CN107016354B - Method and system for extracting characteristic pattern of aluminum electrolysis anode current sequence - Google Patents
Method and system for extracting characteristic pattern of aluminum electrolysis anode current sequence Download PDFInfo
- Publication number
- CN107016354B CN107016354B CN201710158649.XA CN201710158649A CN107016354B CN 107016354 B CN107016354 B CN 107016354B CN 201710158649 A CN201710158649 A CN 201710158649A CN 107016354 B CN107016354 B CN 107016354B
- Authority
- CN
- China
- Prior art keywords
- characteristic
- sequence
- target
- mode
- group
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- C—CHEMISTRY; METALLURGY
- C25—ELECTROLYTIC OR ELECTROPHORETIC PROCESSES; APPARATUS THEREFOR
- C25C—PROCESSES FOR THE ELECTROLYTIC PRODUCTION, RECOVERY OR REFINING OF METALS; APPARATUS THEREFOR
- C25C3/00—Electrolytic production, recovery or refining of metals by electrolysis of melts
- C25C3/06—Electrolytic production, recovery or refining of metals by electrolysis of melts of aluminium
- C25C3/16—Electric current supply devices, e.g. bus bars
-
- C—CHEMISTRY; METALLURGY
- C25—ELECTROLYTIC OR ELECTROPHORETIC PROCESSES; APPARATUS THEREFOR
- C25C—PROCESSES FOR THE ELECTROLYTIC PRODUCTION, RECOVERY OR REFINING OF METALS; APPARATUS THEREFOR
- C25C3/00—Electrolytic production, recovery or refining of metals by electrolysis of melts
- C25C3/06—Electrolytic production, recovery or refining of metals by electrolysis of melts of aluminium
- C25C3/20—Automatic control or regulation of cells
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Organic Chemistry (AREA)
- Metallurgy (AREA)
- Materials Engineering (AREA)
- Electrochemistry (AREA)
- Theoretical Computer Science (AREA)
- Signal Processing (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Electrolytic Production Of Metals (AREA)
Abstract
The invention discloses a characteristic pattern extraction method and a characteristic pattern extraction system for an aluminum electrolysis anode current sequence, which aim to solve the problems of low speed and low extraction characteristic value of the existing characteristic pattern extraction method. The method comprises the following steps: determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be retrieved; sequentially determining a target characteristic mode, an item set interval constraint condition corresponding to the target characteristic mode and a sequence span constraint condition; compressing a target sequence to be retrieved, segmenting and positioning compressed sequence groups corresponding to the target characteristic mode, traversing position information of potential mode groups corresponding to the target characteristic mode in each compressed sequence group, positioning and retrieving a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode groups, and counting the reproduction rate of the target characteristic mode in the characteristic mode set to analyze the working condition of the target anode guide rod.
Description
Technical Field
The invention relates to an analysis technology of anode current in an aluminum electrolysis production process, in particular to a characteristic pattern extraction method and a characteristic pattern extraction system of an aluminum electrolysis anode current sequence.
Background
Aluminum electrolysis is a complex industrial process. In the production process, the aluminum electrolysis cell is always in a high-temperature state, and the high-temperature melt has the characteristics of strong corrosivity and the like. With the improvement and development of current on-line detection technology, the detection of anode current signals on the anode guide rod of the aluminum electrolytic cell becomes possible. The anode current is a current signal flowing through the anode guide rod, is obtained by measuring and calculating equidistant voltage drop signals and is distributed around the aluminum electrolytic cell, so that the state information of the whole electrolytic cell and each area can be efficiently reflected in real time. Therefore, the characteristics of the anode current signal are efficiently analyzed by combining the characteristics of the anode current, which is particularly important for improving the actual production efficiency.
The current feature extraction speed of the existing aluminum electrolysis anode is slow, the overlapping rate between sequence positions in a feature sequence is high, the feature value of the feature sequence is relatively low, and the feature and the credibility of a feature mode cannot be satisfied.
