CN117725398A - Equipment operation information maintenance method and system based on multiple heterogeneous data - Google Patents
Equipment operation information maintenance method and system based on multiple heterogeneous data Download PDFInfo
- Publication number
- CN117725398A CN117725398A CN202311719770.7A CN202311719770A CN117725398A CN 117725398 A CN117725398 A CN 117725398A CN 202311719770 A CN202311719770 A CN 202311719770A CN 117725398 A CN117725398 A CN 117725398A
- Authority
- CN
- China
- Prior art keywords
- data
- representing
- parameter
- class
- operation information
- 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.)
- Pending
Links
- 238000012423 maintenance Methods 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000002159 abnormal effect Effects 0.000 claims abstract description 18
- 230000004927 fusion Effects 0.000 claims abstract description 17
- 239000013598 vector Substances 0.000 claims abstract description 12
- 238000007477 logistic regression Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 9
- 239000002245 particle Substances 0.000 claims description 47
- 238000004891 communication Methods 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000005856 abnormality Effects 0.000 claims description 11
- 230000001133 acceleration Effects 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 4
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 claims 4
- 238000009825 accumulation Methods 0.000 claims 1
- 238000001514 detection method Methods 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 description 6
- 230000009286 beneficial effect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for overhauling equipment operation information based on multi-element heterogeneous data, which belong to the technical field of data processing, wherein the method comprises the following steps: a unified command interface is constructed for a plurality of different manufacturers and types of equipment; acquiring multi-element heterogeneous operation data of equipment to be detected through a unified command interface; according to the fluctuation parameter, the continuity parameter, the isochronal parameter and the relevance parameter of the operation data, the effectiveness of the operation data is evaluated, the operation data is reserved when the operation data is effective, and the operation data is removed when the operation data is ineffective; carrying out standardization processing on the reserved operation data; extracting data characteristics of the standardized operation data; feature fusion is carried out on the data features to obtain fusion feature vectors; and judging whether the equipment operation information is abnormal or not through an equipment operation information maintenance model based on logistic regression according to the fusion feature vector.
Description
Technical Field
The invention belongs to the technical field of data processing, and particularly relates to a device operation information overhauling method and system based on multi-element heterogeneous data.
Background
The equipment operation state is overhauled regularly, so that the reliability, availability and safety of the equipment can be improved, equipment faults and downtime are reduced, the production efficiency is improved, the operation cost is reduced, and the equipment can be ensured to stably operate according to the design requirement.
However, there is a problem that there is often incompatibility between different manufacturers and different types of devices at present, and an overhaul scheme is often required to be customized for each device separately, and when the device needs to be updated or a new device needs to be introduced, due to lack of versatility and standardized overhaul scheme, a maintenance policy needs to be formulated and adjusted again, which increases management complexity. And during maintenance, the manual maintenance is mainly carried out by relying on engineers with abundant experience, the maintenance period is longer, the maintenance period is easily influenced by subjective factors such as the experience of practise, the professional level and the like, human errors can be introduced, and the equipment maintenance accuracy is lower.
Disclosure of Invention
In order to solve the problem that incompatibility often exists between different manufacturers and different types of equipment at present, an overhaul scheme is often required to be independently customized for each equipment, and when the equipment is required to be updated or new equipment is required to be introduced, maintenance strategies are required to be re-formulated and adjusted due to lack of generality and standardized overhaul schemes, so that the complexity of management is increased. And during maintenance, manual maintenance is mainly carried out by relying on engineers with abundant experience, the maintenance period is long, the maintenance period is easily influenced by subjective factors such as practitioner experience, professional level and the like, and human errors can be introduced, so that the technical problem of low equipment maintenance accuracy is solved.
First aspect
The invention provides a device operation information overhauling method based on multi-element heterogeneous data, which comprises the following steps:
s1: a unified command interface is constructed for a plurality of different manufacturers and types of equipment;
s2: acquiring multi-element heterogeneous operation data of equipment to be detected through the unified command interface;
s3: evaluating the validity of the operation data according to the fluctuation parameter, the continuity parameter, the isochronicity parameter and the relevance parameter of the operation data, reserving the operation data when the operation data is valid, and removing the operation data when the operation data is invalid;
s4: carrying out standardization processing on the reserved operation data;
s5: extracting data characteristics of the standardized operation data;
s6: performing feature fusion on the data features to obtain fusion feature vectors;
s7: and judging whether the equipment operation information is abnormal or not through an equipment operation information maintenance model based on logistic regression according to the fusion feature vector.
Second aspect
The invention provides an equipment operation information maintenance system based on multi-element heterogeneous data, which comprises a processor and a memory for storing executable instructions of the processor, wherein the processor is used for storing executable instructions of the processor; the processor is configured to invoke the instructions stored by the memory to perform the equipment operation information overhaul method based on the multivariate heterogeneous data in the first aspect.
Compared with the prior art, the invention has at least the following beneficial technical effects:
(1) In the invention, a unified command interface is constructed for various devices of different manufacturers and types, and the multiple heterogeneous operation data of the device to be detected is obtained through the unified command interface, so that the device is overhauled according to the operation data, and when the device is required to be updated or new devices are introduced, maintenance strategies are not required to be formulated and adjusted again, thereby reducing the complexity of management.
