CN117236572B - Method and system for evaluating performance of dry powder extinguishing equipment based on data analysis - Google Patents
Method and system for evaluating performance of dry powder extinguishing equipment based on data analysis Download PDFInfo
- Publication number
- CN117236572B CN117236572B CN202311510615.4A CN202311510615A CN117236572B CN 117236572 B CN117236572 B CN 117236572B CN 202311510615 A CN202311510615 A CN 202311510615A CN 117236572 B CN117236572 B CN 117236572B
- Authority
- CN
- China
- Prior art keywords
- real
- time performance
- performance parameter
- parameters
- dry powder
- 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
- 239000000843 powder Substances 0.000 title claims abstract description 150
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000007405 data analysis Methods 0.000 title claims abstract description 22
- 238000011156 evaluation Methods 0.000 claims abstract description 25
- 230000000694 effects Effects 0.000 claims abstract description 13
- 238000010586 diagram Methods 0.000 claims description 48
- 230000002159 abnormal effect Effects 0.000 claims description 30
- 239000011159 matrix material Substances 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 18
- 239000000203 mixture Substances 0.000 claims description 17
- 238000009826 distribution Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 11
- 238000000926 separation method Methods 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 description 7
- 238000005507 spraying Methods 0.000 description 6
- 239000000243 solution Substances 0.000 description 5
- 230000007704 transition Effects 0.000 description 5
- 230000007547 defect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of fire extinguishing equipment evaluation, in particular to a method and a system for evaluating the performance of dry powder fire extinguishing equipment based on data analysis, wherein a plurality of initial real-time performance parameter subsets are obtained by clustering real-time performance parameters in a database; acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, verifying and correcting clustering effects of each initial real-time performance parameter subset, and obtaining a plurality of final real-time performance parameter subsets; and comparing and analyzing the final real-time performance parameter subset with the corresponding standard performance parameter subset to judge the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index, so that the performance parameters with high accuracy and high integrity can be obtained, the performance of the dry powder extinguishing equipment can be evaluated more accurately, and the evaluation efficiency is effectively improved.
Description
Technical Field
The invention relates to the technical field of fire extinguishing equipment evaluation, in particular to a dry powder fire extinguishing equipment performance evaluation method and system based on data analysis.
Background
With the advancement of technology, conventional fire extinguishing apparatus evaluation methods have failed to meet the refined production requirements. Therefore, the utilization of data analysis technology, especially big data and machine learning algorithm, to evaluate the performance of dry powder extinguishing equipment becomes a prospective solution. Although the intelligent evaluation method for the performance of the dry powder extinguishing equipment based on data analysis has a plurality of advantages, one of the technical defects is a problem of data quality, the accuracy of an evaluation model is highly dependent on the quality and the integrity of input data, and if the data is wrong, inconsistent or incomplete, errors of an evaluation result can be caused to influence the correct evaluation of the performance of the dry powder extinguishing equipment. Moreover, the creation and training of data analysis models requires a significant amount of computing resources and time, and complex machine learning algorithms and model training processes may require high performance computing devices, which increases the difficulty and cost of technical implementation.
Disclosure of Invention
The invention overcomes the defects of the prior art and provides a dry powder extinguishing equipment performance evaluation method and system based on data analysis.
The technical scheme adopted by the invention for achieving the purpose is as follows:
the invention discloses a dry powder extinguishing equipment performance evaluation method based on data analysis, which comprises the following steps:
acquiring an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, performing pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and constructing a plurality of standard performance parameter subsets according to the standard performance parameters;
detecting dry powder extinguishing equipment to be evaluated, acquiring real-time performance parameters fed back by sensors in the dry powder extinguishing equipment at a plurality of preset time nodes in the detection process, constructing a database, and inputting the real-time performance parameters into the database;
after the acquisition is finished, clustering the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; wherein, each initial real-time performance parameter subset stores the real-time performance parameters of the same type;
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets;
And comparing and analyzing the final real-time performance parameter subset with the corresponding standard performance parameter subset, judging the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index.
Further, in a preferred embodiment of the present invention, an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated is obtained, and the equipment three-dimensional model diagram is paired and identified to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, which specifically includes:
prefabricating standard three-dimensional model diagrams of various types of dry powder fire extinguishing equipment and prefabricating standard performance parameters of the various types of dry powder fire extinguishing equipment;
binding standard three-dimensional model diagrams of various types of dry powder fire extinguishing equipment with corresponding standard performance parameters to obtain a plurality of standard data packets; constructing a knowledge graph, and importing a plurality of standard data packets into the knowledge graph;
acquiring image information of dry powder extinguishing equipment to be evaluated, and constructing an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated according to the image information;
searching standard three-dimensional model diagrams of various types of dry powder extinguishing equipment from the knowledge graph, and calculating the similarity between the equipment three-dimensional model diagrams and each standard three-dimensional model diagram through a Hausdorff distance algorithm to obtain a plurality of similarities;
And selecting the maximum similarity from the plurality of similarities, acquiring a standard three-dimensional model diagram corresponding to the maximum similarity, acquiring a standard data packet of the standard three-dimensional model diagram corresponding to the maximum similarity, and acquiring standard performance parameters of the dry powder extinguishing equipment to be evaluated from the standard data packet.
Further, in a preferred embodiment of the present invention, the real-time performance parameters in the database are clustered to obtain a plurality of initial real-time performance parameter subsets, specifically:
acquiring the number of performance parameters to be evaluated in dry powder extinguishing equipment to be evaluated, determining the number of Gaussian components according to the number of the performance parameters to be evaluated, and determining a plurality of cluster families according to the number of the Gaussian components;
randomly selecting a real-time performance parameter from the database as an initial average value of the Gaussian component; calculating covariance among all real-time performance parameters in a database, and generating an initial covariance matrix of a Gaussian component according to the covariance; randomly generating a positive number for each cluster group, and carrying out normalization processing on the positive number to obtain the initial weight of the Gaussian component;
determining posterior probability of each piece of real-time performance data according to the initial mean value, the initial covariance matrix and the initial weight of the Gaussian components, and re-determining a new mean value, a new covariance matrix and a new weight of the Gaussian components according to the posterior probability of each piece of real-time performance data;
Determining the log-likelihood value of each real-time performance parameter according to the new mean value, the new covariance matrix and the new weight, summing all the log-likelihood values to obtain a total log-likelihood value, and determining the fitting degree of the Gaussian mixture model according to the total log-likelihood value;
if the fitting degree is larger than the preset fitting degree, saving model parameters to obtain a Gaussian mixture model; if the fitting degree is not greater than the preset fitting degree, continuing iteration until the fitting degree is greater than the preset fitting degree, and storing model parameters to obtain a Gaussian mixture model;
and acquiring posterior probability of each real-time performance parameter belonging to each cluster family again in the Gaussian mixture model, clustering each real-time performance parameter into the cluster family with the largest posterior probability, and obtaining a plurality of initial real-time performance parameter subsets after clustering.