Disclosure of Invention
The invention aims to provide a method and a system for extracting a characteristic mode of an aluminum electrolysis anode current sequence, which aim to solve the technical problems of low characteristic and reliability of the characteristic mode extracted by the existing method.
In order to achieve the purpose, the invention provides a characteristic pattern extraction method of an aluminum electrolysis anode current sequence, which comprises the following steps:
determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be retrieved;
sequentially determining a target characteristic mode, an item set interval constraint condition corresponding to the target characteristic mode and a sequence span constraint condition;
compressing a target sequence to be retrieved, segmenting and positioning a compressed sequence group corresponding to the target characteristic mode, traversing position information of a potential mode group corresponding to the target characteristic mode in each compressed sequence group, positioning and retrieving a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode group, and counting the recurrence rate of the target characteristic mode in the characteristic mode set to analyze the working condition of the target anode guide rod.
Further, the position information of the potential pattern group of the target feature pattern includes the position of each feature character.
Further, the item set interval constraint specific judgment process is as follows:
determining the position in the group of each characteristic character in the potential mode group;
and judging whether any two adjacent characteristic characters in each potential mode group meet the item set interval constraint condition, and if not, rejecting the potential mode group.
Further, the sequence span constraint specific judgment process is as follows:
acquiring the spans of all the remaining potential mode groups, and calculating the median of all the spans;
and calculating the difference value between the span and the median of each potential mode group, removing all the characteristic modes with the difference value larger than the sequence span constraint, and reserving the characteristic modes smaller than or equal to the sequence span constraint.
Further, after the feature pattern set meeting the item set interval constraint and the sequence span constraint is located and retrieved according to the position information of the potential pattern group, the steps further include calculating feature values of feature patterns of a final retrieval result:
wherein the content of the first and second substances,
is the sum of the total spans of all sequences in the feature pattern set,
for each feature pattern span in the feature pattern set, njNumber of corresponding sets of feature patterns, n, for character overlapjThe reciprocal of (c) is the weight.
Based on the method, the invention also provides a characteristic pattern extraction system of the aluminum electrolysis anode current sequence, which comprises the following modules:
a sequence acquisition module: the system is used for determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be retrieved;
a condition input module: the method comprises the steps of sequentially determining a target characteristic mode, an item set interval constraint condition corresponding to the target characteristic mode and a sequence span constraint condition;
an operation module: the device comprises a compression sequence group, a search module and a search module, wherein the compression sequence group is used for compressing a target sequence to be searched, segmenting and positioning a compression sequence group corresponding to the target characteristic mode, traversing position information of a potential mode group corresponding to the target characteristic mode in each compression sequence group, and positioning and searching a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode group;
an output module: the characteristic pattern set is used for counting the recurrence rate of the target characteristic pattern in the characteristic pattern set so as to analyze the working condition of the target anode guide rod.
Further, the position information of the potential pattern group of the target feature pattern in the operation module includes the position of each feature character.
Further, the specific judgment process of the term set interval constraint in the operation module is as follows:
determining an intra-group position of each characteristic pattern in the set of potential patterns;
and judging whether any two adjacent characteristic characters in each potential mode group meet the item set interval constraint condition, and if not, rejecting the potential mode group.
Further, the sequence span constraint in the operation module specifically judges the process as follows:
acquiring the spans of all the remaining potential mode groups, and calculating the median of all the spans;
and calculating the difference value between the span and the median of each potential mode group, removing all the characteristic modes with the difference value larger than the sequence span constraint, and reserving the characteristic modes smaller than or equal to the sequence span constraint.
Further, the system for extracting the characteristic pattern of the aluminum electrolysis anode current sequence of the present invention further includes a characteristic value calculation module, configured to, after locating and retrieving the characteristic pattern set satisfying the item set interval constraint condition and the sequence span constraint condition according to the position information of the potential pattern group, further include a step of calculating a characteristic value of a characteristic pattern of a final retrieval result:
wherein the content of the first and second substances,
is the sum of the total spans of all sequences in the feature pattern set,
for each feature pattern span in the feature pattern set, njNumber of corresponding sets of feature patterns, n, for character overlapjThe reciprocal of (c) is the weight.
The invention has the following beneficial effects:
compressing a target sequence to be retrieved, segmenting and positioning a compressed sequence group corresponding to a target characteristic mode, traversing position information of a potential mode group corresponding to the target characteristic mode in each compressed sequence group, and then positioning and retrieving a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode group, so that the retrieval speed is increased, and the speed is obviously improved compared with that of a traditional method when a longer single sequence is processed. And sequence span constraint is introduced, so that the extraction quality of the characteristic mode is improved, the overlapping property is reduced, and the characteristic and the reliability of the characteristic mode are greatly improved.
The present invention will be described in further detail below with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of a method for extracting a characteristic pattern of an aluminum electrolysis anode current sequence according to a preferred embodiment of the present invention;
FIG. 2 is a chart of the proportion occupied by consecutive characters according to the preferred embodiment of the present invention.
Detailed Description
The embodiments of the invention will be described in detail below with reference to the drawings, but the invention can be implemented in many different ways as defined and covered by the claims.
The embodiment of the invention discloses a method for extracting a characteristic pattern of an aluminum electrolysis anode current sequence, which comprises the following steps as shown in figure 1:
step S1: determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be retrieved.
The aluminum electrolysis cell is divided into different capacities, the number of anodes is different, the aluminum electrolysis cell selected in the embodiment comprises 24 groups of cathodes and anodes separated by a cell body, 24 groups of anode current signals are collected in real time in a multi-channel mode on site, and the anode current data of several groups of anode guide rods are selected for analysis. The selected anode current data for this example were taken from anode rods A6-A8 and B6-B8. Data samples are shown in table 1:
table 1:
the acquired anode current sequence data is processed by adopting an SAX (systematic aggregate approximation, a symbolic time sequence similarity measurement method) algorithm, namely, the time sequence is subjected to segmented mean values, and then the mean values are converted into discretized characters, so that the purposes of reducing dimension and noise are achieved. The specific process is as follows: suppose that continuous 30 data are selected as data samples, and the average value of every three collected data is taken as a character point to form a characterA length 10 data set. Then, the data set is normalized, namely, the data set is substituted into a data normalization algorithm
In (3), x is data of character points, max is the maximum value in the data set, min is the minimum value, and the data of the data set A6 is substituted to obtain (0.497326203209, 0.465240641711, 0.326203208556, 0.390374331551, 0.272727272727 and … …); and determining a target sequence S by obtaining character splitting points through Gaussian distribution, wherein the splitting points are shown in a table 2:
table 2:
table 2 shows that the interval from-1 to 1 is divided into α sub-intervals by Gaussian distribution, wherein β is the boundary point of each interval, the area of the data between-1 and the first split point corresponds to character A, the area of the data between the first split point and the second split point corresponds to character B, and so on.
For example, assuming that a data set is { -0.5, -0.3, 0.1, 0.3, 0.6, 0.9}, if the column (-0.97, -0.43, 0, 0.43, 0.97) with α being 6 is selected, if-0.5 is between-0.97 and-0.43 for the upper region, a character B is corresponded, and-0.3 is between-0.43 and 0 for the upper region, a character C is corresponded, and the character sequence is analogized from left to right to obtain the target sequence s, wherein when the interval value of α is determined, the length of the data set can be reasonably determined according to the length of the data set, preferably, the length of the data set is consistent with the interval value of α, and generally, the larger the interval value is, the related data processing precision is improved accordingly.
Step S2: and sequentially determining a target characteristic mode, an item set interval constraint condition corresponding to the target characteristic mode and a sequence span constraint condition.
In the invention, optionally, due to the high time-lag characteristic of the aluminum electrolysis anode current, sequential data of consecutive characters (namely scenes in which the same character is repeated and appears continuously) occupies a large proportion, and the proportion of the sequential characters in the anode current sequences of the anode guide rods A6-A8 and B6-B8 is shown in FIG. 2, as can be seen from the figure, the sequential characters in the sequence occupy a certain proportion no matter whether the sequential characters in the interval [ 1, 2], the sequential characters in (2, 5 ], or characters appearing continuously for more than 5 times, so that the target sequence to be searched needs to be compressed, and the partial sequence at the anode guide rod A6 is assumed as:
S＝BBBBAACEEEBBGAAAJDDDDHHHBAAA；
replacing the format of continuous characters in the original sequence with the format of characters and numbers, and compressing to obtain a sequence:
S’＝B4A2C1E3B2G1A3J1D4H3B1A3。
and S3, compressing the target sequence to be retrieved, dividing and positioning the compressed sequence groups corresponding to the target characteristic patterns, traversing the position information of the potential mode groups corresponding to the target characteristic patterns in each compressed sequence group, positioning and retrieving the characteristic pattern set meeting the item set interval constraint condition and the sequence span constraint condition according to the position information of the potential mode groups, and counting the recurrence rate of the target characteristic patterns in the characteristic pattern set to analyze the working condition of the target anode guide rod.
In this step, the relevant condition analysis may be any one or any combination of the corrosion condition of the anode rod and the working temperature.
In this embodiment, the position information of the potential pattern group includes the position of each characteristic character; for the compressed target sequence, the representation of the relevant position can refer to the following table 3, and an alternative storage structure is: first character position, sequence of characteristic character positions, offset, span set. The characteristic character position sequence records the specific position of each characteristic character in the compressed sequence; the offset records offset information of each characteristic character in each potential pattern group in all potential pattern groups with the same sequence of the first character position and the characteristic character position in the compressed sequence (in other words, in the same group of compressed sequences), and the span set records span values occupied by the first characteristic character and the last characteristic character in each potential pattern group in all potential pattern groups with the same sequence of the first character position and the characteristic character position in the compressed sequence (in other words, in the same group of compressed sequences). For example, the candidate set storage structure for pattern P ═ BA is shown in table 3:
table 3:
in table 3, the term set interval constraint is g ═ 0, 2, and the target feature pattern P ═ BA.
Corresponding to table 3, also taking S' ═ B4A2C1E3B2G1A3J1D4H3B1A3 as an example, in the compressed sequence, the compressed sequence group corresponding to the target feature pattern is divided in the order from left to right, and the first character positions of the three stored sequence groups are 1, 5, and 7, respectively; the characteristic character position sequences are (1, 2), (5, 7) and (11, 12) respectively. The offsets of the corresponding first character position 1 in table 3 are (1, 0) (2, 0) (2, 1) (3, 0) (3, 1), respectively, because the first group of compressed sequences corresponding to the target feature pattern is: B4A2 (its complete expression is: BBBBAA), in the compressed sequence group, there are several combined potential mode groups due to the existence of overlapped characters; the interval of the first B relative to the first A is 3, the interval relative to the 2 nd A is 4, the condition that the set interval is limited by [0, 2] is not met, and the first B is removed; the interval of the 2 nd B with respect to the first a is 2, the condition that the set interval adopts the constraint g ═ 0, 2] is satisfied, the B sequences are arranged in (0, 1, 2, 3), and the a sequences are arranged in (0, 1), so the offset of the potential mode group formed by the 2 nd B and the first a is (1, 0), and so on, the intra-group position corresponding to the first group BA feature is (1, 0) (2, 0) (2, 1) (3, 0) (3, 1). In the first group B4a2 of compressed sequence groups corresponding to the target feature pattern, for example, the BA pattern length formed by the 2 nd B and the first a is 4, the BA pattern length formed by the 4 th B and the 2 nd a is 3, and so on, the span set corresponding to the first group BA feature is 4, 3, 4, 2, 3.
