CN114139648A - Intelligent detection method and system for abnormity of tailing filling pipeline - Google Patents
Intelligent detection method and system for abnormity of tailing filling pipeline Download PDFInfo
- Publication number
- CN114139648A CN114139648A CN202111488714.8A CN202111488714A CN114139648A CN 114139648 A CN114139648 A CN 114139648A CN 202111488714 A CN202111488714 A CN 202111488714A CN 114139648 A CN114139648 A CN 114139648A
- Authority
- CN
- China
- Prior art keywords
- pipeline
- parameters
- parameter
- pseudo
- sequence
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Fuzzy Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Pipeline Systems (AREA)
Abstract
The invention discloses a tailing filling pipeline abnormity intelligent detection method and a system, wherein the method comprises the following steps: collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in a tailing filling process, and constructing time sequence characteristic sample data; wherein the pipeline parameters include flow and pressure; constructing a generation countermeasure network model, and training the generation countermeasure network model by using time sequence characteristic sample data to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter; generating a pseudo pipeline parameter by using a pipeline parameter generation model, and comparing the actual measurement pipeline parameter of the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the current pipeline are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline. The invention can realize intelligent and accurate detection of the abnormity of the tailing filling pipeline under the condition of no negative sample.
Description
Technical Field
The invention relates to the technical field of intersection of mineral engineering and artificial intelligence technology, in particular to an intelligent detection method and system for abnormity of a tailing filling pipeline.
Background
The 'green mining, deep mining and intelligent mining' are three major topics and future directions for ensuring sustainable and efficient development of mineral resources, wherein the green and deep mining and filling are inseparable, and one of key core technologies of the filling method mining is a pipeline conveying technology of filling slurry. Due to the danger, complexity and invisibility of mine filling operation conditions, once serious accidents such as leakage, pipe blockage, pipe explosion and the like occur to a filling pipeline, the whole filling system is directly paralyzed, and filling drill holes are scrapped, so that huge economic loss is caused to mine enterprises, and meanwhile, the production continuity is seriously influenced.
However, limited by the complexity of solid-liquid-gas three-phase mixed flow of the tailing filling slurry in the pipeline and the actual operation situation of the large-scale industrial site of the mine, the following two problems mainly exist, so that the training requirement of the existing anomaly detection model based on deep learning cannot be met: firstly, an effective abnormity detection index system such as negative pressure waves, flow balance, infrasonic wave data and the like cannot be formed; secondly, the negative sample data obtained by the abnormity occurrence or the field experiment is extremely limited or even no.
Therefore, for the abnormity detection of the tailing filling pipeline, the mine generally adopts a manual regular inspection method to eliminate hidden dangers at present, and a quantitative and accurate automatic abnormity detection method for the tailing filling pipeline is not formed.
Disclosure of Invention
The invention provides an intelligent detection method and system for abnormity of a tailing filling pipeline, and aims to solve the technical problem that the existing quantitative, accurate and intelligent tailing filling abnormity detection technology is lacked.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the invention provides an intelligent detection method for abnormity of a tailing filling pipeline, which comprises the following steps:
collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in a tailing filling process, and constructing time sequence characteristic sample data based on the collected pipeline parameters; wherein the pipeline parameters include flow and pressure;
constructing a generated confrontation network model, and training the constructed generated confrontation network model by using the time sequence characteristic sample data to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter;
generating a pseudo pipeline parameter by using the pipeline parameter generation model, and comparing the actual measurement pipeline parameter of the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline.
Further, the collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in a tailing filling process, and constructing time sequence characteristic sample data based on the collected pipeline parameters includes:
the method comprises the following steps that a plurality of monitoring nodes are arranged on a filling pipeline, a pressure sensor and a flowmeter are respectively installed at each monitoring node, and the actual measurement pressure and the actual measurement flow of the pipeline at each monitoring node under the normal operation state of the pipeline are collected through the pressure sensor and the flowmeter, so that the segmented and full-coverage monitoring of the filling pipeline is realized;
respectively preprocessing the acquired actually measured pressure of the pipeline and the acquired actually measured flow of the pipeline; wherein the preprocessing comprises data denoising processing and data standardization processing;
establishing a time sequence segmentation model based on a sliding window, and respectively constructing a pipeline flow and a pressure time sequence pipe transmission characteristic sequence in the filling process based on the pre-processed actual measurement pressure of the pipeline and the actual measurement flow of the pipeline according to the monitoring node number and the time tag corresponding to each actual measurement pressure and flow through the time sequence segmentation model.
Further, an algorithm adopted by the data denoising processing is a wavelet threshold denoising algorithm.
Further, the method for constructing the pipeline flow and pressure time sequence pipe transmission characteristic sequence in the filling process based on the pre-processed actual measurement pressure of the pipeline and the actual measurement flow of the pipeline respectively comprises the following steps:
dividing a sequence formed by the preprocessed data into a plurality of sliders according to the time sequence and the size of a sliding window; selecting a plurality of sliding blocks with continuous time to form a sequence, and using the sequence as the input and the output of the model;
and moving the whole sequence forward by one slide block, and then selecting a plurality of slide blocks with continuous time to form the sequence again until the whole sequence is traversed, thereby completing the construction of the time sequence pipe transmission characteristic sequence.
Further, the generating the countermeasure network model is a gated round unit-based dual generation countermeasure network model.
Further, training the constructed generated confrontation network model by using the time sequence feature sample data, including:
splicing the constructed pipeline flow and pressure time sequence pipe transmission characteristic sequences serving as two nodes which are mutually dual to form a multi-dimensional flow and pressure characteristic matrix;
respectively taking multidimensional flow and pressure characteristic matrixes as input, and constructing two dual generation countermeasure networks with opposite task targets; each generation countermeasure network comprises a generator and a discriminator; wherein the content of the first and second substances,
the generator adopts a gated circulation unit as a main structure and is used for generating a pseudo pipeline parameter according to a real pipeline parameter;
the discriminator adopts a linear layer as a trunk structure and is used for discriminating the pseudo pipeline parameters generated by the generator from the real pipeline parameters and guiding the generation of the pseudo pipeline parameters of the generator according to the error loss of the discrimination result and the real result;
in the training process of generating the countermeasure network, firstly, the parameters of the discriminator are fixed, and the generator optimizes the parameters according to the feedback of the discriminator; then, the discriminator carries out self-updating of parameters according to the latest generated data of the generator; and repeatedly iterating the generator and the discriminator until the generated pseudo pipeline parameters are consistent with the real pipeline parameters and the requirements of the discriminator are met.
Further, each generation countermeasure network is trained by adopting a generation countermeasure network training method of Wasserstein divergence.
Further, in the training process of generating the countermeasure network, a mean square error loss term between the generated time sequence tube output characteristic sequence and the real time sequence tube output characteristic sequence is added to guide the generation process of the generator time sequence tube output characteristic sequence.
Further, generating a pseudo pipeline parameter by using the pipeline parameter generation model, and comparing the actual measurement pipeline parameter of the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline, and the method comprises the following steps:
calculating an error err between an actually measured pipeline parameter corresponding to the current pipeline to be detected and a pseudo pipeline parameter generated by the pipeline parameter generation model, wherein the calculation formula is as follows:
wherein x isiFor actually measuring the pipeline parameters, giIs a pseudo-pipeline parameter, and n is the length of the sequence;
calculating the confidence conf of the filling pipeline in the abnormal state according to the error between the actual measurement pipeline parameter corresponding to the current pipeline to be detected and the pseudo pipeline parameter generated by the pipeline parameter generating model, wherein the calculation formula is as follows:
and judging whether the confidence conf of the filling pipeline in the abnormal state is within a preset range or not so as to determine whether the pipeline parameters of the current pipeline to be detected are within a normal range or not, thereby realizing the abnormal detection of the tailing filling pipeline.
On the other hand, the invention also provides an intelligent detection system for the abnormity of the tailing filling pipeline, which comprises the following components:
the pipeline parameter time sequence characteristic sample construction module is used for collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in the tailing filling process and constructing time sequence characteristic sample data based on the collected pipeline parameters; wherein the pipeline parameters include flow and pressure;
the pipeline parameter generation model construction module is used for constructing a generated confrontation network model, and training the constructed generated confrontation network model by using the time sequence characteristic sample data constructed by the pipeline parameter time sequence characteristic sample construction module to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter;
the filling pipeline abnormity detection module is used for generating a pseudo pipeline parameter by using the pipeline parameter generation model constructed by the pipeline parameter generation model construction module, and comparing the actual measurement pipeline parameter corresponding to the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline.
In yet another aspect, the present invention also provides an electronic device comprising a processor and a memory; wherein the memory has stored therein at least one instruction that is loaded and executed by the processor to implement the above-described method.
In yet another aspect, the present invention also provides a computer-readable storage medium having at least one instruction stored therein, the instruction being loaded and executed by a processor to implement the above method.
The technical scheme provided by the invention has the beneficial effects that at least:
the invention provides an intelligent detection method for abnormity of a tailing filling pipeline aiming at the problems of deficiency of an abnormity detection index system, insufficient quantity of negative sample data and the like in the conventional intelligent detection research and development of the abnormity of the tailing filling pipeline, wherein in the tailing filling process, the actually measured flow and pressure data of each monitoring node are collected and preprocessed, a flow and pressure time sequence pipe transmission characteristic sequence of filling slurry of each monitoring node is constructed, a dual generation model based on dual node characteristic sequences of flow and pressure monitoring data is established through the training of time sequence characteristic samples under the normal running state of the pipeline, the confidence coefficient of the pipeline in an abnormal state is calculated for the generated data obtained by the dual generation model and the real monitoring data thereof by using a confidence coefficient calculation method based on the correlation and difference between sequence spans, so that the intelligent detection of the abnormity of the tailing filling pipeline under the condition of no negative sample training is realized, the method provides necessary technical support for the quantitative, accurate and efficient abnormality detection under complex working conditions.
