CN114755515A - Energy storage power station key equipment fault diagnosis and early warning method based on data mining - Google Patents
Energy storage power station key equipment fault diagnosis and early warning method based on data mining Download PDFInfo
- Publication number
- CN114755515A CN114755515A CN202210331092.6A CN202210331092A CN114755515A CN 114755515 A CN114755515 A CN 114755515A CN 202210331092 A CN202210331092 A CN 202210331092A CN 114755515 A CN114755515 A CN 114755515A
- Authority
- CN
- China
- Prior art keywords
- early warning
- data
- energy storage
- power station
- fault
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- 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
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B19/00—Alarms responsive to two or more different undesired or abnormal conditions, e.g. burglary and fire, abnormal temperature and abnormal rate of flow
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/001—Methods to deal with contingencies, e.g. abnormalities, faults or failures
- H02J3/0012—Contingency detection
Abstract
The invention discloses a fault diagnosis and early warning method for key equipment of an energy storage power station based on data mining, which adopts a K-neighbor mutual information and Apriori association rule algorithm to realize fault diagnosis and early warning of the key equipment of the energy storage power station. Compared with other intelligent algorithm diagnosis and early warning strategies, the method aims at the problems of various types of key equipment parameters of the energy storage power station, strong coupling, huge historical data and the like, combines engineering experience, adopts a K-neighbor mutual information algorithm to screen main fault influence factors, can reduce manual screening workload and errors, and improves execution efficiency.
Description
Technical Field
The invention relates to the field of fault diagnosis of energy storage equipment, in particular to a fault diagnosis and early warning method for key equipment of an energy storage power station based on data mining.
Background
In order to better realize the aim of 'double-carbon' and constructing a novel power system with new energy as a main body, renewable energy sources such as photovoltaic energy, wind power energy and the like are mainly used for replacing emission reduction in the future. However, new energy power generation has the characteristics of intermittence, randomness and poor schedulability, large-scale new energy power generation grid connection provides a severe test for a power system, the existing flexible resources gradually have the weakness of supporting a power grid to accept fluctuating energy with high proportion, stored energy is used as a higher-quality flexible resource, new energy output can be effectively smoothed, and auxiliary services such as frequency modulation and peak shaving can be provided.
The existing energy storage system mostly adopts a lithium battery, and due to the characteristics of the lithium battery, the safety problem of equipment can be caused by improper use. Energy storage safety is often a result of interaction of various factors, leading to battery abuse and thermal runaway, ultimately leading to accidents. The method can be divided into the aspects of battery body faults, operating environment defects, energy storage system comprehensive management system defects and the like. In the aspect of lack of an energy storage system comprehensive management system, when information sharing of a Battery Management System (BMS), an energy storage converter (PCS) and an Energy Management System (EMS) is incomplete or not in time, the PCS and the battery are not properly configured and coordinated, the PCS is abnormal after fault clearing, conflict occurs between a measuring device and the management system, and other system management problems may cause that faults cannot be timely and effectively controlled, and finally evolve into accidents.
The key equipment of the energy storage power station comprises a battery body, a battery management system BMS, an energy storage converter PCS and the like, the equipment is various in types, large in quantity and high in energy density of the energy storage battery, and the efficiency of the original operation and maintenance mode of the energy storage power station is lower due to the potential safety hazard. The conventional key equipment fault diagnosis and early warning method can only provide abnormal state warning, has low response speed, can only reflect the current abnormal state information of equipment, cannot analyze the root fault cause of the equipment reflected by the abnormal state, and is difficult to meet the on-site refined operation and maintenance requirements of a power station and even cause energy storage safety accidents.
Disclosure of Invention
The invention aims to solve the problems that the existing method for diagnosing the faults of the key equipment of the energy storage power station has low response speed, cannot analyze the root fault reasons of the equipment reflected by abnormal states and is difficult to meet the field refined operation and maintenance requirements of the power station, and provides a method for diagnosing and warning the faults of the key equipment of the energy storage power station based on data mining.
In order to solve the technical problem, the technical scheme adopted by the invention is as follows: a fault diagnosis and early warning method for key equipment of an energy storage power station based on data mining comprises the following steps: s01, establishing a diagnosis early warning rule base according to historical data, and establishing a responsive operation and maintenance suggestion base aiming at various rules; s02, collecting and processing abnormal signals in a period of time, carrying out fuzzy matching with the rule base, sending out fault diagnosis early warning information if matching is successful, or only carrying out warning for the abnormal state; and S03, if the system sends out diagnosis early warning information, corresponding operation and maintenance suggestions are given according to different fault types deduced in the step S02.