In this step, the term set interval constraint is a constraint condition of the interval between adjacent feature characters of the feature pattern, and the potential pattern group smaller than or equal to the interval constraint is reserved, otherwise, the potential pattern group is removed. And screening out a characteristic pattern candidate set which meets the item set interval constraint through retrieval.
In the step, after the interval constraint conditions of the item set are screened, the spans of all the remaining potential mode groups are further acquired, and the median of all the spans is calculated; and calculating the difference value between the span and the median of each potential mode group, removing all the characteristic modes with the difference value larger than the sequence span constraint, and reserving the characteristic modes smaller than or equal to the sequence span constraint. For example: in the corresponding target characteristic pattern ABC, there may be more or less other interference characters spaced between the characteristic characters a and B, and between B and C, for example, the maximum span corresponding to the target characteristic pattern ABC in the potential pattern group is 15, and the minimum span is 3, then the median may be 9, and thus, assuming that the potential pattern group aertbutyc exists, the span of the pattern group is 8, it is determined that the span constraint condition is satisfied.
In this embodiment, a target sequence to be retrieved is compressed, (see table 3 above) a compressed sequence group corresponding to a target feature pattern is segmented and located, then position information of a potential pattern group corresponding to the target feature pattern in each compressed sequence group is traversed, and a feature pattern set satisfying an item set interval constraint condition and a sequence span constraint condition is located and retrieved according to the position information of the potential pattern group, for example: the related positioning retrieval can refer to the table 3, firstly, the compressed sequence group corresponding to the target feature mode is positioned, and then, the subsequent constraint condition judgment is carried out on various combinable potential mode groups in the compressed sequence group, so that the retrieval speed is accelerated, and the speed is obviously improved compared with the speed of the traditional method when a longer single sequence is processed. And sequence span constraint is introduced, so that the extraction quality of the characteristic mode is improved, the overlapping property is reduced, and the characteristic and the reliability of the characteristic mode are greatly improved.
Further, after the feature pattern set meeting the item set interval constraint condition and the sequence span constraint condition is located and retrieved according to the position information of the potential pattern group, the method further comprises the following steps of calculating feature values of feature patterns of a final retrieval result:
wherein the content of the first and second substances,
is the sum of the total spans of all sequences in the feature pattern set,
for each feature pattern span in the feature pattern set, njNumber of corresponding sets of feature patterns, n, for character overlapjThe reciprocal of (c) is the weight. Examples are as follows:
assuming that the sequences are BFGAGGBACGA, which are 1 to 11 in the order from left to right, and the target feature pattern is BAG, if the feature character sequences associated with the reproduction ratio in the feature pattern set finally screened are (1, 4, 5), (1, 4, 6), (7, 8, 10), respectively, where there is BA overlap between (1, 4, 5), (1, 4, 6), the calculation formula of η is as follows:
thus, the lower the overlap ratio of the characters, the greater the eigenvalue η, and conversely, the smaller the eigenvalue η, if there is no overlap between all positions, the value is 1, indicating that the pattern is very characteristic and periodic.
In the experimental demonstration of the applicant of the present application, aiming at the anode guide rod a8 in table 1, the average values of the feature patterns extracted by the conventional algorithm and the MAI L algorithm are only 17.1% and 41.1%, and it can be seen that the overlap rate between the positions of the sequence in the candidate set of feature patterns is high, and the improved algorithm of the present invention increases the average value to 77.1%, thereby effectively highlighting the feature of the feature pattern.