(2) In the invention, the operation data are evaluated according to the fluctuation, continuity, isochrony and relevance parameters, the operation data are reserved when valid, and the operation data are removed when invalid, so that misjudgment caused by invalid or abnormal data is reduced, and the accurate monitoring of the operation state of the equipment is enhanced.
(3) According to the invention, through the equipment operation information maintenance model based on logistic regression, whether the equipment operation information is abnormal or not is automatically judged, so that automatic equipment maintenance is realized, manual intervention is not needed, the influence of subjective factors such as practitioner experience, professional level and the like is avoided, the maintenance period is reduced, and the accuracy and consistency of equipment maintenance are improved.
Drawings
The above features, technical features, advantages and implementation of the present invention will be further described in the following description of preferred embodiments in a clear and easily understood manner with reference to the accompanying drawings.
Fig. 1 is a schematic flow chart of a method for overhauling equipment operation information based on multi-element heterogeneous data.
Fig. 2 is a schematic structural diagram of an equipment operation information maintenance system based on multi-element heterogeneous data.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will explain the specific embodiments of the present invention with reference to the accompanying drawings. It is evident that the drawings in the following description are only examples of the invention, from which other drawings and other embodiments can be obtained by a person skilled in the art without inventive effort.
Example 1
In one embodiment, referring to fig. 1 of the specification, a flow diagram of a method for overhauling equipment operation information based on multiple heterogeneous data provided by the invention is shown.
The invention provides a device operation information overhauling method based on multi-element heterogeneous data, which comprises the following steps:
s1: a unified command interface is built for a variety of different vendors and types of devices.
The unified command interface is a standardized and universal interface, and aims to provide consistent command and control modes for various different manufacturers and types of equipment. The unified command interface can solve the problem of communication and control heterogeneity existing between different devices, so that the devices can communicate, control and exchange data through the same interface.
In one possible embodiment, S1 specifically includes substeps S101 to S108:
s101: a generic device command set is defined.
Wherein the device command set is a standardized set of commands for controlling and communicating with the device. These commands typically include various operations that the device may perform, such as start, stop, restart, acquire status, and so forth. The definition of the device command set is to realize the universality and consistency among devices, and ensure that various devices can execute commands according to the same rule.
S102: a common communication protocol is selected.
Wherein the communication protocol is a rule and convention for communication between devices. The communication protocol is chosen to ensure that data exchange between the devices is performed correctly and reliably. Common communication protocols include HTTP, MQTT, coAP, and the like. The selection of the appropriate protocol is typically based on communication requirements, performance requirements, and interoperability considerations between devices.
S103: a common data format is adopted.
Wherein the data format defines a data structure and a representation method used in the communication. A common data format is typically employed, such as JSON (JavaScript Object Notation) or XML (eXtensible Markup Language). This helps ensure that the same data format is shared between different devices, simplifying the complexity of data parsing and processing.
It should be noted that, by defining a general device command set, selecting a general communication protocol, and adopting a general data format, the universality and consistency between devices are realized. Regardless of the manufacturer or model of the device, they follow the same communication rules and data formats, making integration and co-operation easier.
S104: description and metadata are added for each different vendor and type of device.
Where the device description is text or structured information about the device, it generally includes detailed descriptions of the function, characteristics, purpose, specification, etc. of the device.
Metadata is data about data that describes attributes, features, and other information of the data. In device communication, metadata is used to describe data of the device, including sensor data, command formats, data units, and the like.
S105: the REST architecture network service is constructed using a generic communication protocol.
Where REST (Representational State Transfer) is a resource-based architectural style and RESTful API is an application programming interface designed based on this style. The RESTful API is constructed to enable devices to communicate via HTTP requests. The RESTful API uses standard HTTP methods (GET, POST, PUT, DELETE, etc.) to perform operations and represent data states in the form of resources.
It should be noted that, the REST architecture network service (RESTful API) is a lightweight communication manner, and supports flexible and stateless communication.
S106: a proxy service is introduced through which communication and protocol conversion are performed with the underlying device.
The proxy service is a middle layer introduced in the system and is responsible for communication with the bottom layer equipment and processing details such as protocol conversion. The introduction of proxy services helps to solve the problem of the underlying devices using different protocols, making the system more flexible. Proxy services may also provide caching, security, and monitoring functions.
It should be noted that the introduction of proxy services helps to solve the problem of using different protocols by the underlying devices, making the system more adaptive.
S107: using the configuration file, a mapping relationship between the device interface and the device commands is defined.
Wherein the configuration file is used for defining a mapping relation between the device interface and the device command. Through the configuration file, the system can dynamically adapt to the command structures of different devices. This makes the system more convenient when introducing new equipment, modifying equipment configuration or deleting equipment, has improved the maintainability of system.
S108: the adapter is configured for each different vendor and type of device to complete the construction of a unified command interface for a plurality of different vendors and types of devices.
Wherein the adapter is a component for implementing device interface adaptation and protocol conversion. The adapters are configured for each of the different vendors and types of devices to enable them to communicate and operate in accordance with a unified command interface. The adapter is responsible for translating generic commands into device specific commands and translating device responses into generic format. The configuration of the adapter makes the system more scalable and compatible.
It should be noted that, by introducing proxy service, defining mapping relation by using configuration file and configuring adapter for device, protocol conversion and device interface adaptation are realized. This enables the system to communicate with devices using different communication protocols, increasing the compatibility of the system with heterogeneous devices.
In the invention, a unified command interface is constructed for various devices of different manufacturers and types, and the multiple heterogeneous operation data of the device to be detected is obtained through the unified command interface, so that the device is overhauled according to the operation data, and when the device is required to be updated or new devices are introduced, maintenance strategies are not required to be formulated and adjusted again, thereby reducing the complexity of management.
S2: and acquiring the multi-element heterogeneous operation data of the equipment to be detected through the unified command interface.
Wherein the operation data includes: temperature data, humidity data, pressure data, vibration data, CPU usage data, memory usage data, network bandwidth usage data, voltage data, current data, load data, power data, network connection status data, battery status data, and the like.
S3: and evaluating the validity of the operation data according to the fluctuation parameter, the continuity parameter, the isochronicity parameter and the relevance parameter of the operation data, reserving when the operation data is valid, and removing when the operation data is invalid.
Wherein the fluctuation parameter refers to an index for measuring the fluctuation degree or instability of time series data.
The continuity parameter is an index for measuring the degree of data continuity or deletion in time-series data.
The isochrony parameter refers to an index for measuring time sequence synchronicity among different types of operation data.
Wherein the relevance parameter refers to an index for measuring similarity between different types of operation data.
In the invention, the operation data are evaluated according to the fluctuation, continuity, isochrony and relevance parameters, the operation data are reserved when valid, and the operation data are removed when invalid, so that misjudgment caused by invalid or abnormal data is reduced, and the accurate monitoring of the operation state of the equipment is enhanced.
In one possible embodiment, S3 specifically includes: substeps S301 to S303:
s301: and calculating validity parameters of various operation data according to the following formula:
wherein τ i A validity parameter representing i-th class operation data, A i A represents a volatility parameter, alpha, of class i operational data 1 Weight coefficient representing volatility parameter, B i Continuity parameter, alpha, representing class i operational data 2 Weight coefficient representing continuity parameter, T i Isochronic parameter, alpha, representing class i operational data 3 Weight coefficient representing isochronic parameter, C i Relevance parameter, alpha, representing class i operational data 4 And a weight coefficient representing the relevance parameter.
Wherein, the person skilled in the art can according toThe actual situation sets the weight coefficient alpha of the fluctuation parameter 1 Weight coefficient alpha of continuity parameter 2 Weight coefficient alpha of isochronic parameter 3 And a weight coefficient alpha of a relevance parameter 4 The size of (3) is not limited in the present invention.
The importance of different parameters can be flexibly adjusted by introducing the weight coefficient according to actual conditions. Different parameters may have different influence in different scenes, and by adjusting the weight coefficient, the effectiveness evaluation can be flexibly performed according to specific requirements.
In the invention, parameters of multiple dimensions such as volatility, continuity, isochrony, relevance and the like are considered, so that the evaluation of the effectiveness of the operation data is more comprehensive, the comprehensive consideration of the operation data characteristics in different aspects is facilitated, and the accuracy of the operation state of the equipment is improved.
S302: when the validity parameter of the operation data is larger than a preset validity value, the operation data is determined to be valid and reserved.
The size of the preset validity value can be set by a person skilled in the art according to actual conditions, and the invention is not limited.
S303: and when the validity parameter of the operation data is smaller than or equal to the preset validity value, determining that the operation data is invalid, and removing.
It should be noted that, introducing a preset validity value allows setting a standard according to specific requirements, when the validity parameter exceeds the standard, determining that the operation data is valid, otherwise, determining that the operation data is invalid. The method is favorable for carrying out the customized validity standard according to the requirements of specific application scenes, and improves the flexibility of the system.
According to the invention, by comprehensively considering a plurality of parameters and setting reasonable weights and preset values, the validity of the operation data can be evaluated more comprehensively and flexibly, and the method is beneficial to improving the accurate monitoring and judgment of the equipment operation information maintenance system on the equipment state.
In one possible implementation, the volatility parameter is calculated by:
calculating fluctuation values of all data points in the operation data:
a ij ＝|x ij -μ i |
wherein a is ij A fluctuation value, x, representing a j-th data point of i-th type operation data ij Values, μ representing the jth data point of the ith class of operational data i Representing the mean of class i operational data.
Counting the number of data points with a fluctuation value larger than a preset fluctuation value, wherein the preset fluctuation value is 3 sigma i ，σ i Representing the standard deviation of the class i operational data.
By calculating the fluctuation value and comparing the calculated fluctuation value with a preset fluctuation value, abnormal fluctuation in the equipment operation data can be rapidly detected.
Calculating fluctuation parameters according to the number of data points with fluctuation values larger than a preset fluctuation value:
wherein A is i A represents a volatility parameter of class i operation data, a i Representing the number of data points with fluctuation values larger than preset fluctuation values in the i-th class of operation data, S i Representing the total number of data points in the class i operational data.
In the invention, false alarms are reduced by considering the number of data points with the fluctuation value larger than the preset fluctuation value, not just the fluctuation value itself. The number of data points is considered, so that the fluctuation parameter has more practical significance, and the influence of short fluctuation on the whole fluctuation is eliminated.