Further, the process of establishing the diagnostic fault rule base and the operation and maintenance suggestion base in step S01 is as follows: s11, preliminarily screening out influence parameters possibly causing equipment failure from the operation and maintenance parameters of the original energy storage power station according to engineering experience, and then selecting and preprocessing corresponding influence parameter data containing the corresponding influence parameters in the complete period from a historical database; s12, analyzing the correlation between the preliminarily screened influence parameters and the faults according to historical data, and determining the parameters with the correlation meeting the requirements as main influence factors; s13), discretizing the historical data of the main influence factors according to a set threshold value, and establishing an equipment fault diagnosis early warning rule base by adopting a data mining algorithm; s14), making operation and maintenance suggestions according to the fault types corresponding to different rules in the rule base, and establishing an operation and maintenance suggestion base.
Further, a K-neighbor mutual information algorithm is adopted to analyze the correlation between a certain parameter and a fault, and the specific method comprises the following steps: firstly, setting a correlation threshold value theta, then calculating mutual information MI between a certain fault and the fault by using a K-neighbor mutual information algorithm, comparing the correlation threshold value theta with the mutual information MI, if MI is less than theta, abandoning the parameter, otherwise, recording the parameter.
Further, a group of influence parameters related to the fault is determined by adopting an Apriori association rule algorithm, so that an equipment fault diagnosis and early warning rule base is established, and the specific method comprises the following steps: a. setting a minimum support degree and a minimum confidence degree; b. scanning the discretized historical data of the main influence factors, and finding out all items meeting the minimum support degree, wherein the items are called a frequent 1 item set; c. scanning the history data of the discretized main influence factors on the basis of the frequent 1 item set to find a frequent 2 item set, and circulating the process until no new frequent k +1 item set exists, wherein k is the number of the types of the main influence factors; d. and calculating whether the confidence coefficient among the item sets is not less than the minimum confidence coefficient, and determining strong association.
Furthermore, in the process of generating the frequent k +1 item set by the frequent k item set, a connection step and a pruning step are adopted to improve the efficiency, wherein the connection step is to find out the frequent k +1 item set Lk+1From L to LkThe medium-frequency k item set connection generates a candidate set Ck+1To ensure that the resulting item sets are not related, the same party can be connected if and only if the top k-1 items after the ordering of the 2 frequent k item sets; the pruning step is to find out Ck+1The item set meeting the minimum support degree directly scans the discrete calendar of the main influence factors The history data calculates the support of each item set when Ck+1When the value is larger than a set threshold value, the priori knowledge is adopted to compress Ck+1。
Further, step S02 is specifically: s21, collecting the real-time data of the parameters screened in the step S01, and recording the signal data if an abnormal signal occurs; s22, timing is started by collecting first data from S21, abnormal signals occurring in a period of time are counted, and data of the abnormal signals are recorded; s23, discretizing the data recorded in the steps S21 and S22 according to a set threshold value, and fuzzy matching the processed parameter real-time data with a diagnosis and early warning rule base; and S24, if the matching is successful, sending out diagnosis and early warning information aiming at different fault types and levels, otherwise, only vibrating to warn the abnormal state.
Further, step S03 includes 2 steps, specifically: s31, if the matching in the step S02 is successful, the diagnosis and early warning information is sent out, and meanwhile, corresponding operation and maintenance suggestions are accurately matched in the operation and maintenance library according to different fault types; and S32, pushing the operation and maintenance suggestions matched in the S31 through the picture.
The invention has the beneficial effects that: according to the method, the fault diagnosis and early warning of the key equipment of the energy storage power station are realized by adopting the K-neighbor mutual information and Apriori association rule algorithm, compared with the traditional warning strategy, the method can deeply dig the implicit association between each operation parameter and the fault information of the key equipment of the energy storage power station, effectively shortens the fault diagnosis time, provides simple and efficient maintenance suggestions, and provides fine guidance for operation and maintenance personnel. Compared with other intelligent algorithm diagnosis and early warning strategies, the method aims at the problems of various types of key equipment parameters of the energy storage power station, strong coupling, huge historical data and the like, combines engineering experience, adopts a K-neighbor mutual information algorithm to screen main fault influence factors, can reduce manual screening workload and errors, and improves execution efficiency. Meanwhile, the expert database established by the method supports functions of rule expansion, modification, deletion and the like, and can provide support for new operation and maintenance requirements in the later period of the energy storage power station.
