry pump downtime early warning method and apparatus, electronic device, storage medium, and program | |
| CN113609901A (en) | Power transmission and transformation equipment fault monitoring method and system | |
| CN113360656A (en) | Abnormal data detection method, device, equipment and storage medium | |
| CN113918642A (en) | Data filtering, monitoring and early warning method based on power Internet of things equipment | |
| CN114056599B (en) | Airplane climate environment adaptive capacity evaluation method based on grey correlation degree | |
| CN117540325B (en) | Business database anomaly detection method and system based on data variation capture | |
| CN108090635B (en) | Road performance prediction method based on cluster classification | |
| CN114357870A (en) | Metering equipment operation performance prediction analysis method based on local weighted partial least squares | |
| CN117522864A (en) | European pine plate surface flaw detection method based on machine vision | |
| US20220114460A1 (en) | Apparatus of Identifying Heterogeneous Time-Series Data Expression with High Efficiency | |
| TWI710960B (en) | Image classification system and method | |
| CN118152829B (en) | Health state assessment method and system for oil type iron core reactor | |
| CN116661522B (en) | Intelligent temperature regulation and control method for temperature change test box based on data processing | |
| CN117648657B (en) | Urban planning multi-source data optimization processing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |