CN110443376B - State analysis method based on non-supervision machine learning algorithm and application module thereof - Google Patents
State analysis method based on non-supervision machine learning algorithm and application module thereof Download PDFInfo
- Publication number
- CN110443376B CN110443376B CN201910815181.6A CN201910815181A CN110443376B CN 110443376 B CN110443376 B CN 110443376B CN 201910815181 A CN201910815181 A CN 201910815181A CN 110443376 B CN110443376 B CN 110443376B
- Authority
- CN
- China
- Prior art keywords
- point
- alpha
- sample point
- points
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000010801 machine learning Methods 0.000 title claims abstract description 35
- 238000004458 analytical method Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 5
- 230000001174 ascending effect Effects 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 5
- 230000000087 stabilizing effect Effects 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 4
- 238000003745 diagnosis Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 238000013450 outlier detection Methods 0.000 description 3
- 230000032683 aging Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001143 conditioned effect Effects 0.000 description 1
- 230000003750 conditioning effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention discloses a state analysis method based on an unsupervised machine learning algorithm, which comprises the following steps: 1. numbering the remote modules; 2. collecting output signals of the remote module; 3. the method comprises the steps of sorting acquired output signals of a remote module, establishing a sample point set, and enabling the acquired output signals of the remote module to correspond to points in the sample point set one by one; 4. substituting points in the sample point set into an unsupervised machine learning algorithm to determine the working state of the remote module. The invention also discloses an application module for realizing the state analysis method based on the non-supervision machine learning algorithm, which comprises the following steps: energy management device, analog signal acquisition device and digital signal processing device. The method has the advantages of solving the problem that the forward and reverse samples cannot be obtained, making up the defect that the supervised machine learning algorithm cannot be adopted for fault modeling and working state prejudgment, and the like.
Description
Technical Field
The invention relates to the technical field of state analysis, in particular to a state analysis method based on an unsupervised machine learning algorithm and an application module thereof.
Background
When the optical measurement system is in operation, the remote module lacks data before failure, and cannot output data after failure, so that forward and reverse samples cannot be obtained, and therefore, a supervised machine learning algorithm cannot be adopted to perform failure modeling and prognosis, so that a failure diagnosis machine learning algorithm is needed to collect samples of output data of the current field normal remote module, and a feasibility scheme capable of solving the technical problems is formed in a plurality of failure diagnosis machine learning algorithms to predict failures of the remote module, which is a current urgent problem to be solved.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a state analysis method based on an unsupervised machine learning algorithm, which adopts the unsupervised machine learning algorithm to analyze sample characteristics of output signals of a remote module in different working states aiming at output signals of a remote module of a direct-current transmission optical measurement system, so that the working states such as aging of the remote module and the like are predicted.
In view of the shortcomings of the prior art, another object of the present invention is to provide a state analysis method and an application module thereof for implementing an unsupervised machine learning algorithm.
In order to achieve the purpose of the invention, the following technical scheme is adopted: a state analysis method based on an unsupervised machine learning algorithm can comprise the following steps:
Step 1, numbering n far-end modules, wherein n is a positive integer;
Step2, respectively collecting output signals of the n remote modules;
Step 3, the collected output signals of the n remote modules are arranged, a sample point set is established, and the collected output signals of the n remote modules are in one-to-one correspondence with points in the sample point set;
and 4, substituting points in the sample point set into an unsupervised machine learning algorithm to judge the working state of the remote module.
In step 4, the non-supervised machine learning algorithm may specifically include the following steps:
step a1, calculating Euclidean distances among sample points and arranging the Euclidean distances in an ascending order;
Step a2, calculating a kth distance and a kth field of each sample point;
and a3, calculating the reachable density and the reachable distance of each sample point.
In step 4, the determining the working state of the remote module may specifically include the following steps:
Step b1, calculating local outlier factors of each sample point in the sample point set;
step b2, if the local outlier factor of the sample point is greater than 1, the sample point is an outlier, and the remote module corresponding to the sample point is an outlier remote module;
And b3, if the local outlier factor of the sample point is very close to 1 or equal to 1, the point is a normal point, and the corresponding remote module is a normal remote module.
The step 4 specifically comprises the following steps:
step 41, calculating Euclidean distances among sample points and arranging in ascending order: let sample point set G, total n detection samples, data dimension s, for For any two data points in sample point set G, β= (x j1,xj2,…,xjs), the euclidean distance from point α to point β is d (α, β):
Wherein α= (x i1,xi2,…,xis),β＝(xj1,xj2,…,xjs) is the coordinates of two points in n-dimensional euclidean space, and x i1,…,xis and x j1,…,xjs respectively represent the specific positions of the points α and β in the coordinate system;
Step 42, calculating a kth distance of each sample point: defining d k (α) as the kth distance of point α, d k (α) =d (α, β), satisfies the following condition:
(1) At least k points excluding the point alpha are arranged in the sample point set G, and beta 'epsilon G { x is not equal to alpha }, so that d (alpha, beta')isnot more than d (alpha, beta);
(2) At most, k-1 points excluding the point alpha exist in the sample point set G, and beta 'epsilon G { x is not equal to alpha }, so that d (alpha, beta')isnot more than d (alpha, beta); namely, the point beta is the kth point closest to the point alpha, excluding the point alpha;
Step 43, calculating the kth field of each sample point: let N k (α) be the kth distance neighborhood of point α, satisfy:
Nk(α)＝{β'∈{G{x≠α}|d(α,β')≤d(α,β)}}，
Wherein, beta ' is the point excluding the point alpha in the sample point set G, d (alpha, beta ') and d (alpha, beta) are the distances from the point alpha to the point beta ' and from the point alpha to the point beta, respectively, the sample point set G contains all the points with the distance to the point alpha smaller than the kth neighborhood distance of the point alpha, and then N k (alpha) is more than or equal to k;
Step 44, calculating the reachable distance d k (α, β) of each sample point:
dk(α,β)＝max{dk(α),d(α,β)}，
Where d k (α) is the kth distance of point α, and max represents the maximum value of d k (α, β) and d k (α);
I.e., the kth reachable distance from point β to point α is at least the kth distance from point α, the k points nearest to point α are considered equal to the reachable distance from point α and are all equal to d k (α);
step 45, calculating the reachable density ρ k (α) of the point α:
Where d k (α, β) is the reachable distance of each sample point, N k (α) is the kth distance neighborhood of point α, representing the average reachable distance from point α to all points within the kth neighborhood of point α;
Step 46, calculating and arranging the local outlier factors LOF of each sample point in descending order, and defining the local outlier factor LOF k (α) of the point α as:
Wherein ρ k(α),ρk (β) is the achievable density of points α, β; LOF k (α) represents the average of the ratio of the local reachable densities of other points within neighborhood N k (α) of point α to the local reachable density of point α. If the ratio of ρ k (α) to ρ k (β) is equal to 1, indicating that the density of points within the neighborhood of point α is similar, then point α and neighborhood are clustered together; if the ratio of ρ k (α) to ρ k (β) is less than 1, indicating that the density of the point α is higher than the density of its neighborhood points, the point α is a dense point; if the ratio of ρ k (α) to ρ k (β) is greater than 1, indicating that the density of point α is less than its neighborhood point density, then point α is an outlier;
If the local outlier factor LOF k (α) > >1 of the point α, the sample point set G is an outlier set, and the corresponding remote module is an outlier remote module, otherwise, the sample point set G is a normal point set, and the corresponding remote module is a normal remote module.
In order to achieve the other purpose of the invention, the following technical scheme is adopted: a state analysis method based on an unsupervised machine learning algorithm and an application module thereof are realized, and the method comprises the following steps: the energy management device comprises an energy management device, an analog signal acquisition device and a digital signal processing device, wherein one end of the energy management device is connected with one end of the digital signal processing device, the other end of the energy management device is connected with one end of the analog signal acquisition device, the other end of the digital signal processing module is connected with the other end of the analog signal acquisition module, the analog signal acquisition device comprises an overvoltage protection unit, a voltage lifting unit and an amplitude adjusting unit which are sequentially connected, the digital signal processing device comprises an A/D converter and an FPGA logic processor which are sequentially connected, and the energy management device comprises a PPC battery, a voltage stabilizing unit and a voltage reference unit which are sequentially connected.
The application module can also comprise an LED, and the LED is connected with the digital signal processing unit.
The ultra-high voltage or current passes through a voltage dividing resistor or a shunt resistor to obtain a voltage drop within the range of-10V to +10V; the measuring function is to collect and condition the input analog voltage, in order to accurately measure the voltage drop, the front-stage circuit part of the sensor uses a low-noise operational amplifier to condition the signal, so that the signal can be collected and converted by the A/D converter better.
Overvoltage protection unit: in order to prevent the sensor circuit from being impacted and damaged by abnormal input signals, the input signals need to be limited, namely an overvoltage protection circuit is added to the signal input end. Typically implemented using a zener diode, which is a semiconductor device having high resistance characteristics up to a critical reverse breakdown voltage.
A voltage lifting unit: the part comprises a two-stage amplifying circuit and a filter circuit, wherein the first-stage amplifying circuit forms an inverse attenuator, an offset signal is added at the same time, and the second-stage amplifying circuit forms an inverse amplifier, and the two-stage amplifying circuits act at the same time, so that the range of an input signal is between-10V and +10V.
A voltage adjustment unit: the input signal is designed to be-10V to +10V, and the signal range which can be received by the A/D converter of the sensor is 0V to 2.5V, so that the input signal needs to be conditioned. We raised the lowest voltage to 0V and the highest voltage to 2.5V. Thus, a-10V voltage signal corresponds to a "0" output by the A/D converter and a +10V voltage corresponds to a "4096" output by the A/D converter.
The A/D converter is used for converting the analog signal into a digital signal, and the conversion bit number is 12 in order to meet the requirement of measurement precision.
The data processing device selects a logic processor FPGA (Field-Programmable GATE ARRAY), and the data processing device is mainly used for calculating, encoding and transmitting signals acquired by the A/D converter, and finally converting the analog voltage from 0V to 2.5V into a digital sequence from 0V to 4096. In addition, before sending the digital signal, the logic processor FPGA needs to add the start bit, the undervoltage identification bit, and the check bit to the digital signal. The high and low levels are modulated onto an LED infrared light emitting tube, and the emission wavelength of the LED is 850nm. The electrical signal is converted into an optical signal and finally transmitted through an optical fiber.
The energy of the whole far-end module is derived from a laser, and the principle of the laser is as follows: the photovoltaic energy converter (PPC) converts the received laser energy into an electrical signal and then supplies power to the whole module. Meanwhile, the laser also comprises a synchronous clock signal; the energy management device is used for reducing and reversing the voltage of the photovoltaic energy converter to obtain the working voltage of the driving device.
The invention has the advantages and beneficial effects that:
1. The invention analyzes the sample characteristics of the output signals of the remote module in different working states by adopting an unsupervised machine learning algorithm aiming at the output signals of the remote module of the direct current transmission light measuring system, and realizes the prejudgment of the aging state of the remote module, thereby solving the problem that the forward and reverse samples cannot be obtained, and overcoming the defect that the supervised machine learning algorithm cannot be adopted for fault modeling and working state prejudgment.
2. According to the state analysis method based on the non-supervision machine learning algorithm, the LOF outlier detection algorithm is adopted for carrying out working state analysis on the remote module aiming at the sample characteristics of the output data signals of the remote module, so that the output signal characteristics before the remote module fails are found, and further the judgment is carried out before the remote module fails, so that the defects that the remote module fails to output data before the failure, and the forward and reverse samples cannot be obtained and the supervised machine learning algorithm cannot be adopted for carrying out failure modeling and judgment are overcome.
3. The remote module mainly comprises three devices: the energy management device, the analog signal acquisition device and the digital signal processing device, wherein the analog signal acquisition device is used for acquiring and conditioning the input analog voltage, so that the signal can be acquired and converted by the A/D converter better; the digital signal processing device is used for converting the analog signal into a digital signal, and the energy management device is used for supplying power to the whole remote module. Among numerous fault diagnosis machine learning algorithms, an unsupervised machine learning algorithm is a typical density-based high-precision fault diagnosis machine learning algorithm, and can be widely applied to various fields, such as telecommunication fraud analysis, credit card fraud detection, network attack behavior detection, medical diagnosis, extreme weather forecast, and the like. If the outlier factor is much greater than 1, the data point is an outlier; if the outlier factor is close to 1, then the data point is a normal data point.