The effects are as follows: by monitoring the running state of key equipment of the energy storage power station in real time, when a fault is about to occur or has occurred, a plurality of collected early warning or alarm signals in a certain time range are subjected to fuzzy matching with a rule base, potential hazards possibly existing in the running process of the system are analyzed, the abnormality or fault type of the energy storage system is judged timely and accurately, abnormal working condition limitation, fault protection and acousto-optic alarm are automatically implemented, and corresponding operation and maintenance suggestions are given.
Drawings
FIG. 1 is an overall flow diagram of the present invention;
FIG. 2 is a flow chart of screening the primary failure contributors using K-neighbor mutual information;
FIG. 3 is a flow chart of data mining using the Apriori algorithm;
FIG. 4 is a diagram of a partial fault and corresponding operation and maintenance recommendations;
FIG. 5 is a diagram of the rule base triggering mechanism and the specific triggering path according to the present invention.
Detailed Description
The invention is further described with reference to the following figures and specific examples.
Example 1
When a power grid fails, the EMS can receive action event sequence records of power grid switches and relay protection devices sent by different RTUs, substation automation systems and other systems, arrange the events according to time sequence and store the events as historical information in a historical database. According to different key equipment of the energy storage power station, the method can be mainly divided into two parts, namely PCS diagnosis and early warning and battery system diagnosis and early warning.
(1) PCS diagnostic early warning
And diagnosing and early warning according to the operation state of the PCS and real-time data, and performing operations of shutdown maintenance, fastening a connector, replacing a fault module, even quitting maintenance and the like on the PCS. The main fault types include the following:
a) over-high or under-low AC voltage
b) AC frequency too high or too low
c) Over-high or under-low DC voltage
d) Overload, overheat, short-circuit of converter
e) Radiator overheating
f) Converter island
g) DSP fault
(2) Battery system diagnosis early warning
When the battery system runs, if analog quantity of the battery, such as voltage, current, temperature and the like, exceeds the safety protection limit value, local fault isolation can be implemented, the battery pack with the problem is quitted from running, and meanwhile, protection information is reported. The main failure types are as follows:
a) over-voltage of battery
b) Under voltage of battery
c) Over-temperature of battery
d) Low temperature of battery
e) Overcurrent of battery
f) SOC out-of-limit
As shown in fig. 1, a method for diagnosing and warning a fault of a key device of an energy storage power station based on data mining mainly includes the following specific steps:
1. the establishment and maintenance of the two libraries: and establishing a diagnosis early warning rule base according to historical data, and establishing a corresponding operation and maintenance suggestion base aiming at various rules. The rule base and the operation and maintenance base can carry out background maintenance such as supplement, deletion, modification and the like according to different application scene requirements;
2. Equipment fault diagnosis and early warning: processing abnormal signals within a period of time, carrying out fuzzy matching with the rule base, sending fault diagnosis early warning information if the matching is successful, and otherwise, only giving an alarm for the abnormal state;
3. pushing operation and maintenance suggestions: and if the system sends out diagnosis early warning information, giving out a corresponding operation and maintenance suggestion according to the fault type obtained in the step 2, as shown in fig. 4.
According to the specific implementation manner of the present invention, the process of establishing and maintaining the two libraries in step 1 includes four steps, which are specifically as follows:
(1) according to engineering experience, influence parameters causing the fault of the energy storage power station equipment are preliminarily screened out from a large number of operation and maintenance parameters. Selecting data in a complete period from a historical database and preprocessing the data, including processing abnormal values and vacancy values;
(2) analyzing the correlation between the preliminarily screened parameters and the faults according to historical data, and determining the parameters with the correlation meeting the requirements as main influence factors;
(3) discretizing the historical data of the parameters according to a set threshold value, and establishing a rule base by adopting a data mining algorithm. The rule base can be subsequently expanded according to needs;
(4) And making an operation and maintenance suggestion according to the fault types corresponding to different rules in the rule base, and establishing an operation and maintenance suggestion base. The operation and maintenance library can be subsequently expanded according to needs.
The correlation analysis method in the step 2 is a K-neighbor mutual information algorithm, and the data mining algorithm for establishing the rule base is an Apriori association rule algorithm.