Corresponding to the method, the invention also provides a characteristic pattern extraction system of the aluminum electrolysis anode current sequence, which comprises the following modules:
a sequence acquisition module: the system is used for determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be retrieved;
a condition input module: the method comprises the steps of sequentially determining a target characteristic mode, an item set interval constraint condition corresponding to the target characteristic mode and a sequence span constraint condition;
an operation module: the device comprises a compression sequence group, a search module and a search module, wherein the compression sequence group is used for compressing a target sequence to be searched, segmenting and positioning a compression sequence group corresponding to the target characteristic mode, traversing position information of a potential mode group corresponding to the target characteristic mode in each compression sequence group, and positioning and searching a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode group;
an output module: the characteristic pattern set is used for counting the recurrence rate of the target characteristic pattern in the characteristic pattern set so as to analyze the working condition of the target anode guide rod.
Further, the position information of the potential pattern group of the target feature pattern in the operation module includes a first character position of the feature pattern and a feature character position sequence of the feature pattern.
Further, the specific judgment process of the term set interval constraint in the operation module is as follows:
determining an intra-group position of each characteristic pattern in the set of potential patterns;
and adjacent two bits in the group positions of each characteristic mode meet the requirement of giving reservation of the item set interval constraint condition, and do not meet rejection.
Further, the sequence span constraint in the operation module specifically judges the process as follows:
acquiring the spans of positions in all groups in the potential mode group, and calculating the median of all the spans;
and calculating the difference value between the span of the position in the group and the median in the potential mode group of each target characteristic mode, removing all the characteristic modes with the difference value larger than the sequence span constraint, and reserving the characteristic modes smaller than or equal to the sequence span constraint.
Further, the characteristic pattern extraction system for the aluminum electrolysis anode current sequence of the invention further comprises a characteristic value calculation module, wherein the calculation formula of the characteristic value in the characteristic value calculation module is as follows:
wherein the content of the first and second substances,
is the sum of the total spans of all sequences in the standard feature pattern set,
sum of products of related feature pattern spans and weights, n, for the presence of overlapping characters in a feature pattern setjThe number of overlapping subsequences.
Similarly, the system of the embodiment compresses the target sequence to be retrieved, divides and positions the compressed sequence groups corresponding to the target feature mode, then traverses the position information of the potential mode group corresponding to the target feature mode in each compressed sequence group, and then positions and retrieves the feature mode set meeting the item set interval constraint condition and the sequence span constraint condition according to the position information of the potential mode group, so that the retrieval speed is increased, and the speed is obviously improved compared with the traditional method when a longer single sequence is processed. And sequence span constraint is introduced, so that the extraction quality of the characteristic mode is improved, the overlapping property is reduced, and the characteristic and the reliability of the characteristic mode are greatly improved.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (4)
1. A characteristic pattern extraction method of an aluminum electrolysis anode current sequence is characterized by comprising the following steps:
determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be retrieved;
sequentially determining a target characteristic mode, an item set interval constraint condition corresponding to the target characteristic mode and a sequence span constraint condition;
compressing a target sequence to be retrieved, segmenting and positioning a compressed sequence group corresponding to the target characteristic mode, traversing position information of a potential mode group corresponding to the target characteristic mode in each compressed sequence group, positioning and retrieving a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode group, and counting the recurrence rate of the target characteristic mode in the characteristic mode set to analyze the working condition of the target anode guide rod;
wherein the position information of the potential pattern group of the target feature pattern comprises the position of each feature character;
the specific judgment process of the item set interval constraint is as follows: determining the position in the group of each characteristic character in the potential mode group; judging whether any two adjacent characteristic characters in each potential mode group meet the item set interval constraint condition, and if not, rejecting the potential mode group;
after the item set interval constraint conditions are screened, the sequence span constraint is specifically judged as follows:
acquiring the spans of all the remaining potential mode groups, and calculating the median of all the spans; and calculating the difference value between the span and the median of each potential mode group, removing all the characteristic modes with the difference value larger than the sequence span constraint, and reserving the characteristic modes smaller than or equal to the sequence span constraint.