Drawings
FIG. 1 is an overall flow chart of a state analysis method based on an unsupervised machine learning algorithm of the present invention.
FIG. 2 is a flow chart of an embodiment of a state analysis method based on an unsupervised machine learning algorithm of the present invention.
Fig. 3 is a block diagram of an application module according to the present invention.
Detailed Description
Examples
The invention is further described below in connection with the following detailed description.
As shown in fig. 1, a state analysis method based on an unsupervised machine learning algorithm includes the following steps:
Step 1, numbering n far-end modules, wherein n is a positive integer;
Step2, respectively collecting output signals of the n remote modules;
Step 3, the collected output signals of the n remote modules are arranged, a sample point set is established, and the collected output signals of the n remote modules are in one-to-one correspondence with points in the sample point set;
and 4, substituting points in the sample point set into an unsupervised machine learning algorithm to judge the working state of the remote module.
As shown in fig. 2, an outlier detection algorithm of local anomaly factor (LOF) is adopted to analyze the working state of a remote module of an optical measurement system, and the state analysis specifically includes the following steps:
Step 1: numbering a plurality of remote modules, namely a remote module numbered 1, a remote module numbered … … of a remote module numbered n-1 and a remote module numbered n of a remote module numbered 2;
Step 2: respectively collecting output signals of the n remote modules;
Step 3: the collected signals are subjected to preliminary arrangement: establishing a sample point set G, wherein the acquired output signals of the n remote modules are in one-to-one correspondence with points in the sample point set G, namely a sample point 1, a sample point 2 … …, a sample point n-1 and a sample point n;
Step 4: substituting points in the sample point set G into an outlier detection algorithm of a local anomaly factor (LOF) to determine an operating state of the remote module:
1. Calculating Euclidean distances among sample points and arranging in ascending order: let sample point set G, total n detection samples, data dimension s, for For any two data points in sample point set G, β= (x j1,xj2,…,xjs), the euclidean distance from point α to point β is:
Wherein α= (x i1,xi2,…,xis),β＝(xj1,xj2,…,xjs) is the coordinates of two points in n-dimensional euclidean space;
2. Calculating a kth distance for each sample point: defining d k (α) as the kth distance of point α, d k (α) =d (α, β), satisfies the following condition:
1) At least k points which do not include alpha are arranged in the set, and beta 'epsilon G { x is not equal to alpha }, so that d (alpha, beta')isnot more than d (alpha, beta);
2) At most, k-1 points not including alpha exist in the set, and beta 'epsilon G { x is not equal to alpha }, and d (alpha, beta')isnot more than d (alpha, beta); that is, point β is the kth point closest to point α, excluding α;
3. Calculate the kth field for each sample point: let N k (α) be the kth distance neighborhood of point α, satisfy:
Nk(α)＝{β'∈{G{x≠α}|d(α,β')≤d(α,β)}}，
Where beta ' is a point in the collection excluding alpha, d (alpha, beta ') d (alpha, beta) is the distance from the point alpha to beta ' and beta, respectively,
The set contains all points with the distance to the point alpha less than the k neighborhood distance of the point alpha, and N k (alpha) is easy to know and is more than or equal to k;
4. Calculating the reachable distance of each sample point:
dk(α,β)＝max{dk(α),d(α,β)}，
Wherein d k (α) is the kth distance of point α;
I.e., the kth reachable distance from point β to point α is at least the kth distance from point α, the k points nearest to point α are considered equal to the reachable distance from α and are all equal to d k (α);
5. The achievable density for each sample point is calculated:
Where d k (α, β) is the reachable distance of each sample point, N k (α) is the kth distance neighborhood of point α, representing the average reachable distance from point α to all points within the kth neighborhood of point α;
6. The local outlier factor LOF for each sample point is calculated and arranged in descending order: the local outlier factor for point α is defined as:
Wherein ρ k(α),ρk (β) is the achievable density of points α, β; LOF k (α) represents the average of the ratio of the local reachable densities of other points within neighborhood N k (α) of point α to the local reachable density of point α. If the ratio of ρ k (α) to ρ k (β) is equal to 1, indicating that the density of points within the neighborhood of point α is similar, then point α and neighborhood are clustered together; if the ratio of ρ k (α) to ρ k (β) is less than 1, indicating that the density of the point α is higher than the density of its neighborhood points, the point α is a dense point; if the ratio of ρ k (α) to ρ k (β) is greater than 1, indicating that the density of point α is less than its neighborhood point density, then point α is an outlier;
If the local outlier factor LOF k (α) > >1 of the point α, the sample point set G is an outlier set, and the corresponding remote module is an outlier remote module, otherwise, the sample point set G is a normal point set, and the corresponding remote module is a normal remote module.