Mutual Information (MI) is a method for measuring the degree of interdependence of two random variables, and is based on the concept of entropy in information theory. The larger the mutual information of two random variables, the stronger the correlation. The main influence factors of the faults are selected according to the mutual information, the influence of multivariable coupling on the equipment faults is not considered, and the defects of large high-dimensional calculation amount, low precision and the like exist. The K-nearest neighbor mutual information method can avoid direct calculation of probability density, and has good evaluation effect on complex nonlinear relations.
From the definition of mutual information, mutual information MI: (x, y) The smaller, the sequenceXAnd sequenceYThe smaller the amount of common information, the smaller the sequence correlation; in contrast, mutual information MI (x, y) The larger, the sequenceXAnd sequenceYThe greater the amount of common information, the greater the sequence correlation. Therefore, the main influence factors of the fault of the key equipment of the energy storage power station can be determined according to the MI.
As shown in fig. 2, the process of screening the main influence factors of the fault by using the K-nearest neighbor mutual information method is as follows: firstly, setting a correlation threshold value theta, then calculating mutual information MI between a certain fault and the fault by using a K-neighbor mutual information algorithm, comparing the correlation threshold value theta with the mutual information MI, if MI is less than theta, abandoning the parameter, otherwise, recording the parameter. If a plurality of main influence factors are calculated, the process is repeated, and the correlation between other factors and the fault is calculated.
According to the above description of the K-neighbor mutual information algorithm, steps 1 (1) and (2) are further explained:
the key equipment of the energy storage power station comprises a battery body, a BMS (battery management system), a PCS (personal communication system) and the like, the equipment parameters are various, the coupling is strong, the historical data volume is huge, and the main influence factors causing faults are difficult to determine. Therefore, the invention preliminarily screens out the influence parameters causing the equipment fault from various operation and maintenance parameters such as current, voltage, frequency, temperature, pressure and the like by combining engineering experience. And selecting data in the complete period from the historical database for preprocessing, including processing abnormal values and vacancy values. A correlation threshold is then set, i.e. exceeding the threshold value considers the parameter to be sufficiently strongly correlated with the fault as a major cause of the occurrence of the fault. And finally, analyzing the correlation between the preliminarily screened parameters and the equipment fault by adopting a K-neighbor mutual information algorithm according to historical data, and determining 3-5 parameters exceeding a correlation threshold value as final parameters for establishing a rule base.
The association rule may find an internal association mechanism between 2 or more variables and predict the occurrence of an event by analyzing the established association mechanism. When key equipment such as PCS and the like has faults, fault related information is recorded in an EMS system log and stored in a historical database, and correlation rules can be applied to extract the dependency and the correlation between the fault type and the fault phenomenon by analyzing historical data so as to obtain mode characteristics of various faults and guide the diagnosis and early warning process.
Two key indexes of support and confidence are usually adopted to measure the correlation analysis result. In particular with item setsXAnd item setYExample, degree of supportSSet of finger itemsXAndYprobability of coincidence, confidenceC isNamed item setXWhen it happens, item setYThe probability of occurrence. Wherein the set of items is a collection of items, includingkThe item set of individual items is calledkA set of items, such as the set { A, B, C } is a set of 3 items.
In general, minimum supports_minAnd minimum confidencec_minThe threshold value used for defining the support degree and the confidence degree, the item set meeting the support degree not less than the minimum support degree is called a frequent item set, and simultaneously the support degree is metRules that are not less than the minimum support and confidence not less than the minimum confidence are referred to as strong rules.
The mining of the association rule in this embodiment is performed to obtain mode characteristics of various faults, that is, a group of signals corresponding to the faults, as shown in fig. 3, that is, a specific process is as follows:
a. setting a minimum support degree and a minimum confidence degree; b. scanning the discretized historical data of the main influence factors, and finding out all items meeting the minimum support degree, wherein the items are called a frequent 1 item set; c. scanning the history data of the discretized main influence factors on the basis of the frequent 1 item set to find a frequent 2 item set, and circulating the process until no new frequent k +1 item set exists, wherein k is the number of the types of the main influence factors; d. and calculating whether the confidence coefficient among the item sets is not less than the minimum confidence coefficient, and determining strong association.