In one possible implementation, the continuity parameter is calculated in the following manner:
and counting the number of missing data points in the i-th class of operation data.
It should be noted that, by counting the number of missing data points, the missing condition in the equipment operation data can be detected rapidly.
Calculating a continuity parameter according to the number of missing data points:
wherein B is i A continuity parameter representing class i operational data, b i Representing the number of missing data points in the i-th class of operation data, S i Representing the total number of data points in the class i operational data.
According to the invention, the number of missing data points is counted, so that the data missing condition in the running data of the equipment can be intuitively reflected, the problems in data acquisition or transmission can be found in time, and the integrity and reliability of the data are improved.
In one possible implementation, the isochronal parameter is calculated by:
calculating the time length difference rate between various operation data:
wherein t is ij Representing a time length difference rate between the i-th class operation data and the j-th class operation data, t i Representing the total duration of the i-th class of operation data, t j Representing the total duration of the j-th class of operation data, max () represents taking the maximum value.
It should be noted that, by calculating the time length difference rate between various kinds of operation data, the isochronal parameter can reflect the time synchronicity of different kinds of operation data, so as to help to know whether the operation data of different devices or systems have consistent time sequence characteristics.
Calculating isochronal parameters according to the time length difference rate among various operation data:
wherein T is i An isochronal parameter representing class i operational data.
According to the invention, the isochronal parameters are obtained by summarizing the time length difference rate among various operation data. The isochronic parameter can provide a global evaluation index, reflects the overall synchronicity level and facilitates the validity evaluation of the operation data. If the time duration of certain operational data differs too much from other operational data, this may be problematic.
In one possible implementation, the relevance parameter is calculated in the following manner:
calculating the distance between each data point in the two types of operation data, and constructing a distance matrix d ij ]。
Wherein d ij Representing the distance between the i-th data point and the j-th data point.
Starting from the upper left corner of the distance matrix, recording the minimum accumulated distance of each data point in the distance matrix, and forming a dynamic programming matrix [ D ] ij ]：
D ij ＝d ij +min[D i-1,j ,D i,j-1 ,D i-1,j-1 ]
Wherein D is ij Representing the minimum cumulative distance to the ith row, jth column data point in the distance matrix. min []Representing taking the minimum value.
Starting from the lower right corner, backtracking the dynamic programming matrix from the upper left corner, and determining the optimal regular path with the smallest accumulated distance.
And taking the accumulated distance of the optimal regular path as the association degree between the two types of operation data.
It should be noted that, selecting the optimal regular path with the smallest cumulative distance as the measure of the association degree can more accurately reflect the similarity of the two time series data on the time axis. Such a measure is more practical and helps to accurately evaluate the degree of correlation between data.
According to the degree of association between various types of operation data, calculating association parameters:
wherein C is i A relevance parameter representing class i operational data, c ij And representing the degree of association between the i-th class of operation data and the j-th class of operation data.
According to the invention, the flexibility and the change of the time sequence data can be fully considered, and the association relation between the equipment operation data can be more accurately described, so that the accuracy and the adaptability of the equipment operation information maintenance system are improved.
S4: and (5) carrying out standardization processing on the reserved operation data.
The normalization process may specifically be a normalization process. The dimensional differences between different features are eliminated, and better comparability and stability of the data are ensured when further analysis or modeling is performed.
In the invention, the normalization processing is performed on the reserved operation data, which is helpful to ensure the consistency among all the characteristics and improve the stability and performance of model training.
S5: and extracting the data characteristics of the standardized operation data.
Optionally, the data features may include: temperature characteristics, humidity characteristics, pressure characteristics, vibration characteristics, CPU usage characteristics, memory usage characteristics, network bandwidth usage characteristics, voltage characteristics, current characteristics, load characteristics, power characteristics, network connection status characteristics, battery status characteristics, and the like.
In particular, statistical features, time domain features, frequency domain features, etc. of the operational data may be extracted.
Further, statistical tools, neural networks, etc. may be employed to extract data features. The invention is not limited to the specific way in which the data features are extracted.
S6: and carrying out feature fusion on the data features to obtain fusion feature vectors.
Wherein the fusion feature vector is specifically [ beta ] 1 e 1 ,β 2 e 2 ,…,β n e n ]，e i Data characteristic, beta, representing class i operational data i Weight coefficient representing i-th class of operation data, i=1, 2…, n, n represent the total number of categories of operational data.
In the invention, the feature fusion can comprehensively consider the data features of a plurality of categories, and not only the characteristics of each category but also the relation between the data features. This helps to more fully capture multifaceted information of the device's operating state.
S7: and judging whether the equipment operation information is abnormal or not through an equipment operation information maintenance model based on logistic regression according to the fusion feature vector.
According to the invention, through the equipment operation information maintenance model based on logistic regression, whether the equipment operation information is abnormal or not is automatically judged, so that automatic equipment maintenance is realized, manual intervention is not needed, the influence of subjective factors such as practitioner experience, professional level and the like is avoided, the maintenance period is reduced, and the accuracy and consistency of equipment maintenance are improved.