As shown in fig. 3, an application module for implementing the state analysis method based on the non-supervised machine learning algorithm includes: the energy management device comprises an energy management device, an analog signal acquisition device, a digital signal processing device and an LED, wherein one end of the energy management device is connected with one end of the digital signal processing device, the other end of the energy management device is connected with one end of the analog signal acquisition device, the other end of the digital signal processing module is connected with the other end of the analog signal acquisition module, the analog signal acquisition device comprises an overvoltage protection unit, a voltage lifting unit and an amplitude adjusting unit which are sequentially connected, the digital signal processing device comprises an A/D converter and an FPGA logic processor which are sequentially connected, the energy management device comprises a PPC battery, a voltage stabilizing unit and a voltage reference unit which are sequentially connected, and the LED is connected with the digital signal processing unit.
The foregoing detailed description is directed to embodiments of the invention which are not intended to limit the scope of the invention, but rather to cover all modifications and variations within the scope of the invention.
Claims (5)
1. The state analysis method based on the non-supervision machine learning algorithm is characterized by comprising the following steps of:
Step 1, numbering n far-end modules, wherein n is a positive integer;
Step2, respectively collecting output signals of the n remote modules;
Step 3, the collected output signals of the n remote modules are arranged, a sample point set is established, and the collected output signals of the n remote modules are in one-to-one correspondence with points in the sample point set;
and 4, substituting points in the sample point set into an unsupervised machine learning algorithm to judge the working state of the remote module, wherein the step 4 specifically comprises the following steps:
step 41, calculating Euclidean distances among sample points and arranging in ascending order: let sample point set G, total n detection samples, data dimension s, for For any two data points in sample point set G, β= (x j1,xj2,...,xjs), the euclidean distance from point α to point β is d (α, β):
Wherein α= (x i1,xi2,...,xis),β＝(xj1,xj2,...,xjs) is the coordinates of two points in n-dimensional euclidean space, and x i1,...,xis and x j1,...,xjs respectively represent the specific positions of the points α and β in the coordinate system;
Step 42, calculating a kth distance of each sample point, defining d k (α) as the kth distance of the point α, and d k (α) =d (α, β), where the following conditions are satisfied:
(1) At least k points excluding the point alpha are arranged in the sample point set G, and beta 'epsilon G { x is not equal to alpha }, so that d (alpha, beta')isnot more than d (alpha, beta);
(2) At most, k-1 points excluding the point alpha exist in the sample point set G, and beta 'epsilon G { x is not equal to alpha }, so that d (alpha, beta')isnot more than d (alpha, beta); namely, the point beta is the kth point closest to the point alpha, excluding the point alpha;
Step 43, calculating the kth field of each sample point: let N k (α) be the kth distance neighborhood of point α, satisfy:
Nk(α)＝{β′∈{G{x≠α}|d(α，β′)≤d(α，β)}}，
Wherein, beta ' is the point excluding the point alpha in the sample point set G, d (alpha, beta ') and d (alpha, beta) are the distances from the point alpha to the point beta ' and from the point alpha to the point beta, respectively, the sample point set G contains all the points with the distance to the point alpha smaller than the kth neighborhood distance of the point alpha, and then N k (alpha) is more than or equal to k;