2. The method for extracting the characteristic pattern of the aluminum electrolysis anode current sequence according to claim 1, wherein after the positioning and searching the characteristic pattern set satisfying the item set interval constraint and the sequence span constraint according to the position information of the potential pattern group, the steps further comprise calculating the characteristic value of the characteristic pattern of the final search result:
wherein the content of the first and second substances,
is the sum of the total spans of all sequences in the feature pattern set,
for each feature pattern span in the feature pattern set, njNumber of corresponding sets of feature patterns, n, for character overlapjThe reciprocal of (c) is the weight.
3. A characteristic pattern extraction system of an aluminum electrolysis anode current sequence is characterized by comprising the following modules:
a sequence acquisition module: the system comprises a target anode guide rod, a target search module and a search module, wherein the target anode guide rod is used for determining a target anode guide rod, carrying out data acquisition on a current signal of the target anode guide rod, and carrying out symbolization processing on the acquired current signal to obtain a target sequence to be searched;
a condition input module: the method comprises the steps of sequentially determining a target feature mode, an item set interval constraint condition corresponding to the target feature mode and a sequence span constraint condition;
an operation module: the device comprises a compression sequence group, a search module and a search module, wherein the compression sequence group is used for compressing a target sequence to be searched, segmenting and positioning a compression sequence group corresponding to the target characteristic mode, traversing position information of a potential mode group corresponding to the target characteristic mode in each compression sequence group, and positioning and searching a characteristic mode set meeting an item set interval constraint condition and a sequence span constraint condition according to the position information of the potential mode group; the position information of the potential mode group of the target characteristic mode in the operation module comprises the position of each characteristic character; the specific judgment process of the term set interval constraint in the operation module is as follows: determining an intra-group position of each characteristic pattern in the set of potential patterns; judging whether any two adjacent characteristic characters in each potential mode group meet the item set interval constraint condition, and if not, rejecting the potential mode group; the specific judgment process of the sequence span constraint in the operation module is as follows: acquiring the spans of all the remaining potential mode groups, and calculating the median of all the spans; calculating the difference between the span and the median of each potential mode group, eliminating all the characteristic modes with the difference larger than the sequence span constraint, and reserving the characteristic modes smaller than or equal to the sequence span constraint;
an output module: the characteristic pattern set is used for counting the recurrence rate of the target characteristic pattern in the characteristic pattern set so as to analyze the working condition of the target anode guide rod.
4. The system for extracting the characteristic pattern of the aluminum electrolysis anode current sequence according to claim 3, further comprising a characteristic value calculating module for calculating the characteristic value of the characteristic pattern of the final search result after locating and searching the characteristic pattern set satisfying the item set interval constraint and the sequence span constraint according to the position information of the potential pattern group:
wherein the content of the first and second substances,
total span of all sequences in a set of feature patternsThe sum of the degrees is obtained by the following steps,
for each feature pattern span in the feature pattern set, njNumber of corresponding sets of feature patterns, n, for character overlapjThe reciprocal of (c) is the weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710158649.XA CN107016354B (en) | 2017-03-16 | 2017-03-16 | Method and system for extracting characteristic pattern of aluminum electrolysis anode current sequence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710158649.XA CN107016354B (en) | 2017-03-16 | 2017-03-16 | Method and system for extracting characteristic pattern of aluminum electrolysis anode current sequence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107016354A CN107016354A (en) | 2017-08-04 |