In one possible implementation, S7 specifically includes:
s701: according to the fusion feature vector, calculating an operation state value through an equipment operation information maintenance model based on logistic regression:
σ＝β 0 +β 1 e 1 +β 2 e 2 +…+β n e n
wherein σ represents an operation state value, β 0 Representing the bias term.
Wherein, by calculating the operation state value sigma, the system quantifies the state quantity of the equipment operation information into a specific numerical value, thereby helping to understand the operation state of the equipment more clearly.
S702: determining the probability of abnormality of equipment operation information according to the operation state value:
where P represents the probability of abnormality of the device operation information, exp () represents an exponential function based on e.
By using the logistic regression model, the probability P of the abnormality occurring in the device operation information can be calculated, and the reliability measure for the abnormality of the device state is provided, which is helpful for determining the possibility of the abnormality.
S703: when the probability of abnormality of the equipment operation information is larger than the preset probability P 0 And when the equipment operation information is abnormal, determining the equipment operation information.
Wherein, a person skilled in the art can set the preset probability P according to the actual situation 0 The size of (3) is not limited in the present invention.
S704: when the probability of abnormality of the equipment operation information is smaller than or equal to the preset probability P 0 And when the equipment operation information is determined to be normal.
It should be noted that the preset probability threshold value P is introduced 0 The method can flexibly determine when the equipment operation information is considered to be abnormal according to actual requirements. This approach allows the system to be customized according to the specific application scenario, improving the adaptability of the system.
In the invention, by comparing the abnormal probability of the equipment operation information with the preset probability threshold, the system can automatically judge whether the operation state of the equipment is normal, thereby reducing the requirement for manual intervention and improving the automation level of the system. The automatic equipment maintenance is realized, manual intervention is not needed, the influence of subjective factors such as practitioner experience, professional level and the like is avoided, the maintenance period is reduced, and the accuracy and consistency of equipment maintenance are improved.
In one possible implementation manner, the training method of the equipment operation information maintenance model comprises the following steps:
and constructing a sample data set, wherein the sample data set comprises a plurality of training samples, and inputting the sample data set into the equipment operation information maintenance model to detect the equipment operation information.
Constructing a loss function of an equipment operation information overhaul model:
where L () represents a loss function, θ represents a model parameter, θ= [ β ] 0 ,β 1 ,β 2 ,…,β n ,P 0 ]N+2 model parameters in total, lambda represents the weight coefficient of the mean square error loss, y i Representing the classification result of the ith training sample,classification labels representing the i-th training sample, i=1, 2, …, N representing the total number of training samples in the sample dataset.
The size of the weight coefficient λ of the mean square error loss can be set by a person skilled in the art according to practical situations, and the invention is not limited.
In the invention, the loss function comprises two parts of mean square error loss and cross entropy loss, and can be regarded as linear combination of square error loss and logarithmic loss. Such a design can find a suitable balance point between the smoothness and classification accuracy of the balanced regression model.
And taking the loss function as an adaptability function of the particle swarm optimization algorithm.
Initializing a population comprising a plurality of particles, each particle representing a viable model parameter, the ith particle having a position X in n+2 dimensions i ＝[x i1 ,x i2 ,…,x i,n+2 ]The speed of the ith particle is V i ＝[v i1 ,v i2 ,…,v i,n+2 ]。
After each iteration, particles having a distance from the current particle less than a preset distance are taken as neighboring particles.
And determining the optimal neighbor with the smallest fitness value in the neighbor particles.
Each particle updates speed and position according to its own extremum and population extremum:
wherein V is i k+1 Representing the velocity of the ith particle at the k+1th iteration, V i k Represents the velocity of the ith particle at the kth iteration, ω represents the inertial weight factor, c 1 Representing the individual optimal acceleration factor, c 2 Represents a globally optimal acceleration factor, r 1 And r 2 All represent random numbers between 0 and 1, P i k Representing the individual optimal position of the ith particle at the kth iteration,represents the global optimum position, X, of the ith particle at the kth iteration i k Represents the position of the ith particle, X, at the kth iteration i k+1 Represents the position of the ith particle, X, at the (k+1) th iteration i best Represents the optimal neighbor of the ith particle, η represents the attractive force factor.
Wherein the inertia weight factor ω controls the inertia of the particle, i.e. the ability to maintain the current state of motion of the particle. A larger inertial weight makes the particles more prone to continue moving in the current direction, while a smaller inertial weight makes the particles more susceptible to being attracted by the locally or globally optimal solution to change the direction of motion.
Wherein the individual optimal acceleration factor c 1 For controlling the extent to which the particles are affected by their own experience.
Wherein the global optimum acceleration factor c 2 For controlling the extent to which the particles are affected by the global information.
It should be noted that, by introducing the information of individual optimum and global optimum, it is helpful to avoid sinking into local minimum, and the robustness of algorithm and global searching capability are improved.
Further, the optimal neighbors and the attractive force factors are introduced, so that the current particles can be close to the direction of the optimal neighbors when the position is updated, the convergence rate of the algorithm is improved, the global searching capacity is enhanced, the situation that the current particles are converged to the local minimum value too early is avoided, and the global searching performance of the algorithm is improved.
When the iteration number reaches the maximum iteration number, outputting the model parameter represented by the particle with the lowest adaptability as the optimal model parameter, and completing the training of the equipment operation information maintenance model.