Step 44, calculating the reachable distance d k (α, β) of each sample point:
dk(α,β)＝max{dk(α)，d(α，β)}，
Where d k (α) is the kth distance of point α, and max represents the maximum value of d k (α, β) and d k (α);
I.e., the kth reachable distance from point β to point α is at least the kth distance from point α, the k points nearest to point α are considered equal to the reachable distance from point α and are all equal to d k (α);
step 45, calculating the reachable density ρ k (α) of the point α:
where d k (α, β) is the reachable distance of each sample point, N k (α) is the kth distance neighborhood of point α, representing the average reachable distance from point α to all points within the kth neighborhood of point α;
Step 46, calculating local outlier factors LOF of each sample point and arranging in descending order: the local outlier factor LOF k (α) at point α is defined as:
Wherein ρ k(α),ρk (β) is the achievable density of points α, β; LOF k (α) represents the average of the ratio of the local reachable densities of other points within neighborhood N k (α) of point α to the local reachable density of point α; if the ratio of ρ k (α) to ρ k (β) is equal to 1, then the point α and the neighborhood are co-clustered; if the ratio of ρ k (α) to ρ k (β) is less than 1, then point α is a dense point; if the ratio of ρ k (α) to ρ k (β) is greater than 1, then point α is an outlier;
If the local outlier factor LOF k (alpha) > 1 of the point alpha, the sample point set G is an outlier set, the corresponding remote module is an outlier remote module, otherwise, the sample point set G is a normal point set, and the corresponding remote module is a normal remote module.
2. The method for analyzing a state based on an unsupervised machine learning algorithm according to claim 1, wherein in step 4, the unsupervised machine learning algorithm specifically comprises the steps of:
step a1, calculating Euclidean distances among sample points and arranging the Euclidean distances in an ascending order;
Step a2, calculating a kth distance and a kth field of each sample point;
and a3, calculating the reachable density and the reachable distance of each sample point.
3. The method for analyzing a state based on an unsupervised machine learning algorithm according to claim 1, wherein in step 4, the determining the working state of the remote module specifically includes the steps of:
Step b1, calculating local outlier factors of each sample point in the sample point set;
step b2, if the local outlier factor of the sample point is greater than 1, the sample point is an outlier, and the remote module corresponding to the sample point is an outlier remote module;
And b3, if the local outlier factor of the sample point is equal to 1, the point is a normal point, and the corresponding remote module is a normal remote module.
4. An application module for implementing the non-supervised machine learning algorithm based state analysis method of claim 1, comprising: the energy management device comprises an energy management device, an analog signal acquisition device and a digital signal processing device, wherein one end of the energy management device is connected with one end of the digital signal processing device, the other end of the energy management device is connected with one end of the analog signal acquisition device, the other end of the digital signal processing module is connected with the other end of the analog signal acquisition module, the analog signal acquisition device comprises an overvoltage protection unit, a voltage lifting unit and an amplitude adjusting unit which are sequentially connected, the digital signal processing device comprises an A/D converter and an FPGA logic processor which are sequentially connected, and the energy management device comprises a PPC battery, a voltage stabilizing unit and a voltage reference unit which are sequentially connected.