The intelligent detection method for the abnormity of the tailing filling pipeline based on the dual generation countermeasure network can realize intelligent and accurate intelligent detection of the abnormity of the tailing filling pipeline under the conditions of no clear abnormity detection index system and no negative sample data. On one hand, the intelligent detection method for the abnormity of the tailing filling pipeline can timely and automatically detect the abnormity of the filling pipeline, so that the serious accident of the complex pipeline is effectively avoided, the safety of mine filling operation is improved, and the economic loss of mine enterprises is reduced; on the other hand, the method provides a new idea and a general technical support for constructing an abnormal intelligent detection system under the conditions of an incomplete abnormal detection index system and insufficient accumulation of abnormal data, thereby effectively assisting the intelligent development of the abnormal detection of the state of the industrial equipment.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flow chart of an intelligent detection method for abnormality of a tailing filling pipeline according to an embodiment of the invention;
fig. 2 is a schematic diagram illustrating monitoring node deployment according to an embodiment of the present invention;
fig. 3a is a graph of measured flow of each monitoring node according to the embodiment of the present invention;
fig. 3b is a graph of measured pressure of each monitoring node according to the embodiment of the present invention;
FIG. 4 is a schematic diagram of a process for constructing a time-series pipe transmission characteristic sequence according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a dual-generation countermeasure network according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
First embodiment
The embodiment provides an intelligent detection method for the abnormity of a tailing filling pipeline, which can be realized by electronic equipment, wherein the electronic equipment can be a terminal or a server.
The method is characterized in that flow and pressure data measured by monitoring nodes of a tailing filling field pipeline are used as a basis, and flow and pressure time sequence pipe transmission characteristic sequences in the filling slurry process are respectively constructed on the basis of a time sequence segmentation model of a sliding window after data preprocessing; then, the flow and pressure time sequence pipe transmission characteristic sequences are respectively used as two nodes which are mutually dual, the characteristic sequences are spliced into a multidimensional flow and pressure characteristic matrix which is used as input, two dual generation countermeasure networks with opposite task targets are constructed, and utilizes the multidimensional flow and pressure characteristics of the pipeline in the normal operation state as a dual generation model constructed by training of a training set to establish a flow and pressure dual mode in the normal operation state, and discovering abnormal pipeline modes through a confidence calculation method based on the correlation and the difference between the sequence spans, in the actual anomaly detection application process, the confidence coefficient between the input real characteristic and the generation characteristic of the dual generation module is calculated according to the difference between the input real characteristic and the generation characteristic of the dual generation module, so that the intelligent and accurate detection of the anomaly of the tailing filling pipeline is realized under the conditions of no anomaly detection index system and no negative sample data.
Based on the above, the execution flow of the method of the embodiment is shown in fig. 1, and includes the following steps:
s1, collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in the tailing filling process, and constructing time sequence characteristic sample data based on the collected pipeline parameters; wherein the pipeline parameters include flow and pressure.
Specifically, in this embodiment, the implementation process of S1 is as follows:
s11, establishing a filling pipeline parameter measuring system covering the number, the time label, the flow and the pressure of the filling monitoring node; based on the established filling pipeline parameter measurement system, the positions of the monitoring nodes are reasonably set, and the actually measured flow and pressure data of the pipeline in the tailing filling process are collected through 4 pressure sensors and 4 flow meters (shown in figure 2) arranged on the monitoring nodes, as shown in figures 3a and 3b, so that the pipeline information including the monitoring node numbers, the time labels, the flow and the pressure is obtained, and the segmented and full-coverage monitoring of the filling pipeline is realized. And further provides a necessary data base for realizing the construction of the time sequence pipe transmission characteristics.
It should be noted that, the collection of the actually measured flow and pressure data is not limited to the above specific form, the monitoring nodes of the flow and pressure may be fixedly disposed at any position of the pipeline, and the setting criteria are as follows: the monitoring nodes of the same type of data are distributed as much as possible, and the monitoring nodes of different types of data can be properly concentrated to ensure that the characteristics of different types have good diversity and correlation; furthermore, it should be noted that all the collected data should maintain good time synchronization and have the same time scale.
S12, respectively preprocessing the acquired actually measured pressure of the pipeline and the actually measured flow of the pipeline; obtaining flow and pressure information of each monitoring node of a tailing filling pipeline;
the data preprocessing process specifically comprises the following steps: carrying out noise reduction processing based on a wavelet threshold denoising algorithm on the collected actually measured flow and pressure data to obtain filtered flow and pressure information; and carrying out normalization processing, namely normalization operation on the filtered flow and pressure information, so that the data can be better represented.
Specifically, in this embodiment, the above S12 includes the following steps:
s121, performing wavelet decomposition on the flow and pressure two-dimensional signals acquired by the information acquisition module;
s122, selecting a threshold value for each layer of signals subjected to wavelet decomposition, and performing soft thresholding on the high-frequency coefficient of the layer;
s123, performing wavelet reconstruction of the flow and pressure two-dimensional signals according to the low-frequency coefficient of the wavelet decomposition and the high-frequency coefficient after the soft thresholding;
and S124, carrying out standardization processing on the data subjected to the wavelet threshold denoising processing.
And S13, establishing a time sequence segmentation model based on a sliding window, and respectively constructing a pipeline flow and a pressure time sequence pipe transmission characteristic sequence in the filling process based on the pre-processed actual measurement pressure of the pipeline and the actual measurement flow of the pipeline according to the monitoring node number and the time label corresponding to each actual measurement pressure and flow through the time sequence segmentation model.
Specifically, the step S13 is to perform segmentation based on a sliding window on the flow and pressure data obtained by the preprocessing, so as to form a time sequence pipe transmission characteristic sequence of the flow and pressure of each monitoring node; the construction process of the time sequence pipe transmission characteristic sequence is shown in fig. 4, and comprises the following steps:
s131, dividing the preprocessed data into a plurality of sliders according to the time sequence and the size (slider _ len) of the sliding window;
s132, selecting m sliding block composition sequences of continuous time as input and output of the model.
S133, moving the whole sequence forward by one sliding block and repeating S132 until the whole sequence is traversed, thereby completing the construction of the time sequence pipe transmission characteristic sequence.
And S2, constructing a generated confrontation network model, and training the constructed generated confrontation network model by using the time sequence characteristic sample data to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter.
Specifically, in the present embodiment, the network model is a Dual-generation adaptive network model (Dual generic adaptive Networks) based on gated cyclic units as shown in fig. 5.
The step S2 is to establish a dual generation model between dual node feature sequences based on flow and pressure monitoring data by using the time-series pipe transmission feature sequence established in the step S1; the method specifically comprises the following steps:
s21, splicing the constructed flow and pressure time sequence pipe transmission characteristic sequences serving as two nodes which are dual to each other respectively to form a multi-dimensional flow and pressure characteristic matrix;
it should be noted that, the selection of the dual node is not limited to the above form, and the dual node may be arbitrarily set, and the setting criteria are as follows: carrying out correlation analysis on different characteristic sequences so as to ensure good correlation of data between dual nodes; good diversity and independence are ensured for data in dual nodes.
S22, respectively taking multidimensional flow and pressure feature matrixes as input, constructing two dual generation countermeasure Networks A, B with opposite task targets, wherein each generation countermeasure network (GAN) A, B comprises a generator and a discriminator,
a generator, which uses a Gated Recirculation Unit (GRU) as a backbone structure, and is used for generating a pseudo pressure/pseudo flow characteristic according to a real flow/pressure characteristic;
the discriminator adopts a Linear Layer (Linear Layer) as a main structure and is used for discriminating the pseudo pressure/pseudo flow characteristic generated by the generator from the real pressure/flow characteristic and guiding the generation of the pseudo pressure/pseudo flow characteristic of the generator according to the error loss of the discrimination result and the real result;
s23, in the training process of generating the countermeasure network A, B, firstly, the parameters of the discriminator are fixed, and the generator optimizes the parameters according to the feedback of the discriminator; then, the discriminator carries out self-updating of parameters according to the latest generated data of the generator; and the generator and the discriminator are iterated repeatedly until the generated pseudo pressure/pseudo flow characteristic is consistent with the real pressure/flow characteristic and the requirement of the discriminator is met.
In the embodiment, a confrontation network training process is generated in a dual mode, and training samples all adopt multidimensional flow and pressure characteristics of tailings filling pipelines in a normal state; the generation countermeasure network A, B is trained by adopting a Wasserstein Divergence generation countermeasure network training method (WGAN-div); the dual-generation countermeasure network formed by the generation countermeasure network A, B adopts the idea based on dual learning, and the original generation countermeasure network training process is improved by introducing the reconstruction error loss obtained by data reconstruction of two generators.
In addition, in this embodiment, in the training process of generating the countermeasure network generator A, B, a mean square error loss term between the generated time-series tube characteristic sequence and the real time-series tube characteristic sequence is added to guide the generation process of the generator time-series tube characteristic sequence.
Specifically, the training process of the dual generation countermeasure network based on the gated round robin unit is as follows:
it should be noted that the generation mode of the pseudo pressure/pseudo flow characteristic is not limited to reconstructing the time-series pipe transmission characteristic sequence according to the above example: a single generator may be employed to generate the data; for the conditions that the noise is small and the error distinguishing degree is not obvious, the real characteristic data can be reconstructed and amplified for multiple times to obtain the generated characteristic data of the pseudo pressure/pseudo flow characteristic.
Further, in this embodiment, in order to evaluate the reconstruction quality of the pseudo pressure/pseudo flow characteristic data, the model uses a Root Mean Square Error (RMSE) index to evaluate the difference between the pseudo pressure/pseudo flow characteristic data and the real characteristic data, and the results are compared and shown in table 1 under different model parameters.
TABLE 1 reconstruction quality of pseudo-pressure/pseudo-flow characteristic data under different parameter selection
| train_len | slider_len | batch_size | RMSE |
| 32 | 1 | 1 | 0.2270 |
| 32 | 4 | 1 | 0.2427 |
| 32 | 1 | 16 | 0.2455 |
| 64 | 1 | 1 | 0.2350 |
| 128 | 1 | 1 | 0.2477 |
In table 1, train _ len indicates the input data sequence length, slider _ len indicates the output data sequence length/slider movement sequence length, and batch _ size indicates the batch size.
S3, generating a pseudo pipeline parameter by using the pipeline parameter generation model, and comparing the actual measurement pipeline parameter of the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline.
Specifically, in this embodiment, the above-mentioned S3 is implemented by constructing a confidence calculation method based on the correlation and difference between sequence spans, and after feature generation is performed on flow and pressure data of each monitoring node in a tailing filling site by using a dual generation model, the difference between the input real feature and the generated feature of the dual generation model is compared, the confidence between the two is calculated, and the confidence between the obtained generated data and the real monitoring data is used as the confidence that a pipeline is in an abnormal state, so as to implement intelligent detection of tailing filling pipeline abnormality.
Specifically, the above S3 includes the following steps:
s31, inputting the pseudo pressure/pseudo flow characteristic data generated by the dual generation module;
and S32, calculating the error err between the input real pressure/flow characteristic data and the pseudo pressure/pseudo flow characteristic data, wherein the calculation formula is as follows:
wherein x isiAs true characteristic data, giThe data is pseudo pressure/pseudo flow characteristic data, and n is the length of the sequence;
s33, calculating the confidence conf of the filling pipeline in the abnormal state according to the error between the real characteristic sequence and the generated characteristic sequence of the dual generation module, wherein the calculation formula is as follows:
s34, judging whether the confidence conf of the filling pipeline in the abnormal state is in a preset range or not to determine whether the pipeline parameters of the pipeline to be detected are in a normal range or not, thereby realizing the abnormal detection of the tailing filling pipeline.
Further, because the multidimensional flow and pressure characteristic data of the field-measured tailing filling pipeline are label-free data in a normal state, in order to evaluate the generalization capability and effectiveness of the model, two sequences are randomly selected from eight time-sequence pipe transmission characteristic sequences and noise with different scales is added:
wherein the content of the first and second substances,
representing a sequence of time-series pipe characteristics with noise added, xiRepresenting the noise-added time-series pipe characteristic sequence, weight representing the weight of the noise, weighting the intensity of the noise to represent, niRepresenting a noise sequence, in the textIn the example, the noise sequence employs a (-1,1) uniformly distributed noise and a standard normally distributed noise, respectively. Table 2 compares the results of determining the confidence of each monitoring point in an abnormal state by the intelligent detection method for detecting abnormality of a tailing filling pipeline provided by this embodiment under noises of different scales.
TABLE 2 confidence output of intelligent detection method for tailing filling pipeline abnormality under different scale noises
In table 2, 1# LL, 2# LL, 3# LL, 4# LL, 1# YL, 2# YL, 3# YL, and 4# YL represent time-series tubing characteristic sequences of different monitoring points, respectively, where LL represents flow and YL represents pressure. The two randomly selected sequences are 4# LL and 4# YL respectively, and it can be seen that after the randomly sampled noise sequences are added to 4# LL and 4# YL, the model can accurately detect the monitoring point where the abnormal state is located along with the increase of the noise intensity and output higher confidence, the influence on the output of other sequences without noise is smaller, and the effectiveness of the judgment result of the tailing filling pipeline abnormal intelligent detection method is explained.
In summary, the present embodiment provides an intelligent detection method capable of performing real-time anomaly detection without a clear anomaly detection index system and without past negative sample data. The related measured data come from a pressure sensor and a flowmeter which are arranged at a pipeline monitoring node on a tailing filling site, and a time sequence pipe transmission characteristic sequence is respectively constructed for flow and pressure information after data preprocessing operations such as wavelet threshold denoising and standardization are carried out; then, establishing two dual generation confrontation networks with opposite task targets by respectively taking flow and pressure time sequence pipe transmission sequences as dual two nodes, and training a network model by taking multidimensional flow and pressure characteristics of the existing tailing filling pipeline in a normal state as a training set; and finally, constructing a confidence coefficient calculation method based on the reconstruction error of the dual generation module, comparing the difference between the input real features and the generated features of the dual generation module in the actual anomaly detection application process, and calculating the confidence coefficient between the input real features and the generated features of the dual generation module, namely the confidence coefficient that the filling pipeline belongs to an anomaly state, so as to realize intelligent detection of the anomaly of the tailing filling pipeline. The intelligent detection method for the abnormity of the tailing filling pipeline provides necessary technical support for the quantitative, accurate and efficient abnormity detection under complex working conditions, and can be used for green, deep and intelligent mining construction of mineral resources well.
The detection method can be migrated and expanded in the aspects of abnormal detection of large-scale industrial equipment and the like on the basis of the abnormal detection application of the tailing filling pipeline, assists the intelligent development of the industry, and has a positive promoting effect on the development of a real-time, quantitative and accurate novel method, a novel theory and a novel technology for the abnormal intelligent detection.
Second embodiment
The embodiment provides an unusual intelligent detection system of tailing filling pipeline, and it includes following module:
the pipeline parameter time sequence characteristic sample construction module is used for collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in the tailing filling process and constructing time sequence characteristic sample data based on the collected pipeline parameters; wherein the pipeline parameters include flow and pressure;
the pipeline parameter generation model construction module is used for constructing a generated confrontation network model, and training the constructed generated confrontation network model by using the time sequence characteristic sample data constructed by the pipeline parameter time sequence characteristic sample construction module to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter;
the filling pipeline abnormity detection module is used for generating a pseudo pipeline parameter by using the pipeline parameter generation model constructed by the pipeline parameter generation model construction module, and comparing the actual measurement pipeline parameter corresponding to the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline.
The intelligent detection system for the abnormity of the tailing filling pipeline of the embodiment corresponds to the intelligent detection method for the abnormity of the tailing filling pipeline of the first embodiment; the functions realized by the functional modules in the intelligent detection system for the abnormality of the tailing filling pipeline of the embodiment correspond to the flow steps in the intelligent detection method for the abnormality of the tailing filling pipeline of the first embodiment one by one; therefore, it is not described herein.
Third embodiment
The present embodiment provides an electronic device, which includes a processor and a memory; wherein the memory has stored therein at least one instruction that is loaded and executed by the processor to implement the method of the first embodiment.
The electronic device may have a relatively large difference due to different configurations or performances, and may include one or more processors (CPUs) and one or more memories, where at least one instruction is stored in the memory, and the instruction is loaded by the processor and executes the method.
Fourth embodiment
The present embodiments provide a computer-readable storage medium having stored therein at least one instruction, which is loaded and executed by a processor, to implement the method of the first embodiment. The computer readable storage medium may be, among others, ROM, random access memory, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like. The instructions stored therein may be loaded by a processor in the terminal and perform the above-described method.
Furthermore, it should be noted that the present invention may be provided as a method, apparatus or computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied in the medium.
Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should also be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or terminal that comprises the element.
Finally, it should be noted that while the above describes a preferred embodiment of the invention, it will be appreciated by those skilled in the art that, once the basic inventive concepts have been learned, numerous changes and modifications may be made without departing from the principles of the invention, which shall be deemed to be within the scope of the invention. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the embodiments of the invention.
Claims (10)
1. An intelligent detection method for abnormity of a tailing filling pipeline is characterized by comprising the following steps:
collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in a tailing filling process, and constructing time sequence characteristic sample data based on the collected pipeline parameters; wherein the pipeline parameters include flow and pressure;
constructing a generated confrontation network model, and training the constructed generated confrontation network model by using the time sequence characteristic sample data to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter;
generating a pseudo pipeline parameter by using the pipeline parameter generation model, and comparing the actual measurement pipeline parameter of the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline.
2. The method for intelligently detecting the abnormality of the tailing filling pipeline according to claim 1, wherein the step of collecting pipeline parameters of each monitoring node in the normal operation state of the pipeline in the tailing filling process and constructing time sequence characteristic sample data based on the collected pipeline parameters comprises the following steps:
the method comprises the following steps that a plurality of monitoring nodes are arranged on a filling pipeline, a pressure sensor and a flowmeter are respectively installed at each monitoring node, and the actual measurement pressure and the actual measurement flow of the pipeline at each monitoring node under the normal operation state of the pipeline are collected through the pressure sensor and the flowmeter, so that the segmented and full-coverage monitoring of the filling pipeline is realized;
respectively preprocessing the acquired actually measured pressure of the pipeline and the acquired actually measured flow of the pipeline; wherein the preprocessing comprises data denoising processing and data standardization processing;
establishing a time sequence segmentation model based on a sliding window, and respectively constructing a pipeline flow and a pressure time sequence pipe transmission characteristic sequence in the filling process based on the pre-processed actual measurement pressure of the pipeline and the actual measurement flow of the pipeline according to the monitoring node number and the time tag corresponding to each actual measurement pressure and flow through the time sequence segmentation model.
3. The tailing filling pipeline anomaly intelligent detection method as claimed in claim 2, wherein the algorithm adopted by the data denoising process is a wavelet threshold denoising algorithm.
4. The method for intelligently detecting the abnormality of the tailing filling pipeline according to claim 2, wherein the step of respectively constructing a pipeline flow rate and pressure time sequence pipe transmission characteristic sequence in the filling process based on the pre-processed measured pressure of the pipeline and the pre-processed measured flow rate of the pipeline comprises the following steps:
dividing a sequence formed by the preprocessed data into a plurality of sliders according to the time sequence and the size of a sliding window; selecting a plurality of sliding blocks with continuous time to form a sequence, and using the sequence as the input and the output of the model;
and moving the whole sequence forward by one slide block, and then selecting a plurality of slide blocks with continuous time to form the sequence again until the whole sequence is traversed, thereby completing the construction of the time sequence pipe transmission characteristic sequence.
5. The tailing filling pipeline anomaly intelligent detection method as claimed in claim 2, wherein the generation of the antagonistic network model is a gated cyclic unit-based dual generation antagonistic network model.
6. The tailing filling pipeline anomaly intelligent detection method of claim 5, wherein the training of the constructed generation countermeasure network model by using the time sequence feature sample data comprises the following steps:
splicing the constructed pipeline flow and pressure time sequence pipe transmission characteristic sequences serving as two nodes which are mutually dual to form a multi-dimensional flow and pressure characteristic matrix;
respectively taking multidimensional flow and pressure characteristic matrixes as input, and constructing two dual generation countermeasure networks with opposite task targets; each generation countermeasure network comprises a generator and a discriminator; wherein the content of the first and second substances,
the generator adopts a gated circulation unit as a main structure and is used for generating a pseudo pipeline parameter according to a real pipeline parameter;
the discriminator adopts a linear layer as a trunk structure and is used for discriminating the pseudo pipeline parameters generated by the generator from the real pipeline parameters and guiding the generation of the pseudo pipeline parameters of the generator according to the error loss of the discrimination result and the real result;
in the training process of generating the countermeasure network, firstly, the parameters of the discriminator are fixed, and the generator optimizes the parameters according to the feedback of the discriminator; then, the discriminator carries out self-updating of parameters according to the latest generated data of the generator; and repeatedly iterating the generator and the discriminator until the generated pseudo pipeline parameters are consistent with the real pipeline parameters and the requirements of the discriminator are met.
7. The method for intelligently detecting the abnormality of the tailing filling pipeline according to claim 6, wherein each generative confrontation network is trained by a training method of the generative confrontation network with Wasserstein divergence.
8. The method for intelligently detecting the abnormality of the tailing filling pipeline according to claim 7, wherein a mean square error loss term between the generated time-series pipe transmission characteristic sequence and the real time-series pipe transmission characteristic sequence is added in the training process of generating the countermeasure network so as to guide the generation process of the generator time-series pipe transmission characteristic sequence.
9. The tailing filling pipeline anomaly intelligent detection method of claim 1, characterized in that the pipeline parameter generation model is used to generate pseudo pipeline parameters, and the currently measured pipeline parameters of the pipeline to be detected are compared with the pseudo pipeline parameters generated by the pipeline parameter generation model; judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline, and the method comprises the following steps:
calculating an error err between an actually measured pipeline parameter corresponding to the current pipeline to be detected and a pseudo pipeline parameter generated by the pipeline parameter generation model, wherein the calculation formula is as follows:
wherein x isiFor actually measuring the pipeline parameters, giIs a pseudo-pipeline parameter, and n is the length of the sequence;
calculating the confidence conf of the filling pipeline in the abnormal state according to the error between the actual measurement pipeline parameter corresponding to the current pipeline to be detected and the pseudo pipeline parameter generated by the pipeline parameter generating model, wherein the calculation formula is as follows:
and judging whether the confidence conf of the filling pipeline in the abnormal state is within a preset range or not so as to determine whether the pipeline parameters of the current pipeline to be detected are within a normal range or not, thereby realizing the abnormal detection of the tailing filling pipeline.
10. The utility model provides a tailing fills unusual intelligent detection system of pipeline which characterized in that includes:
the pipeline parameter time sequence characteristic sample construction module is used for collecting pipeline parameters of each monitoring node in a normal operation state of a pipeline in the tailing filling process and constructing time sequence characteristic sample data based on the collected pipeline parameters; wherein the pipeline parameters include flow and pressure;
the pipeline parameter generation model construction module is used for constructing a generated confrontation network model, and training the constructed generated confrontation network model by using the time sequence characteristic sample data constructed by the pipeline parameter time sequence characteristic sample construction module to obtain a pipeline parameter generation model for generating a pseudo pipeline parameter;
the filling pipeline abnormity detection module is used for generating a pseudo pipeline parameter by using the pipeline parameter generation model constructed by the pipeline parameter generation model construction module, and comparing the actual measurement pipeline parameter corresponding to the current pipeline to be detected with the pseudo pipeline parameter generated by the pipeline parameter generation model; and judging whether the pipeline parameters of the pipeline to be detected are in a normal range according to the comparison result so as to realize the abnormal detection of the tailing filling pipeline.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111488714.8A CN114139648B (en) | 2021-12-07 | 2021-12-07 | Intelligent detection method and system for abnormity of tailing filling pipeline |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111488714.8A CN114139648B (en) | 2021-12-07 | 2021-12-07 | Intelligent detection method and system for abnormity of tailing filling pipeline |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114139648A true CN114139648A (en) | 2022-03-04 |
| CN114139648B CN114139648B (en) | 2022-08-02 |
Family
ID=80384620
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111488714.8A Active CN114139648B (en) | 2021-12-07 | 2021-12-07 | Intelligent detection method and system for abnormity of tailing filling pipeline |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114139648B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598851A (en) * | 2019-08-29 | 2019-12-20 | 北京航空航天大学合肥创新研究院 | Time series data abnormity detection method fusing LSTM and GAN |
| CN110909046A (en) * | 2019-12-02 | 2020-03-24 | 上海舵敏智能科技有限公司 | Time series abnormality detection method and device, electronic device, and storage medium |
| CN111669373A (en) * | 2020-05-25 | 2020-09-15 | 山东理工大学 | Network anomaly detection method and system based on space-time convolutional network and topology perception |
| CN111784596A (en) * | 2020-06-12 | 2020-10-16 | 北京理工大学 | General endoscope image enhancement method and device based on generation of antagonistic neural network |
| CN111967618A (en) * | 2019-05-20 | 2020-11-20 | 武汉剑心科技有限公司 | Online diagnosis method for voltage regulator based on deep learning |
| CN112102306A (en) * | 2020-09-25 | 2020-12-18 | 西安交通大学 | Dual-GAN-based defect detection method for edge repair feature fusion |