The inertia weight factors are specifically as follows:
wherein omega min Representing the minimum value of the inertia weight factor omega max Represents the maximum value of the inertia weight factor, L i Indicating the fitness value of the ith particle, L avg Represents the average fitness value, L min Representing a minimum fitness value.
In the present invention, when the fitness value is small, a large inertial weight is used to help particles move more rapidly in the search space, accelerating convergence around the global minimum. This helps to explore the entire search space more quickly in the initial stages. When the fitness value is larger, the small inertia weight is adopted, so that fine searching in a region with a relatively larger fitness value is facilitated, and particles are prevented from crossing possible local minima due to larger inertia. The dynamic inertia weight factor can balance global search and local search in the search space, so that the algorithm has the characteristic of global exploration, and the possibility of finding a global optimal solution is increased.
Compared with the prior art, the invention has at least the following beneficial technical effects:
(1) In the invention, a unified command interface is constructed for various devices of different manufacturers and types, and the multiple heterogeneous operation data of the device to be detected is obtained through the unified command interface, so that the device is overhauled according to the operation data, and when the device is required to be updated or new devices are introduced, maintenance strategies are not required to be formulated and adjusted again, thereby reducing the complexity of management.
(2) In the invention, the operation data are evaluated according to the fluctuation, continuity, isochrony and relevance parameters, the operation data are reserved when valid, and the operation data are removed when invalid, so that misjudgment caused by invalid or abnormal data is reduced, and the accurate monitoring of the operation state of the equipment is enhanced.
(3) According to the invention, through the equipment operation information maintenance model based on logistic regression, whether the equipment operation information is abnormal or not is automatically judged, so that automatic equipment maintenance is realized, manual intervention is not needed, the influence of subjective factors such as practitioner experience, professional level and the like is avoided, the maintenance period is reduced, and the accuracy and consistency of equipment maintenance are improved.
Example 2
In one embodiment, referring to fig. 2 of the specification, a schematic structural diagram of an equipment operation information maintenance system based on multiple heterogeneous data is shown.
The invention provides an equipment operation information maintenance system based on multi-element heterogeneous data, which comprises a processor 201 and a memory 202 for storing executable instructions of the processor 201. The processor 201 is configured to call the instructions stored in the memory 202 to execute the apparatus operation information overhaul method based on the multi-heterogeneous data in embodiment 1.
The equipment operation information maintenance system based on the multi-element heterogeneous data provided by the invention can realize the steps and effects of the equipment operation information maintenance method based on the multi-element heterogeneous data in the embodiment 1, and the invention is not repeated for avoiding repetition.
Compared with the prior art, the invention has at least the following beneficial technical effects:
(1) In the invention, a unified command interface is constructed for various devices of different manufacturers and types, and the multiple heterogeneous operation data of the device to be detected is obtained through the unified command interface, so that the device is overhauled according to the operation data, and when the device is required to be updated or new devices are introduced, maintenance strategies are not required to be formulated and adjusted again, thereby reducing the complexity of management.
(2) In the invention, the operation data are evaluated according to the fluctuation, continuity, isochrony and relevance parameters, the operation data are reserved when valid, and the operation data are removed when invalid, so that misjudgment caused by invalid or abnormal data is reduced, and the accurate monitoring of the operation state of the equipment is enhanced.
(3) According to the invention, through the equipment operation information maintenance model based on logistic regression, whether the equipment operation information is abnormal or not is automatically judged, so that automatic equipment maintenance is realized, manual intervention is not needed, the influence of subjective factors such as practitioner experience, professional level and the like is avoided, the maintenance period is reduced, and the accuracy and consistency of equipment maintenance are improved.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
Claims (10)
1. The equipment operation information overhauling method based on the multi-element heterogeneous data is characterized by comprising the following steps of:
s1: a unified command interface is constructed for a plurality of different manufacturers and types of equipment;
s2: acquiring multi-element heterogeneous operation data of equipment to be detected through the unified command interface;
s3: evaluating the validity of the operation data according to the fluctuation parameter, the continuity parameter, the isochronicity parameter and the relevance parameter of the operation data, reserving the operation data when the operation data is valid, and removing the operation data when the operation data is invalid;
s4: carrying out standardization processing on the reserved operation data;
s5: extracting data characteristics of the standardized operation data;
s6: performing feature fusion on the data features to obtain fusion feature vectors;
s7: and judging whether the equipment operation information is abnormal or not through an equipment operation information maintenance model based on logistic regression according to the fusion feature vector.
2. The equipment operation information overhauling method based on multi-element heterogeneous data according to claim 1, wherein the S1 specifically comprises:
s101: defining a general device command set;
s102: selecting a general communication protocol;
s103: a general data format is adopted;
s104: adding descriptions and metadata for each different vendor and type of device;
s105: constructing a network service of a REST architecture by using a general communication protocol;
s106: introducing a proxy service, and carrying out communication and protocol conversion with the bottom layer equipment through the proxy service;
s107: defining a mapping relation between the device interface and the device command by using the configuration file;
s108: the adapter is configured for each different vendor and type of device to complete the construction of a unified command interface for a plurality of different vendors and types of devices.