Further, in a preferred embodiment of the present invention, the coordinate values of the real-time performance parameters in the low-dimensional space in each initial real-time performance parameter subset are obtained, specifically:
extracting corresponding real-time performance parameters from each initial real-time performance parameter subset, measuring the similarity between the real-time performance parameters in each initial real-time performance parameter subset through a polynomial kernel function, and obtaining a similarity matrix of each initial real-time performance parameter subset based on the similarity between the real-time performance parameters of each initial real-time performance parameter subset;
Acquiring the relative distance of the real-time performance parameters in the high-dimensional space in each initial real-time performance parameter subset, and calculating the conditional probability distribution of each real-time performance parameter in the high-dimensional space based on the relative distance and the corresponding similarity matrix;
randomly initializing the relative positions of the real-time performance parameters in the low-dimensional space in each initial real-time performance parameter subset, and calculating according to the relative positions and the corresponding similarity matrix to obtain the conditional probability distribution of each real-time performance parameter in the low-dimensional space;
calculating KL divergence between the high-dimensional real-time performance parameters and the low-dimensional real-time performance parameters according to the conditional probability distribution of each real-time performance parameter in the high-dimensional space and the conditional probability distribution of each real-time performance parameter in the low-dimensional space;
the KL divergence between the high-dimensional real-time performance parameter and the low-dimensional real-time performance parameter is minimized through a gradient descent optimization algorithm, so that the position of the real-time performance parameter in the low-dimensional space is adjusted; and repeating the steps until convergence conditions are reached, determining the position of each real-time performance parameter in the low-dimensional space, and acquiring the coordinate value of each real-time performance parameter in the low-dimensional space.
Further, in a preferred embodiment of the present invention, the Davies-Bouldin index of each initial real-time performance parameter subset is calculated according to the coordinate values, specifically:
Acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset according to the coordinate values, and carrying out average value taking treatment on the Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset to obtain intra-cluster compactness;
calculating Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets according to the coordinate values, and carrying out average value taking processing on the Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets to obtain the cluster separation degree;
calculating the Davies-Bouldin indexes of each initial real-time performance parameter subset according to the intra-cluster compactness and the inter-cluster separation degree, and comparing the Davies-Bouldin indexes of each initial real-time performance parameter subset with a preset threshold;
marking an initial real-time performance parameter subset with the Davies-Bouldin index larger than a preset threshold value as a clustering abnormal parameter subset; and marking the initial real-time performance parameter subset with the Davies-Bouldin index not larger than a preset threshold as a clustering normal parameter subset.
Further, in a preferred embodiment of the present invention, the clustering effect of each initial real-time performance parameter subset is verified and corrected according to the Davies-Bouldin index, so as to obtain a plurality of final real-time performance parameter subsets, which are specifically:
acquiring real-time performance parameters in the clustering abnormal parameter subset, calculating the local density and the relative local density of each real-time performance parameter through an LOF algorithm, and calculating the LOF value of each real-time performance parameter according to the local density and the relative local density of each real-time performance parameter;
comparing LOF values of the real-time performance parameters with preset LOF values, marking the real-time performance parameters with LOF values larger than the preset LOF values in the clustering abnormal parameter subsets as singular parameters, and eliminating the singular parameters in the corresponding clustering abnormal parameter subsets;
the singular parameters are imported into the remaining real-time performance parameter subsets, after the abnormal parameters are imported into the remaining real-time performance parameter subsets, the Davies-Bouldin indexes of the corresponding real-time performance parameter subsets are recalculated, and the recalculated Davies-Bouldin indexes are compared with a preset threshold value;
if the recalculated Davies-Bouldin indexes are all larger than a preset threshold, the singular parameters are thoroughly deleted;
If the recalculated Davies-Bouldin index is not greater than a preset threshold value, sequencing the recalculated Davies-Bouldin index based on numerical values to extract the minimum Davies-Bouldin index, and clustering the singular parameters into a real-time performance parameter subset corresponding to the minimum Davies-Bouldin index;
repeating the steps until all the clustering abnormal parameter subsets are verified and corrected, and updating all the initial real-time performance parameter subsets to obtain a plurality of final real-time performance parameter subsets.
Further, in a preferred embodiment of the present invention, the final real-time performance parameter subset is compared with the corresponding standard performance parameter subset to determine the standard condition of each performance index in the dry powder extinguishing device to be evaluated, and the performance grade of the dry powder extinguishing device is evaluated according to the standard condition of each performance index, specifically:
acquiring real-time performance parameters in each final real-time performance parameter subset, calculating the coincidence degree between the real-time performance parameters in each final real-time performance parameter subset and standard performance parameters in the corresponding standard performance parameter subset through a cosine similarity algorithm, and comparing the coincidence degree with a preset coincidence degree;
If the overlap ratio is not greater than the preset overlap ratio, marking the performance index corresponding to the final real-time performance parameter subset as a substandard index; if the overlap ratio is larger than the preset overlap ratio, marking the performance index corresponding to the final real-time performance parameter subset as a standard reaching index;
the method comprises the steps of constructing an information base according to the corresponding performance grades of all performance indexes of the prefabricated dry powder extinguishing equipment under various standard-reaching or non-standard-reaching combined conditions, and importing the corresponding performance grades of all the performance indexes of the dry powder extinguishing equipment under various standard-reaching or non-standard-reaching combined conditions into the information base;
counting the up-to-standard or non-up-to-standard conditions of all the performance indexes in the current dry powder extinguishing equipment to be evaluated, guiding the up-to-standard or non-up-to-standard conditions of all the performance indexes in the current dry powder extinguishing equipment to be evaluated into the information base for pairing evaluation to obtain the performance grade of the current dry powder extinguishing equipment to be evaluated, and conveying the performance grade of the current dry powder extinguishing equipment to be evaluated to a preset platform for display.
The invention further discloses a dry powder extinguishing equipment performance evaluation system based on data analysis, which comprises a memory and a processor, wherein the memory stores a dry powder extinguishing equipment performance evaluation method program, and when the dry powder extinguishing equipment performance evaluation method program is executed by the processor, the following steps are realized:
Acquiring an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, performing pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and constructing a plurality of standard performance parameter subsets according to the standard performance parameters;
detecting dry powder extinguishing equipment to be evaluated, acquiring real-time performance parameters fed back by sensors in the dry powder extinguishing equipment at a plurality of preset time nodes in the detection process, constructing a database, and inputting the real-time performance parameters into the database;
after the acquisition is finished, clustering the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; wherein, each initial real-time performance parameter subset stores the real-time performance parameters of the same type;
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets;
And comparing and analyzing the final real-time performance parameter subset with the corresponding standard performance parameter subset, judging the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index.
The invention solves the technical defects existing in the background technology, and has the following beneficial effects: clustering real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets; the final real-time performance parameter subset is compared with the corresponding standard performance parameter subset to judge the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and the performance grade of the dry powder extinguishing equipment is evaluated according to the standard condition of each performance index.
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 embodiments of the drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a first method of evaluating the performance of a dry powder extinguishing device based on data analysis;
FIG. 2 is a second method flow chart of a method for evaluating the performance of a dry powder extinguishing device based on data analysis;
FIG. 3 is a third method flow chart of a method for evaluating the performance of a dry powder extinguishing device based on data analysis;
fig. 4 is a system block diagram of a dry powder extinguishing apparatus performance evaluator system based on data analysis.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, in the case of no conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
As shown in fig. 1, the first aspect of the present invention discloses a method for evaluating performance of a dry powder extinguishing apparatus based on data analysis, comprising the steps of:
s102: acquiring an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, performing pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and constructing a plurality of standard performance parameter subsets according to the standard performance parameters;
s104: detecting dry powder extinguishing equipment to be evaluated, acquiring real-time performance parameters fed back by sensors in the dry powder extinguishing equipment at a plurality of preset time nodes in the detection process, constructing a database, and inputting the real-time performance parameters into the database;
s106: after the acquisition is finished, clustering the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; wherein, each initial real-time performance parameter subset stores the real-time performance parameters of the same type;
S108: acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets;
s110: and comparing and analyzing the final real-time performance parameter subset with the corresponding standard performance parameter subset, judging the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index.
Specifically, an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated is obtained, pairing identification is carried out on the equipment three-dimensional model diagram, and standard performance parameters of the dry powder extinguishing equipment to be evaluated are obtained, wherein the standard performance parameters specifically comprise:
prefabricating standard three-dimensional model diagrams of various types of dry powder fire extinguishing equipment and prefabricating standard performance parameters of the various types of dry powder fire extinguishing equipment;
binding standard three-dimensional model diagrams of various types of dry powder fire extinguishing equipment with corresponding standard performance parameters to obtain a plurality of standard data packets; constructing a knowledge graph, and importing a plurality of standard data packets into the knowledge graph;
Acquiring image information of dry powder extinguishing equipment to be evaluated, and constructing an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated according to the image information;
searching standard three-dimensional model diagrams of various types of dry powder extinguishing equipment from the knowledge graph, and calculating the similarity between the equipment three-dimensional model diagrams and each standard three-dimensional model diagram through a Hausdorff distance algorithm to obtain a plurality of similarities;
it is mentioned that the Chinese name of Hausdorff distance algorithm is Hausdorff distance. The Haoskov distance algorithm is widely applied to the fields of three-dimensional model retrieval, object identification, matching and the like;
and selecting the maximum similarity from the plurality of similarities, acquiring a standard three-dimensional model diagram corresponding to the maximum similarity, acquiring a standard data packet of the standard three-dimensional model diagram corresponding to the maximum similarity, and acquiring standard performance parameters of the dry powder extinguishing equipment to be evaluated from the standard data packet.
It should be noted that, the standard performance parameters such as fire extinguishing speed, spraying distance, spraying time, powder consumption of the dry powder extinguishing equipment of different types are different, so before the dry powder extinguishing equipment to be evaluated is evaluated, the standard performance parameters corresponding to the dry powder extinguishing equipment need to be determined according to the type of the dry powder extinguishing equipment. In the step, the image information of the dry powder extinguishing equipment to be evaluated can be obtained through a camera, so that an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated is constructed by utilizing a point cloud reconstruction mode, then the equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated is subjected to matching identification, so that standard performance parameters corresponding to the dry powder extinguishing equipment to be evaluated are quickly matched, and a plurality of standard performance parameter subsets are constructed, such as standard spraying distance, powder consumption and the like of the dry powder extinguishing equipment to be evaluated at each preset time node in a preset test period.
Specifically, the real-time performance parameters in the database are clustered to obtain a plurality of initial real-time performance parameter subsets, specifically:
acquiring the number of performance parameters to be evaluated in dry powder extinguishing equipment to be evaluated, determining the number of Gaussian components according to the number of the performance parameters to be evaluated, and determining a plurality of cluster families according to the number of the Gaussian components;
randomly selecting a real-time performance parameter from the database as an initial average value of the Gaussian component; calculating covariance among all real-time performance parameters in a database, and generating an initial covariance matrix of a Gaussian component according to the covariance; randomly generating a positive number for each cluster group, and carrying out normalization processing on the positive number to obtain the initial weight of the Gaussian component;
determining posterior probability of each piece of real-time performance data according to the initial mean value, the initial covariance matrix and the initial weight of the Gaussian components, and re-determining a new mean value, a new covariance matrix and a new weight of the Gaussian components according to the posterior probability of each piece of real-time performance data;
determining the log-likelihood value of each real-time performance parameter according to the new mean value, the new covariance matrix and the new weight, summing all the log-likelihood values to obtain a total log-likelihood value, and determining the fitting degree of the Gaussian mixture model according to the total log-likelihood value;
If the fitting degree is larger than the preset fitting degree, saving model parameters to obtain a Gaussian mixture model; if the fitting degree is not greater than the preset fitting degree, continuing iteration until the fitting degree is greater than the preset fitting degree, and storing model parameters to obtain a Gaussian mixture model;
and acquiring posterior probability of each real-time performance parameter belonging to each cluster family again in the Gaussian mixture model, clustering each real-time performance parameter into the cluster family with the largest posterior probability, and obtaining a plurality of initial real-time performance parameter subsets after clustering.
The method is characterized in that the dry powder fire extinguishing equipment to be evaluated is detected through the detection equipment, real-time performance parameters fed back by sensors in the dry powder fire extinguishing equipment are collected at a plurality of preset time nodes in the detection process, a database is constructed, and the real-time performance parameters are input into the database, for example, pressure parameter data information of a nozzle is continuously collected through a pressure sensor. After the test is finished, the massive performance parameters acquired by the sensors are concentrated in a database, and the massive performance parameters are required to be classified at the moment so as to distinguish the pressure parameters, the spraying distance parameters and the like. Specifically, these data are classified by a gaussian mixture model, the mean, covariance matrix, and weight of each gaussian component are initialized by random, and the posterior probability, i.e., responsibility, of each parameter data belonging to each gaussian component is calculated; re-estimating the mean value, covariance matrix and weight of each Gaussian component according to responsibility; the change of log likelihood is used for checking whether the Gaussian mixture model is fitted to preset requirements, so that the trained Gaussian mixture model is used for classifying massive parameters, and a plurality of initial real-time performance parameter subsets are obtained; the real-time performance parameters of the same type are stored in each initial real-time performance parameter subset respectively, and if the pressure parameters acquired in the testing process are clustered in one initial real-time performance parameter subset. By the method, massive real-time performance parameters can be clustered, so that different types of performance parameters can be obtained quickly, the robustness of the system can be improved, and the evaluation efficiency is improved.
As shown in fig. 2, specifically, coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset are obtained, specifically:
s202: extracting corresponding real-time performance parameters from each initial real-time performance parameter subset, measuring the similarity between the real-time performance parameters in each initial real-time performance parameter subset through a polynomial kernel function, and obtaining a similarity matrix of each initial real-time performance parameter subset based on the similarity between the real-time performance parameters of each initial real-time performance parameter subset;
s204: acquiring the relative distance of the real-time performance parameters in the high-dimensional space in each initial real-time performance parameter subset, and calculating the conditional probability distribution of each real-time performance parameter in the high-dimensional space based on the relative distance and the corresponding similarity matrix;
s206: randomly initializing the relative positions of the real-time performance parameters in the low-dimensional space in each initial real-time performance parameter subset, and calculating according to the relative positions and the corresponding similarity matrix to obtain the conditional probability distribution of each real-time performance parameter in the low-dimensional space;
s208: calculating KL divergence between the high-dimensional real-time performance parameters and the low-dimensional real-time performance parameters according to the conditional probability distribution of each real-time performance parameter in the high-dimensional space and the conditional probability distribution of each real-time performance parameter in the low-dimensional space;
S210: the KL divergence between the high-dimensional real-time performance parameter and the low-dimensional real-time performance parameter is minimized through a gradient descent optimization algorithm, so that the position of the real-time performance parameter in the low-dimensional space is adjusted; and repeating the steps until convergence conditions are reached, determining the position of each real-time performance parameter in the low-dimensional space, and acquiring the coordinate value of each real-time performance parameter in the low-dimensional space.
It should be noted that, after the massive real-time performance parameters are clustered by the gaussian mixture model, the model is difficult to avoid the occurrence of the local optimal solution phenomenon, so that a clustering error phenomenon may occur, for example, a pressure parameter at a certain moment is clustered into a subset of the injection distance, so that a t-SNE algorithm and a Davies-Bouldin index are introduced to further verify whether the clustering error phenomenon occurs. Firstly, after a plurality of initial real-time performance parameter subsets are obtained through Gaussian mixture model clustering, the real-time performance parameters in the initial real-time performance parameter subsets are subjected to dimension reduction processing through a t-SNE algorithm, so that high-dimensional parameter data are mapped to a low-dimensional space, the parameter data are expressed in a dot form, and the structure and the relation of the parameter data are understood.
Specifically, the Davies-Bouldin index of each initial real-time performance parameter subset is calculated according to the coordinate values, specifically:
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset according to the coordinate values, and carrying out average value taking treatment on the Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset to obtain intra-cluster compactness;
calculating Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets according to the coordinate values, and carrying out average value taking processing on the Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets to obtain the cluster separation degree;
calculating the Davies-Bouldin indexes of each initial real-time performance parameter subset according to the intra-cluster compactness and the inter-cluster separation degree, and comparing the Davies-Bouldin indexes of each initial real-time performance parameter subset with a preset threshold;
marking an initial real-time performance parameter subset with the Davies-Bouldin index larger than a preset threshold value as a clustering abnormal parameter subset; and marking the initial real-time performance parameter subset with the Davies-Bouldin index not larger than a preset threshold as a clustering normal parameter subset.
It should be noted that the process of evaluating cluster quality by Davies-Bouldin index (DBI index) involves calculating the degree of separation between each cluster and the compactness inside the cluster, and then combining these values into a composite index, with lower DBI index indicating tighter inside the cluster and more separation between clusters, which is typically a good clustering result; a higher DBI index indicates a poor cluster quality, a relatively low compactness inside the clusters or a relatively high degree of separation between clusters. By the method, whether each initial real-time performance parameter subset is a subset of clustering abnormality can be judged.
Specifically, verifying and correcting the clustering effect of each initial real-time performance parameter subset according to the Davies-Bouldin index to obtain a plurality of final real-time performance parameter subsets, wherein the method specifically comprises the following steps:
acquiring real-time performance parameters in the clustering abnormal parameter subset, calculating the local density and the relative local density of each real-time performance parameter through an LOF algorithm, and calculating the LOF value of each real-time performance parameter according to the local density and the relative local density of each real-time performance parameter;
comparing LOF values of the real-time performance parameters with preset LOF values, marking the real-time performance parameters with LOF values larger than the preset LOF values in the clustering abnormal parameter subsets as singular parameters, and eliminating the singular parameters in the corresponding clustering abnormal parameter subsets;
The singular parameters are imported into the remaining real-time performance parameter subsets, after the abnormal parameters are imported into the remaining real-time performance parameter subsets, the Davies-Bouldin indexes of the corresponding real-time performance parameter subsets are recalculated, and the recalculated Davies-Bouldin indexes are compared with a preset threshold value;
if the recalculated Davies-Bouldin indexes are all larger than a preset threshold, the singular parameters are thoroughly deleted;
if the recalculated Davies-Bouldin index is not greater than a preset threshold value, sequencing the recalculated Davies-Bouldin index based on numerical values to extract the minimum Davies-Bouldin index, and clustering the singular parameters into a real-time performance parameter subset corresponding to the minimum Davies-Bouldin index;
repeating the steps until all the clustering abnormal parameter subsets are verified and corrected, and updating all the initial real-time performance parameter subsets to obtain a plurality of final real-time performance parameter subsets.
It should be noted that, after the clustering abnormal parameter subset is selected, the LOF value of each real-time performance parameter in the clustering abnormal parameter subset is calculated by the LOF algorithm (local abnormal factor algorithm), and the real-time performance parameter with the LOF value greater than the preset LOF value is marked as a singular parameter, that is, the parameter is the clustering abnormal parameter, and the singular parameter is removed from the corresponding clustering abnormal parameter subset. Meanwhile, the singular parameters are imported into the remaining real-time performance parameter subsets, after the abnormal parameters are imported into the remaining real-time performance parameter subsets, the Davies-Bouldin indexes of the corresponding real-time performance parameter subsets are recalculated, if the recalculated Davies-Bouldin indexes are all larger than a preset threshold value, the singular parameters are not included in any subset, and if the parameters are invalid parameters, the singular parameters are thoroughly deleted; if the recalculated Davies-Bouldin index is not greater than the preset threshold, the singular parameters are clustered into a subset of real-time performance parameters corresponding to the minimum Davies-Bouldin index. The clustering effect of each initial real-time performance parameter subset can be verified and corrected through the step, so that a final real-time performance parameter subset with high reliability and good data quality is obtained, and the reliability of an evaluation result is effectively improved.
As shown in fig. 3, specifically, the final real-time performance parameter subset is compared with the corresponding standard performance parameter subset to determine the standard condition of each performance index in the dry powder extinguishing device to be evaluated, and the performance grade of the dry powder extinguishing device is evaluated according to the standard condition of each performance index, specifically:
s302: acquiring real-time performance parameters in each final real-time performance parameter subset, calculating the coincidence degree between the real-time performance parameters in each final real-time performance parameter subset and standard performance parameters in the corresponding standard performance parameter subset through a cosine similarity algorithm, and comparing the coincidence degree with a preset coincidence degree;
s304: if the overlap ratio is not greater than the preset overlap ratio, marking the performance index corresponding to the final real-time performance parameter subset as a substandard index; if the overlap ratio is larger than the preset overlap ratio, marking the performance index corresponding to the final real-time performance parameter subset as a standard reaching index;
s306: the method comprises the steps of constructing an information base according to the corresponding performance grades of all performance indexes of the prefabricated dry powder extinguishing equipment under various standard-reaching or non-standard-reaching combined conditions, and importing the corresponding performance grades of all the performance indexes of the dry powder extinguishing equipment under various standard-reaching or non-standard-reaching combined conditions into the information base;
S308: counting the up-to-standard or non-up-to-standard conditions of all the performance indexes in the current dry powder extinguishing equipment to be evaluated, guiding the up-to-standard or non-up-to-standard conditions of all the performance indexes in the current dry powder extinguishing equipment to be evaluated into the information base for pairing evaluation to obtain the performance grade of the current dry powder extinguishing equipment to be evaluated, and conveying the performance grade of the current dry powder extinguishing equipment to be evaluated to a preset platform for display.
The performance indexes include, but are not limited to, fire extinguishing speed, spraying distance, spraying time, powder consumption, etc. If the overlap ratio is not greater than the preset overlap ratio, indicating that the difference between the real-time performance parameters in the final real-time performance parameter subset and the standard performance parameters in the corresponding standard performance parameter subset is large, marking the performance index corresponding to the final real-time performance parameter subset as a non-standard index; and if the coincidence degree is larger than the preset coincidence degree, indicating that the difference between the real-time performance parameters in the final real-time performance parameter subset and the standard performance parameters in the corresponding standard performance parameter subset is smaller or completely coincident, marking the performance index corresponding to the final real-time performance parameter subset as a standard index. And then matching the performance grade of the current dry powder extinguishing equipment to be evaluated according to the condition that each performance index in the current dry powder extinguishing equipment to be evaluated meets or does not meet the standard.
In summary, the method can obtain the performance parameters with high accuracy and high integrity, can evaluate the performance of the dry powder extinguishing equipment more accurately, has high data processing analysis efficiency, can save algorithm operation time, and effectively improves evaluation efficiency.
In addition, the image information of the dry powder extinguishing equipment to be evaluated is obtained, and an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated is constructed according to the image information, and the method specifically comprises the following steps:
performing feature extraction processing on the image information to obtain a plurality of boundary points, and calculating the isolated score value of each boundary point based on an isolated forest algorithm;
screening out boundary points with the isolated score value larger than a preset isolated score value to obtain evacuation boundary points; selecting one evacuation boundary as a reference point, establishing a space three-dimensional coordinate system according to the reference point, and acquiring three-dimensional coordinate values of each evacuation boundary point in the space three-dimensional coordinate system;
calculating Chebyshev distances among the evacuation boundary points according to the three-dimensional coordinate values, searching out nearest neighbors of the evacuation boundary points according to the Chebyshev distances among the evacuation boundary points, and pairing the evacuation boundary points with the nearest neighbors to obtain a plurality of pairs of boundary point pairs;
Carrying out connection processing on each boundary point pair in the space three-dimensional coordinate system to obtain a plurality of line segments, and carrying out discrete processing on each line segment based on a discretization method to obtain a plurality of new boundary points;
generating dense boundary points according to the new boundary points and the sparse boundary points, acquiring relative three-dimensional coordinate values of all dense boundary points in the space three-dimensional coordinate system, and importing the relative three-dimensional coordinate values into modeling software for model construction to obtain a device three-dimensional model diagram of dry powder extinguishing device to be evaluated.
It should be noted that, feature extraction processing may be performed on the image information through an ORB algorithm to obtain boundary points of the dry powder fire extinguishing device in the image, and since a part of outliers may be extracted during feature extraction, these outliers are screened out through an isolated forest algorithm, and through performing connection processing on each boundary point pair, a plurality of line segments are obtained, and based on a discretization method, discrete processing is performed on each line segment, so as to obtain a plurality of new boundary points, and obtain a greater number of boundary points, thereby further making up for the algorithm deficiency. By the method, the three-dimensional model diagram with higher precision and better smoothness can be generated, the reliability of the model pairing result is improved, the evaluation precision is further improved, and the pairing error phenomenon is avoided.
Furthermore, the method comprises the following steps:
acquiring performance requirement information of dry powder extinguishing equipment to be evaluated, constructing a hierarchical evaluation system, importing the performance requirement information into a target layer of the hierarchical evaluation system, importing a standard performance parameter subset into a standard layer of the hierarchical evaluation system, and importing the final real-time performance parameter subset into a scheme layer of the hierarchical evaluation system to obtain a performance evaluation system;
comparing the performance parameters of the criterion layer and the scheme layer to judge the similarity of the final real-time performance parameter subset and the standard performance parameter subset, so as to obtain a judgment matrix; processing the judgment matrix through a singular value decomposition algorithm to obtain a feature vector;
obtaining the maximum eigenvalue of the eigenvector in the judgment matrix, and generating performance evaluation weight vector information according to the maximum eigenvalue;
generating state transition probability distribution according to the performance evaluation weight vector information, introducing a Markov model, carrying out simulation deduction on the state transition probability distribution through the Markov model to obtain a state transition probability matrix, and obtaining a time node of the dry powder extinguishing equipment reaching a preset performance stage according to the state transition probability matrix.
It should be noted that, the similarity between the final real-time performance parameter subset and the standard performance parameter subset of the dry powder fire extinguishing device is researched and judged by an analytic hierarchy process, and the state transition probability of the dry powder fire extinguishing device is deduced by a Markov model, so that the time node of the dry powder fire extinguishing device reaching the preset performance stage is estimated, the failure time of the device is accurately determined, and the reliability is improved.
As shown in fig. 4, another aspect of the present invention discloses a system for evaluating performance of a dry powder extinguishing device based on data analysis, the system for evaluating performance of a dry powder extinguishing device includes a memory 11 and a processor 22, the memory 11 stores a program for evaluating performance of a dry powder extinguishing device, and when the program for evaluating performance of a dry powder extinguishing device is executed by the processor 22, the following steps are implemented:
acquiring an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, performing pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and constructing a plurality of standard performance parameter subsets according to the standard performance parameters;
detecting dry powder extinguishing equipment to be evaluated, acquiring real-time performance parameters fed back by sensors in the dry powder extinguishing equipment at a plurality of preset time nodes in the detection process, constructing a database, and inputting the real-time performance parameters into the database;
After the acquisition is finished, clustering the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; wherein, each initial real-time performance parameter subset stores the real-time performance parameters of the same type;
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets;
and comparing and analyzing the final real-time performance parameter subset with the corresponding standard performance parameter subset, judging the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above described device embodiments are only illustrative, e.g. the division of the units is only one logical function division, and there may be other divisions in practice, such as: multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the various components shown or discussed may be coupled or directly coupled or communicatively coupled to each other via some interface, whether indirectly coupled or communicatively coupled to devices or units, whether electrically, mechanically, or otherwise.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units; can be located in one place or distributed to a plurality of network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present invention may be integrated in one processing unit, or each unit may be separately used as one unit, or two or more units may be integrated in one unit; the integrated units may be implemented in hardware or in hardware plus software functional units.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk or an optical disk, or the like, which can store program codes.