5. The application module of claim 4, further comprising an LED, the LED being coupled to the digital signal processing unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910815181.6A CN110443376B (en) | 2019-08-30 | 2019-08-30 | State analysis method based on non-supervision machine learning algorithm and application module thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910815181.6A CN110443376B (en) | 2019-08-30 | 2019-08-30 | State analysis method based on non-supervision machine learning algorithm and application module thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110443376A CN110443376A (en) | 2019-11-12 |
| CN110443376B true CN110443376B (en) | 2024-05-17 |
Family
ID=68438578
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910815181.6A Active CN110443376B (en) | 2019-08-30 | 2019-08-30 | State analysis method based on non-supervision machine learning algorithm and application module thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110443376B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113702894A (en) * | 2021-09-23 | 2021-11-26 | 云南电网有限责任公司电力科学研究院 | Fault diagnosis method and device for remote module and merging unit |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106412841A (en) * | 2016-11-23 | 2017-02-15 | 沈阳航空航天大学 | Indoor location method based on DBSCAN algorithm |
| CN106446500A (en) * | 2016-07-14 | 2017-02-22 | 江苏信息职业技术学院 | Improved supervised orthogonal neighborhood preserving embedding dimension reduction-based fault identification method |
| CN109100143A (en) * | 2018-07-06 | 2018-12-28 | 华中科技大学 | Fault Diagnosis of Roller Bearings and equipment based on CEEMDAN and CFSFDP |
| CN109446184A (en) * | 2018-08-28 | 2019-03-08 | 湖南大唐先科技有限公司 | Power generation big data preprocess method and system based on big data analysis platform |
| CN109934354A (en) * | 2019-03-12 | 2019-06-25 | 北京信息科技大学 | Abnormal deviation data examination method based on Active Learning |
-
2019
- 2019-08-30 CN CN201910815181.6A patent/CN110443376B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446500A (en) * | 2016-07-14 | 2017-02-22 | 江苏信息职业技术学院 | Improved supervised orthogonal neighborhood preserving embedding dimension reduction-based fault identification method |
| CN106412841A (en) * | 2016-11-23 | 2017-02-15 | 沈阳航空航天大学 | Indoor location method based on DBSCAN algorithm |
| CN109100143A (en) * | 2018-07-06 | 2018-12-28 | 华中科技大学 | Fault Diagnosis of Roller Bearings and equipment based on CEEMDAN and CFSFDP |
| CN109446184A (en) * | 2018-08-28 | 2019-03-08 | 湖南大唐先科技有限公司 | Power generation big data preprocess method and system based on big data analysis platform |
| CN109934354A (en) * | 2019-03-12 | 2019-06-25 | 北京信息科技大学 | Abnormal deviation data examination method based on Active Learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110443376A (en) | 2019-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8581608B2 (en) | Apparatus and method for detecting abnormality in solar cell power generation system | |
| JP2020123951A (en) | Method, apparatus, and device, for predicting failure of optical module | |
| CN106501714B (en) | Digital switch detection circuit and method with fault diagnosis function | |
| CN112327100B (en) | Power failure detection method and system based on Internet of things | |
| CN106384400B (en) | A kind of employee for office detection system on duty and method | |
| KR101532163B1 (en) | Evaluation and condition diagnosis system for photovoltaic power generator | |
| CN110443376B (en) | State analysis method based on non-supervision machine learning algorithm and application module thereof | |
| KR101000734B1 (en) | Apparatus sensing shade and breaking down of solar photovaltaic power generation cell line/module/group and monitoring solar photovaltaic power generation amount | |
| CN105044440A (en) | Fuel cell monolithic voltage inspection system based on LTC6803 | |
| CN210488596U (en) | Application module based on unsupervised machine learning algorithm | |
| CN111894814A (en) | Fault processing system and method for power generation system | |
| CN117650621A (en) | Method and system for monitoring network operation state of power system | |
| JP6621000B2 (en) | Method and apparatus for determining deterioration of solar cell module | |
| CN216350976U (en) | Fault detection circuit, direct current power supply system and direct current equipment | |
| CN103616573B (en) | A kind of IGCT QRR measures system and measuring method thereof | |
| CN110222399B (en) | Power health assessment method and device | |
| CN114510397A (en) | Current output detection circuit, method and device, electronic device and storage medium | |
| CN110702972B (en) | Adaptive sampling method and device for analog signals | |
| KR20180124182A (en) | Multi-inverter PV monitoring system | |
| CN110426217B (en) | Fault processing method based on wheel sensor signal processing device | |
| CN112611959A (en) | High voltage generator XPSI board fault diagnosis device | |
| CN111614561A (en) | Intelligent road side data transmission method and device based on edge calculation and distributed system | |
| CN215575347U (en) | Slope intelligent monitoring equipment fault self-diagnosis circuit | |
| CN214200424U (en) | Four-wire system temperature acquisition and self-checking circuit | |
| CN116722758B (en) | Micro inverter control system |
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 |