| CN107016354B true CN107016354B (en) | 2020-07-31 |
Family
ID=59440718
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710158649.XA Active CN107016354B (en) | 2017-03-16 | 2017-03-16 | Method and system for extracting characteristic pattern of aluminum electrolysis anode current sequence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107016354B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109161931B (en) * | 2018-10-22 | 2019-11-08 | 中南大学 | Aluminium electrolysis anode electric current classification method based on shapelet conversion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106326426A (en) * | 2016-08-24 | 2017-01-11 | 四川大学 | Comparison sequence pattern mining method by adopting item sets as sequential elements |
| CN106339609A (en) * | 2016-09-19 | 2017-01-18 | 四川大学 | Heuristic mining method of optimal comparing sequence mode of free interval constraint |
| CN106469171A (en) * | 2015-08-20 | 2017-03-01 | 武汉泰乐奇信息科技有限公司 | Frequent sequential mode mining method in time serieses |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8340917B2 (en) * | 2009-12-09 | 2012-12-25 | Oracle International Corporation | Sequence matching allowing for errors |
-
2017
- 2017-03-16 CN CN201710158649.XA patent/CN107016354B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106469171A (en) * | 2015-08-20 | 2017-03-01 | 武汉泰乐奇信息科技有限公司 | Frequent sequential mode mining method in time serieses |
| CN106326426A (en) * | 2016-08-24 | 2017-01-11 | 四川大学 | Comparison sequence pattern mining method by adopting item sets as sequential elements |
| CN106339609A (en) * | 2016-09-19 | 2017-01-18 | 四川大学 | Heuristic mining method of optimal comparing sequence mode of free interval constraint |
Non-Patent Citations (4)
| Title |
|---|
| Risk Prediction for Acute Hypotensive Patients by Using Gap Constrained Sequential Contrast Patterns;Shameek Ghosh 等;《AMIA Annu Symp Proc》;20141114;第1748-1757页 * |
| 带有间隔约束的多序列模式挖掘;王华东 等;《计算机应用》;20140910;第34卷(第9期);第2612、2616、2634页 * |
| 带通配符和间隔约束的近似频繁模式挖掘研究;项泰宁;《中国优秀硕士学位论文全文数据库信息科技辑》;20160615;第I138-987页 * |
| 序列模式挖掘在企业能耗预警中的应用研究;杨剑峰;《中国优秀硕士学位论文全文数据库信息科技辑》;20130315;第I138-881页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107016354A (en) | 2017-08-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1650683B1 (en) | Search of similar features representing objects in a large reference database | |
| CN110674841B (en) | Logging curve identification method based on clustering algorithm | |
| CN111104398B (en) | Detection method and elimination method for intelligent ship approximate repeated record | |
| CN106687990A (en) | A method for selecting frames from video sequences based on incremental improvement | |
| CN107577605A (en) | A kind of feature clustering system of selection of software-oriented failure prediction | |
| CN109410238A (en) | A kind of fructus lycii identification method of counting based on PointNet++ network | |
| CN103324929B (en) | Based on the handwritten Chinese recognition methods of minor structure study | |
| CN107153670A (en) | The video retrieval method and system merged based on multiple image | |
| CN111368867B (en) | File classifying method and system and computer readable storage medium | |
| CN1783092A (en) | Data analysis device and data analysis method | |
| CN104361135A (en) | Image search method | |
| CN107016354B (en) | Method and system for extracting characteristic pattern of aluminum electrolysis anode current sequence | |
| CN106611180A (en) | Decision tree classifier construction method based on test cost | |
| CN110032585B (en) | Time sequence double-layer symbolization method and device | |
| CN116644378A (en) | Ultra-long multidimensional time sequence classification method based on double-layer nested LSTM | |
| Khare et al. | Keyframe extraction using binary robust invariant scalable keypoint features | |
| CN110751196A (en) | Oil-like drop attachment identification method in oil-water two-phase flow transparent pipe wall | |
| CN112968865B (en) | Network protocol grammatical feature rapid extraction method based on association rule mining | |
| CN110032933B (en) | Image data acquisition method and device, terminal and storage medium | |
| CN113705656A (en) | Method, system and storage medium for screening unequal sequence | |
| CN111079089B (en) | Base station data anomaly detection method based on interval division | |
| CN113707320A (en) | EN (EN) -MPA-SVM (multi-point support vector machine) -based abnormal physical sign miner distinguishing method based on correlation analysis | |
| Li et al. | A novel approach to remote sensing image retrieval with multi-feature VP-tree indexing and online feature selection | |
| CN111143436A (en) | Data mining method for big data | |
| CN109379282B (en) | Network community detection method based on multi-label propagation |
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 |