Alternatively, the above-described integrated units of the present invention may be stored in a computer-readable storage medium if implemented in the form of software functional modules and sold or used as separate products. Based on such understanding, the technical solutions of the embodiments of the present invention may be embodied in essence or a part contributing to the prior art in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, ROM, RAM, magnetic or optical disk, or other medium capable of storing program code.
The foregoing is merely illustrative embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think about variations or substitutions within the technical scope of the present invention, and the invention should be covered. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (6)
1. The method for evaluating the performance of the dry powder extinguishing equipment based on the data analysis is characterized by comprising the following steps of:
acquiring an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, performing pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and constructing a plurality of standard performance parameter subsets according to the standard performance parameters;
detecting dry powder extinguishing equipment to be evaluated, acquiring real-time performance parameters fed back by sensors in the dry powder extinguishing equipment at a plurality of preset time nodes in the detection process, constructing a database, and inputting the real-time performance parameters into the database;
after the acquisition is finished, clustering the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; wherein, each initial real-time performance parameter subset stores the real-time performance parameters of the same type;
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets;
Comparing and analyzing the final real-time performance parameter subset with a corresponding standard performance parameter subset, judging the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index;
the Davies-Bouldin index of each initial real-time performance parameter subset is calculated according to the coordinate values, and specifically comprises the following steps:
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset according to the coordinate values, and carrying out average value taking treatment on the Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset to obtain intra-cluster compactness;
calculating Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets according to the coordinate values, and carrying out average value taking processing on the Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets to obtain the cluster separation degree;
calculating the Davies-Bouldin indexes of each initial real-time performance parameter subset according to the intra-cluster compactness and the inter-cluster separation degree, and comparing the Davies-Bouldin indexes of each initial real-time performance parameter subset with a preset threshold;
Marking an initial real-time performance parameter subset with the Davies-Bouldin index larger than a preset threshold value as a clustering abnormal parameter subset; marking an initial real-time performance parameter subset with the Davies-Bouldin index not larger than a preset threshold value as a clustering normal parameter subset;
the clustering effect of each initial real-time performance parameter subset is verified and corrected according to the Davies-Bouldin index, and a plurality of final real-time performance parameter subsets are obtained, specifically:
acquiring real-time performance parameters in the clustering abnormal parameter subset, calculating the local density and the relative local density of each real-time performance parameter through an LOF algorithm, and calculating the LOF value of each real-time performance parameter according to the local density and the relative local density of each real-time performance parameter;
comparing LOF values of the real-time performance parameters with preset LOF values, marking the real-time performance parameters with LOF values larger than the preset LOF values in the clustering abnormal parameter subsets as singular parameters, and eliminating the singular parameters in the corresponding clustering abnormal parameter subsets;
the singular parameters are imported into the remaining real-time performance parameter subsets, after the abnormal parameters are imported into the remaining real-time performance parameter subsets, the Davies-Bouldin indexes of the corresponding real-time performance parameter subsets are recalculated, and the recalculated Davies-Bouldin indexes are compared with a preset threshold value;
If the recalculated Davies-Bouldin indexes are all larger than a preset threshold, the singular parameters are thoroughly deleted;
if the recalculated Davies-Bouldin index is not greater than a preset threshold value, sequencing the recalculated Davies-Bouldin index based on numerical values to extract the minimum Davies-Bouldin index, and clustering the singular parameters into a real-time performance parameter subset corresponding to the minimum Davies-Bouldin index;
repeating the steps until all the clustering abnormal parameter subsets are verified and corrected, and updating all the initial real-time performance parameter subsets to obtain a plurality of final real-time performance parameter subsets.
2. The method for evaluating the performance of the dry powder extinguishing equipment based on the data analysis according to claim 1, wherein the method is characterized by obtaining an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, and carrying out pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and specifically comprises the following steps:
prefabricating standard three-dimensional model diagrams of various types of dry powder fire extinguishing equipment and prefabricating standard performance parameters of the various types of dry powder fire extinguishing equipment;
binding standard three-dimensional model diagrams of various types of dry powder fire extinguishing equipment with corresponding standard performance parameters to obtain a plurality of standard data packets; constructing a knowledge graph, and importing a plurality of standard data packets into the knowledge graph;
Acquiring image information of dry powder extinguishing equipment to be evaluated, and constructing an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated according to the image information;
searching standard three-dimensional model diagrams of various types of dry powder extinguishing equipment from the knowledge graph, and calculating the similarity between the equipment three-dimensional model diagrams and each standard three-dimensional model diagram through a Hausdorff distance algorithm to obtain a plurality of similarities;
and selecting the maximum similarity from the plurality of similarities, acquiring a standard three-dimensional model diagram corresponding to the maximum similarity, acquiring a standard data packet of the standard three-dimensional model diagram corresponding to the maximum similarity, and acquiring standard performance parameters of the dry powder extinguishing equipment to be evaluated from the standard data packet.
3. The method for evaluating the performance of dry powder extinguishing equipment based on data analysis according to claim 1, wherein the clustering processing is performed on the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets, specifically:
acquiring the number of performance parameters to be evaluated in dry powder extinguishing equipment to be evaluated, determining the number of Gaussian components according to the number of the performance parameters to be evaluated, and determining a plurality of cluster families according to the number of the Gaussian components;
Randomly selecting a real-time performance parameter from the database as an initial average value of the Gaussian component; calculating covariance among all real-time performance parameters in a database, and generating an initial covariance matrix of a Gaussian component according to the covariance; randomly generating a positive number for each cluster group, and carrying out normalization processing on the positive number to obtain the initial weight of the Gaussian component;
determining posterior probability of each piece of real-time performance data according to the initial mean value, the initial covariance matrix and the initial weight of the Gaussian components, and re-determining a new mean value, a new covariance matrix and a new weight of the Gaussian components according to the posterior probability of each piece of real-time performance data;
determining the log-likelihood value of each real-time performance parameter according to the new mean value, the new covariance matrix and the new weight, summing all the log-likelihood values to obtain a total log-likelihood value, and determining the fitting degree of the Gaussian mixture model according to the total log-likelihood value;
if the fitting degree is larger than the preset fitting degree, saving model parameters to obtain a Gaussian mixture model; if the fitting degree is not greater than the preset fitting degree, continuing iteration until the fitting degree is greater than the preset fitting degree, and storing model parameters to obtain a Gaussian mixture model;
And acquiring posterior probability of each real-time performance parameter belonging to each cluster family again in the Gaussian mixture model, clustering each real-time performance parameter into the cluster family with the largest posterior probability, and obtaining a plurality of initial real-time performance parameter subsets after clustering.