3. The equipment operation information overhauling method based on multi-element heterogeneous data according to claim 1, wherein the calculation mode of the fluctuation parameter is as follows:
calculating fluctuation values of all data points in the operation data:
a ij ＝|x ij -μ i |
wherein a is ij A fluctuation value, x, representing a j-th data point of i-th type operation data ij Values, μ representing the jth data point of the ith class of operational data i Representing the mean value of the i-th class of operation data;
counting the number of data points with a fluctuation value larger than a preset fluctuation value, wherein the preset fluctuation value is 3 sigma i ，σ i Representing standard deviation of the i-th class of operation data;
calculating the fluctuation parameter according to the number of data points with the fluctuation value larger than a preset fluctuation value:
wherein A is i A represents a volatility parameter of class i operation data, a i Representing the number of data points with fluctuation values larger than preset fluctuation values in the i-th class of operation data, S i Representing the total number of data points in the class i operational data.
4. The equipment operation information overhauling method based on multi-element heterogeneous data according to claim 1, wherein the calculation mode of the continuity parameter is as follows:
counting the number of missing data points in the i-th class of operation data;
calculating the continuity parameter according to the number of missing data points:
wherein B is i A continuity parameter representing class i operational data, b i Representing the number of missing data points in the i-th class of operation data, S i Representing the total number of data points in the class i operational data.
5. The equipment operation information overhauling method based on multi-heterogeneous data according to claim 1, wherein the calculation mode of the isochronal parameters is as follows:
calculating the time length difference rate between various operation data:
wherein t is ij Representing a time length difference rate between the i-th class operation data and the j-th class operation data, t i Representing the total duration of the i-th class of operation data, t j Representing the total duration of j-th class operation data, and max () represents taking the maximum value;
calculating the isochronal parameter according to the time length difference rate among various operation data:
wherein T is i An isochronal parameter representing class i operational data.
6. The equipment operation information overhauling method based on multi-element heterogeneous data according to claim 1, wherein the calculation mode of the relevance parameter is as follows:
calculating the distance between each data point in the two types of operation data, and constructing a distance matrix d ij ]；
Wherein d ij Representing a distance between an i-th data point and a j-th data point;
starting from the upper left corner of the distance matrix, recording the minimum accumulated distance reaching each data point in the distance matrix to form a dynamic programming matrix [ D ] ij ]：
D ij ＝d ij +min[D i-1,j ,D i,j-1 ,D i-1,j-1 ]
Wherein D is ij Representing a minimum cumulative distance to an ith row and jth column data point in the distance matrix; min []Representing to take the minimum value;
starting from the lower right corner, backtracking the dynamic programming matrix to the upper left corner, and determining an optimal regular path with the smallest accumulation distance;
taking the accumulated distance of the optimal regular path as the association degree between two types of operation data;
according to the degree of association between various types of operation data, calculating the association parameters:
wherein C is i A relevance parameter representing class i operational data, c ij Representing class i operational data and class j operational dataDegree of association between the two.
7. The equipment operation information overhauling method based on multi-element heterogeneous data according to claim 1, wherein the step S3 specifically comprises:
s301: and calculating validity parameters of various operation data according to the following formula:
wherein τ i A validity parameter representing i-th class operation data, A i A represents a volatility parameter, alpha, of class i operational data 1 Weight coefficient representing volatility parameter, B i Continuity parameter, alpha, representing class i operational data 2 Weight coefficient representing continuity parameter, T i Isochronic parameter, alpha, representing class i operational data 3 Weight coefficient representing isochronic parameter, C i Relevance parameter, alpha, representing class i operational data 4 A weight coefficient representing the relevance parameter;
s302: when the validity parameter of the operation data is larger than a preset validity value, determining that the operation data is valid and reserving the operation data;
s303: and when the validity parameter of the operation data is smaller than or equal to the preset validity value, determining that the operation data is invalid, and removing.
8. The equipment operation information overhauling method based on multi-heterogeneous data according to claim 1, wherein the fusion feature vector is [ beta ] 1 e 1 ,β 2 e 2 ,…,β n e n ]，e i Data characteristic, beta, representing class i operational data i A weight coefficient representing i-th class of operation data, i=1, 2, …, n, n representing the total class number of the operation data;
the step S7 specifically comprises the following steps:
s701: according to the fusion feature vector, calculating an operation state value through an equipment operation information maintenance model based on logistic regression:
σ＝β 0 +β 1 e 1 +β 2 e 2 +…+β n e n
wherein σ represents an operation state value, β 0 Representing a bias term;
s702: determining the probability of abnormality of equipment operation information according to the operation state value:
wherein, P represents the probability of abnormality of the equipment operation information, exp () represents an exponential function based on e;
s703: when the probability of abnormality of the equipment operation information is larger than the preset probability P 0 When the equipment operation information is abnormal, determining that the equipment operation information is abnormal;
s704: when the probability of abnormality of the equipment operation information is smaller than or equal to the preset probability P 0 And when the equipment operation information is determined to be normal.