| CN112148955A (en) * | 2020-10-22 | 2020-12-29 | 南京航空航天大学 | Method and system for detecting abnormal time sequence data of Internet of things |
| US20210150684A1 (en) * | 2019-11-15 | 2021-05-20 | Modiface Inc. | Image-to-image translation using unpaired data for supervised learning |
| CN113127705A (en) * | 2021-04-02 | 2021-07-16 | 西华大学 | Heterogeneous bidirectional generation countermeasure network model and time sequence anomaly detection method |
| CN113298133A (en) * | 2021-05-18 | 2021-08-24 | 沈阳航空航天大学 | Supercritical unit boiler tube burst fault diagnosis method |
-
2021
- 2021-12-07 CN CN202111488714.8A patent/CN114139648B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111967618A (en) * | 2019-05-20 | 2020-11-20 | 武汉剑心科技有限公司 | Online diagnosis method for voltage regulator based on deep learning |
| CN110598851A (en) * | 2019-08-29 | 2019-12-20 | 北京航空航天大学合肥创新研究院 | Time series data abnormity detection method fusing LSTM and GAN |
| US20210150684A1 (en) * | 2019-11-15 | 2021-05-20 | Modiface Inc. | Image-to-image translation using unpaired data for supervised learning |
| CN110909046A (en) * | 2019-12-02 | 2020-03-24 | 上海舵敏智能科技有限公司 | Time series abnormality detection method and device, electronic device, and storage medium |
| CN111669373A (en) * | 2020-05-25 | 2020-09-15 | 山东理工大学 | Network anomaly detection method and system based on space-time convolutional network and topology perception |
| CN111784596A (en) * | 2020-06-12 | 2020-10-16 | 北京理工大学 | General endoscope image enhancement method and device based on generation of antagonistic neural network |
| CN112102306A (en) * | 2020-09-25 | 2020-12-18 | 西安交通大学 | Dual-GAN-based defect detection method for edge repair feature fusion |
| CN112148955A (en) * | 2020-10-22 | 2020-12-29 | 南京航空航天大学 | Method and system for detecting abnormal time sequence data of Internet of things |
| CN113127705A (en) * | 2021-04-02 | 2021-07-16 | 西华大学 | Heterogeneous bidirectional generation countermeasure network model and time sequence anomaly detection method |
| CN113298133A (en) * | 2021-05-18 | 2021-08-24 | 沈阳航空航天大学 | Supercritical unit boiler tube burst fault diagnosis method |
Non-Patent Citations (3)
| Title |
|---|
| LI ZHE ET AL.: "RCC-dual-gan：an efficient approach for outlier detection with few identified anomalies", 《ARXIV》 * |
| RUNYUAN GUO ET AL.: "A hybrid mechanism- and data-driven soft sensor based on the generative adversarial network and gated recurrent unit", 《IEEE SENSORS JOURNAL》 * |
| 刘建伟等: "生成对抗网络在各领域应用研究进展", 《自动化学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114139648B (en) | 2022-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110779746A (en) | Diagnosis method for improving composite fault of deep sparse self-encoder network rotating machinery | |
| CN110440148B (en) | Method, device and system for classifying and identifying leakage acoustic signals | |
| Ma et al. | Automated arrival-time picking using a pixel-level network | |
| CN112632680B (en) | Large civil engineering structure water leakage condition reconstruction method based on deep learning | |
| CN116382100B (en) | Oil and gas pipeline detection control system and control method | |
| CN115600513B (en) | Karst collapse monitoring early warning and prevention and control integrated informationized simulation research and judgment system | |
| CN116335925B (en) | Data enhancement-based intelligent regulation and control system for underground coal mine emulsification pump station | |
| CN109034076A (en) | A kind of automatic clustering method and automatic cluster system of mechanical fault signals | |
| CN104038792A (en) | Video content analysis method and device for IPTV (Internet Protocol Television) supervision | |
| CN111898644A (en) | Intelligent identification method for health state of aerospace liquid engine under fault-free sample | |
| CN113822201A (en) | Deep learning method for underwater object shape recognition based on flow field velocity component time course | |
| CN110837111B (en) | Seismic data interpolation method and system | |
| CN114139648B (en) | Intelligent detection method and system for abnormity of tailing filling pipeline | |
| CN117040983A (en) | Data sharing method and system based on big data analysis | |
| CN109945075B (en) | Method and device for detecting leakage degree of water supply pipeline | |
| CN116541771A (en) | Unbalanced sample bearing fault diagnosis method based on multi-scale feature fusion | |
| Britto et al. | Supervised Learning Algorithm for Water Leakage Detection through the Pipelines | |
| CN109886421A (en) | Colony intelligence coalcutter cut mode identifying system based on integrated study | |
| CN117554000A (en) | Intelligent detection method and system for leakage of tailing conveying pipeline | |
| CN117540277B (en) | Lost circulation early warning method based on WGAN-GP-TabNet algorithm | |
| CN116551466B (en) | Intelligent monitoring system and method in CNC (computerized numerical control) machining process | |
| CN114779002B (en) | Power transmission line fault point positioning method, device, equipment and storage medium | |
| CN115062680A (en) | Pipeline integrity evaluation method and device based on artificial intelligence | |
| Hu et al. | TOPOMA: Time-Series Orthogonal Projection Operator with Moving Average for Interpretable and Training-Free Anomaly Detection | |
| Xu et al. | Research on Earthquake Prediction Effect Based on Multi-Prediction Analog Model |
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 |