4. The method for evaluating the performance of a dry powder extinguishing device based on data analysis according to claim 1, wherein the coordinate values of the real-time performance parameters in the low-dimensional space in each initial real-time performance parameter subset are obtained, specifically:
extracting corresponding real-time performance parameters from each initial real-time performance parameter subset, measuring the similarity between the real-time performance parameters in each initial real-time performance parameter subset through a polynomial kernel function, and obtaining a similarity matrix of each initial real-time performance parameter subset based on the similarity between the real-time performance parameters of each initial real-time performance parameter subset;
acquiring the relative distance of the real-time performance parameters in the high-dimensional space in each initial real-time performance parameter subset, and calculating the conditional probability distribution of each real-time performance parameter in the high-dimensional space based on the relative distance and the corresponding similarity matrix;
Randomly initializing the relative positions of the real-time performance parameters in the low-dimensional space in each initial real-time performance parameter subset, and calculating according to the relative positions and the corresponding similarity matrix to obtain the conditional probability distribution of each real-time performance parameter in the low-dimensional space;
calculating KL divergence between the high-dimensional real-time performance parameters and the low-dimensional real-time performance parameters according to the conditional probability distribution of each real-time performance parameter in the high-dimensional space and the conditional probability distribution of each real-time performance parameter in the low-dimensional space;
the KL divergence between the high-dimensional real-time performance parameter and the low-dimensional real-time performance parameter is minimized through a gradient descent optimization algorithm, so that the position of the real-time performance parameter in the low-dimensional space is adjusted; and repeating the steps until convergence conditions are reached, determining the position of each real-time performance parameter in the low-dimensional space, and acquiring the coordinate value of each real-time performance parameter in the low-dimensional space.
5. The method for evaluating the performance of a dry powder extinguishing device based on data analysis according to claim 1, wherein the final real-time performance parameter subset is compared with the corresponding standard performance parameter subset to judge the standard condition of each performance index in the dry powder extinguishing device to be evaluated, and the performance grade of the dry powder extinguishing device is evaluated according to the standard condition of each performance index, specifically:
Acquiring real-time performance parameters in each final real-time performance parameter subset, calculating the coincidence degree between the real-time performance parameters in each final real-time performance parameter subset and standard performance parameters in the corresponding standard performance parameter subset through a cosine similarity algorithm, and comparing the coincidence degree with a preset coincidence degree;
if the overlap ratio is not greater than the preset overlap ratio, marking the performance index corresponding to the final real-time performance parameter subset as a substandard index; if the overlap ratio is larger than the preset overlap ratio, marking the performance index corresponding to the final real-time performance parameter subset as a standard reaching index;
the method comprises the steps of constructing an information base according to the corresponding performance grades of all performance indexes of the prefabricated dry powder extinguishing equipment under various standard-reaching or non-standard-reaching combined conditions, and importing the corresponding performance grades of all the performance indexes of the dry powder extinguishing equipment under various standard-reaching or non-standard-reaching combined conditions into the information base;
counting the up-to-standard or non-up-to-standard conditions of all the performance indexes in the current dry powder extinguishing equipment to be evaluated, guiding the up-to-standard or non-up-to-standard conditions of all the performance indexes in the current dry powder extinguishing equipment to be evaluated into the information base for pairing evaluation to obtain the performance grade of the current dry powder extinguishing equipment to be evaluated, and conveying the performance grade of the current dry powder extinguishing equipment to be evaluated to a preset platform for display.
6. The system for evaluating the performance of the dry powder extinguishing equipment based on data analysis is characterized by comprising a memory and a processor, wherein the memory stores a dry powder extinguishing equipment performance evaluation method program, and when the dry powder extinguishing equipment performance evaluation method program is executed by the processor, the following steps are realized:
acquiring an equipment three-dimensional model diagram of the dry powder extinguishing equipment to be evaluated, performing pairing identification on the equipment three-dimensional model diagram to obtain standard performance parameters of the dry powder extinguishing equipment to be evaluated, and constructing a plurality of standard performance parameter subsets according to the standard performance parameters;
detecting dry powder extinguishing equipment to be evaluated, acquiring real-time performance parameters fed back by sensors in the dry powder extinguishing equipment at a plurality of preset time nodes in the detection process, constructing a database, and inputting the real-time performance parameters into the database;
after the acquisition is finished, clustering the real-time performance parameters in the database to obtain a plurality of initial real-time performance parameter subsets; wherein, each initial real-time performance parameter subset stores the real-time performance parameters of the same type;
Acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Davies-Bouldin indexes of each initial real-time performance parameter subset according to the coordinate values, verifying and correcting clustering effects of each initial real-time performance parameter subset according to the Davies-Bouldin indexes, and obtaining a plurality of final real-time performance parameter subsets;
comparing and analyzing the final real-time performance parameter subset with a corresponding standard performance parameter subset, judging the standard condition of each performance index in the dry powder extinguishing equipment to be evaluated, and evaluating the performance grade of the dry powder extinguishing equipment according to the standard condition of each performance index;
the Davies-Bouldin index of each initial real-time performance parameter subset is calculated according to the coordinate values, and specifically comprises the following steps:
acquiring coordinate values of real-time performance parameters in a low-dimensional space in each initial real-time performance parameter subset, calculating Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset according to the coordinate values, and carrying out average value taking treatment on the Euclidean distances among the real-time performance parameters in the same initial real-time performance parameter subset to obtain intra-cluster compactness;
Calculating Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets according to the coordinate values, and carrying out average value taking processing on the Euclidean distances among all the real-time performance parameters among the different initial real-time performance parameter subsets to obtain the cluster separation degree;
calculating the Davies-Bouldin indexes of each initial real-time performance parameter subset according to the intra-cluster compactness and the inter-cluster separation degree, and comparing the Davies-Bouldin indexes of each initial real-time performance parameter subset with a preset threshold;