9. The equipment operation information maintenance method based on the multivariate heterogeneous data according to claim 8, wherein the training method of the equipment operation information maintenance model comprises the following steps:
constructing a sample data set, wherein the sample data set comprises a plurality of training samples, and inputting the sample data set into the equipment operation information maintenance model for equipment operation information detection;
constructing a loss function of the equipment operation information overhaul model:
where L () represents a loss function, θ represents a model parameter, θ= [ β ] 0 ,β 1 ,β 2 ,…,β n ,P 0 ]N+2 model parameters in total, lambda represents the weight coefficient of the mean square error loss, y i Representing the classification result of the ith training sample,classification labels representing the i-th training sample, i=1, 2, …, N representing the total number of training samples in the sample dataset;
taking the loss function as an adaptability function of a particle swarm optimization algorithm;
initializing a population comprising a plurality of particles, each particle representing a viable model parameter, the ith particle having a position X in an n+2-dimensional space i ＝[x i1 ,x i2 ,…,x i,n+2 ]The speed of the ith particle is V i ＝[v i1 ,v i2 ,…,v i,n+2 ]；
After each iteration, taking the particles with the distance from the current particle smaller than the preset distance as the neighbor particles;
determining an optimal neighbor with the minimum fitness value in the neighbor particles;
each particle updates speed and position according to its own extremum and population extremum:
wherein V is i k+1 Representing the velocity of the ith particle at the k+1th iteration, V i k Represents the velocity of the ith particle at the kth iteration, ω represents the inertial weight factor, c 1 Representing the individual optimal acceleration factor, c 2 Represents a globally optimal acceleration factor, r 1 And r 2 All represent random numbers between 0 and 1, P i k Representing the individual of the ith particle at the kth iterationThe optimal position of the lens is determined,represents the global optimum position, X, of the ith particle at the kth iteration i k Represents the position of the ith particle, X, at the kth iteration i k+1 Represents the position of the ith particle, X, at the (k+1) th iteration i best Representing the optimal neighbor of the ith particle, η representing the attractive force factor;
when the iteration number reaches the maximum iteration number, outputting model parameters represented by particles with the lowest adaptability as optimal model parameters, and completing the training of the equipment operation information maintenance model.
10. An equipment operation information maintenance system based on multi-element heterogeneous data is characterized by comprising a processor and a memory for storing executable instructions of the processor; the processor is configured to invoke the instructions stored in the memory to perform the equipment operation information overhaul method based on the multi-heterogeneous data of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311719770.7A CN117725398A (en) | 2023-12-14 | 2023-12-14 | Equipment operation information maintenance method and system based on multiple heterogeneous data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311719770.7A CN117725398A (en) | 2023-12-14 | 2023-12-14 | Equipment operation information maintenance method and system based on multiple heterogeneous data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117725398A true CN117725398A (en) | 2024-03-19 |
Family
ID=90200984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311719770.7A Pending CN117725398A (en) | 2023-12-14 | 2023-12-14 | Equipment operation information maintenance method and system based on multiple heterogeneous data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117725398A (en) |
-
2023
- 2023-12-14 CN CN202311719770.7A patent/CN117725398A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109347668B (en) | Training method and device for service quality assessment model | |
| CN111694879B (en) | Multielement time sequence abnormal mode prediction method and data acquisition monitoring device | |
| JP2022524244A (en) | Predictive classification of future behavior | |
| CN114585983B (en) | Method, device and system for detecting abnormal operation state of equipment | |
| US7716152B2 (en) | Use of sequential nearest neighbor clustering for instance selection in machine condition monitoring | |
| CN109471698B (en) | System and method for detecting abnormal behavior of virtual machine in cloud environment | |
| US20220156574A1 (en) | Methods and systems for remote training of a machine learning model | |
| CN115877198A (en) | Primary and secondary fusion switch fault diagnosis early warning system based on edge calculation | |
| CN111814557A (en) | Action flow detection method, device, equipment and storage medium | |
| CN117596119A (en) | Equipment data acquisition and monitoring method and system based on SNMP (simple network management protocol) | |
| CN117354171A (en) | Platform health condition early warning method and system based on Internet of things platform | |
| CN117725398A (en) | Equipment operation information maintenance method and system based on multiple heterogeneous data | |
| CN117236559A (en) | Intelligent charging pile data analysis method and device based on Internet of things | |
| CN117291576A (en) | Industrial scene data trend prediction-based method, system, computer equipment and storage medium | |
| CN111614504A (en) | Power grid regulation and control data center service characteristic fault positioning method and system based on time sequence and fault tree analysis | |
| CN116595353A (en) | Remote fault diagnosis and intelligent decision system for camera | |
| CN116089012A (en) | Self-adaptive container anomaly detection method based on container resource index | |
| CN113904933B (en) | Synchronous optimization method for performance and reliability of digital network control system | |
| KR102389317B1 (en) | Method of Determining Whether A Smart Farm Sensor has failed using a Recurrent Neural Network(RNN) | |
| CN113344072B (en) | GPB algorithm-based operation and maintenance multi-mode decision method and system and cloud server | |
| CN111597860B (en) | System and method for CNC cutter breakage prediction | |
| CN117110748A (en) | Transformer substation main equipment operation state abnormality detection method based on fusion terminal | |
| RU2819568C1 (en) | Method of detecting training data for machine learning of computer system of industrial internet of things powered by rechargeable battery | |
| US20220283139A1 (en) | Method, system and paperboard production machine for estimating paperboard quality parameters | |
| CN116415500A (en) | Breathing machine fault prediction method and system based on big data |
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 |