marking an initial real-time performance parameter subset with the Davies-Bouldin index larger than a preset threshold value as a clustering abnormal parameter subset; marking an initial real-time performance parameter subset with the Davies-Bouldin index not larger than a preset threshold value as a clustering normal parameter subset;
the clustering effect of each initial real-time performance parameter subset is verified and corrected according to the Davies-Bouldin index, and a plurality of final real-time performance parameter subsets are obtained, specifically:
acquiring real-time performance parameters in the clustering abnormal parameter subset, calculating the local density and the relative local density of each real-time performance parameter through an LOF algorithm, and calculating the LOF value of each real-time performance parameter according to the local density and the relative local density of each real-time performance parameter;
Comparing LOF values of the real-time performance parameters with preset LOF values, marking the real-time performance parameters with LOF values larger than the preset LOF values in the clustering abnormal parameter subsets as singular parameters, and eliminating the singular parameters in the corresponding clustering abnormal parameter subsets;
the singular parameters are imported into the remaining real-time performance parameter subsets, after the abnormal parameters are imported into the remaining real-time performance parameter subsets, the Davies-Bouldin indexes of the corresponding real-time performance parameter subsets are recalculated, and the recalculated Davies-Bouldin indexes are compared with a preset threshold value;
if the recalculated Davies-Bouldin indexes are all larger than a preset threshold, the singular parameters are thoroughly deleted;
if the recalculated Davies-Bouldin index is not greater than a preset threshold value, sequencing the recalculated Davies-Bouldin index based on numerical values to extract the minimum Davies-Bouldin index, and clustering the singular parameters into a real-time performance parameter subset corresponding to the minimum Davies-Bouldin index;
repeating the steps until all the clustering abnormal parameter subsets are verified and corrected, and updating all the initial real-time performance parameter subsets to obtain a plurality of final real-time performance parameter subsets.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311510615.4A CN117236572B (en) | 2023-11-14 | 2023-11-14 | Method and system for evaluating performance of dry powder extinguishing equipment based on data analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311510615.4A CN117236572B (en) | 2023-11-14 | 2023-11-14 | Method and system for evaluating performance of dry powder extinguishing equipment based on data analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117236572A CN117236572A (en) | 2023-12-15 |
CN117236572B true CN117236572B (en) | 2024-03-29 |
Family
ID=89084627
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311510615.4A Active CN117236572B (en) | 2023-11-14 | 2023-11-14 | Method and system for evaluating performance of dry powder extinguishing equipment based on data analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117236572B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106599601A (en) * | 2016-12-29 | 2017-04-26 | 中国科学院遥感与数字地球研究所 | Remote sensing assessment method and system for ecosystem vulnerability |
CN113408576A (en) * | 2021-05-12 | 2021-09-17 | 上海师范大学 | Learning style identification method based on fusion label and stacked machine learning model |
CN114778684A (en) * | 2022-06-21 | 2022-07-22 | 张家港沙龙精密管业有限公司 | Steel pipe performance evaluation method and system based on service scene |
CN115454990A (en) * | 2022-09-30 | 2022-12-09 | 国网福建省电力有限公司莆田供电公司 | Oil paper insulation data cleaning method based on improved KNN |
CN115822887A (en) * | 2022-12-05 | 2023-03-21 | 中国华能集团清洁能源技术研究院有限公司 | Performance evaluation and energy efficiency diagnosis method and system of wind turbine generator |
CN116776631A (en) * | 2023-07-05 | 2023-09-19 | 深圳市精微康投资发展有限公司 | Connector performance evaluation method and system based on data analysis |
CN116862081A (en) * | 2023-09-05 | 2023-10-10 | 北京建工环境修复股份有限公司 | Operation and maintenance method and system for pollution treatment equipment |
CN116956639A (en) * | 2023-09-15 | 2023-10-27 | 深圳海汇科技有限公司 | High-heat-dissipation storage battery assessment method and system based on data analysis |
-
2023
- 2023-11-14 CN CN202311510615.4A patent/CN117236572B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106599601A (en) * | 2016-12-29 | 2017-04-26 | 中国科学院遥感与数字地球研究所 | Remote sensing assessment method and system for ecosystem vulnerability |
CN113408576A (en) * | 2021-05-12 | 2021-09-17 | 上海师范大学 | Learning style identification method based on fusion label and stacked machine learning model |
CN114778684A (en) * | 2022-06-21 | 2022-07-22 | 张家港沙龙精密管业有限公司 | Steel pipe performance evaluation method and system based on service scene |
CN115454990A (en) * | 2022-09-30 | 2022-12-09 | 国网福建省电力有限公司莆田供电公司 | Oil paper insulation data cleaning method based on improved KNN |
CN115822887A (en) * | 2022-12-05 | 2023-03-21 | 中国华能集团清洁能源技术研究院有限公司 | Performance evaluation and energy efficiency diagnosis method and system of wind turbine generator |
CN116776631A (en) * | 2023-07-05 | 2023-09-19 | 深圳市精微康投资发展有限公司 | Connector performance evaluation method and system based on data analysis |
CN116862081A (en) * | 2023-09-05 | 2023-10-10 | 北京建工环境修复股份有限公司 | Operation and maintenance method and system for pollution treatment equipment |
CN116956639A (en) * | 2023-09-15 | 2023-10-27 | 深圳海汇科技有限公司 | High-heat-dissipation storage battery assessment method and system based on data analysis |
Also Published As
Publication number | Publication date |
---|---|
CN117236572A (en) | 2023-12-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111738351B (en) | Model training method and device, storage medium and electronic equipment | |
CN111353549B (en) | Image label verification method and device, electronic equipment and storage medium | |
CN110928862A (en) | Data cleaning method, data cleaning apparatus, and computer storage medium | |
CN108491302A (en) | A method of detection spark clustered node states | |
CN111338972A (en) | Machine learning-based software defect and complexity incidence relation analysis method | |
CN111581092A (en) | Method for generating simulation test data, computer device and storage medium | |
CN114760098A (en) | CNN-GRU-based power grid false data injection detection method and device | |
CN113918367A (en) | Large-scale system log anomaly detection method based on attention mechanism | |
CN114048468A (en) | Intrusion detection method, intrusion detection model training method, device and medium | |
CN116756594A (en) | Method, system, equipment and medium for detecting abnormal points of power grid data | |
CN115021679A (en) | Photovoltaic equipment fault detection method based on multi-dimensional outlier detection | |
CN115758183A (en) | Training method and device for log anomaly detection model | |
CN114020811A (en) | Data anomaly detection method and device and electronic equipment | |
CN117236572B (en) | Method and system for evaluating performance of dry powder extinguishing equipment based on data analysis | |
CN113822336A (en) | Cloud hard disk fault prediction method, device and system and readable storage medium | |
CN116365519B (en) | Power load prediction method, system, storage medium and equipment | |
CN112257332A (en) | Simulation model evaluation method and device | |
CN113255810B (en) | Network model testing method based on key decision logic design test coverage rate | |
CN112836570B (en) | Equipment abnormality detection method utilizing Gaussian noise | |
CN112465073A (en) | Numerical value distribution anomaly detection method and system based on distance | |
CN114334696A (en) | Quality detection method and device, electronic equipment and computer readable storage medium | |
CN112101468A (en) | Method for judging abnormal sequence in sequence combination | |
CN117687290A (en) | Stopwatch detection and evaluation method and system based on multi-source data | |
CN117668528B (en) | Natural gas voltage regulator fault detection method and system based on Internet of things | |
CN112685324B (en) | Method and system for generating test scheme |
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 |