Because the whole database needs to be scanned completely every time a new item set is searched, the execution efficiency is low, the efficiency is improved by adopting a connecting step and a pruning step in the process of generating a frequent k +1 item set by the frequent k item set, wherein the connecting step is to find out a frequent k +1 item set Lk+1From L to LkThe medium-frequency k item set connection generates a candidate set Ck+1To ensure that the resulting item sets are not related, the same party can be connected if and only if the top k-1 items after the ordering of the 2 frequent k item sets; the pruning step is to find out Ck+1The item set meeting the minimum support degree, the support degree of each item set is calculated by directly scanning the history data of the discretized main influence factors, and when C is used k+1When the value is larger than a set threshold value, the priori knowledge is adopted to compress Ck+1。
Since the association rule algorithm can only process boolean data, and the values of parameters such as current, voltage, temperature and the like of the key equipment of the energy storage power station are continuous, the historical data of the parameters screened in the step (2) need to be discretized according to a set threshold. Taking the current as an example, the current is recorded as "1" when the current exceeds the set threshold, and is recorded as "0" otherwise. And then, establishing a diagnosis and early warning rule base by discretizing data and adopting an Apriori association rule algorithm through a confidence coefficient formula combining a connection step and a pruning step.
According to a specific embodiment of the present invention, the device fault diagnosis and early warning process in step 2 includes 4 steps, which are specifically as follows:
(1) acquiring real-time data of the parameters screened in the step 1, and recording the signal data if an abnormal signal occurs;
(2) collecting first data from (1), starting timing, counting abnormal signals occurring within a period of time, and recording data of the abnormal signals;
(3) discretizing the data recorded in the steps (1) and (2) according to a set threshold value, and fuzzy matching the processed parameter real-time data with a diagnosis and early warning rule base;
(4) If the matching is successful, different diagnosis and early warning information such as sound, light, vibration, pushed pictures and the like shown in fig. 5 is sent out according to different fault types and levels, and otherwise, the abnormal state is only warned through vibration.
According to a specific embodiment of the present invention, the operation and maintenance suggestion pushing process in step 3 includes 2 steps, which are specifically as follows:
(1) if the matching in the step 2 is successful, the diagnosis and early warning information is sent out, and meanwhile, corresponding operation and maintenance suggestions are accurately matched in the operation and maintenance library according to different fault types;
(2) and (4) pushing the operation and maintenance suggestions matched in the step (1) through a picture.
According to the specific implementation of the present invention, the whole process of the method is further explained by taking the battery module failure as an example:
and analyzing the collected historical operation data of signals such as battery current and voltage, battery compartment temperature and pressure, battery system fire-fighting out-of-limit and the like by a K-neighbor mutual information algorithm according to engineering experience to obtain parameters such as battery monomer voltage, battery compartment temperature, battery system fire-fighting and the like and larger than battery module fault MI. Discretizing the parameters according to a set threshold, wherein the discretization result is as follows: the battery single overvoltage is recorded as '1', the battery compartment overtemperature is recorded as '1', the battery system fire-fighting out-of-limit alarm is recorded as '1', and the battery module fault is recorded as '1'. And determining { battery monomer overvoltage, battery temperature overhigh, battery system fire alarm } → battery module fault strong association rule according to an Apriori association rule algorithm.
When the energy storage power station operates, when some of parameters acquired in real time occur { battery monomer overvoltage, battery temperature is too high, and a battery system is in fire alarm }, the system starts timing and automatically carries out fuzzy matching with the rule base. If two other abnormal signals appear within 30 seconds, the rule base is triggered under the action of the combined alarm signal, the early warning is carried out by adopting the modes of sound, light, vibration, image pushing and the like, and information such as maintenance suggestions and the like is pushed, so that reference can be provided for operation and maintenance maintainers.
It should be noted that the above-mentioned contents are only used for understanding the method of the present invention and the core idea thereof, and the protection scope of the present invention is not limited to the above-mentioned examples. Various other modifications and alterations of this invention will become apparent to those skilled in the art from this disclosure without departing from the spirit of this invention, and such changes are intended to be included within the scope of this invention.
Claims (7)
1. A fault diagnosis and early warning method for key equipment of an energy storage power station based on data mining is characterized by comprising the following steps: the method comprises the following steps: s01, establishing a diagnosis early warning rule base according to historical data, and establishing a responsive operation and maintenance suggestion base aiming at various rules; s02, collecting and processing abnormal signals within a period of time, carrying out fuzzy matching with the rule base, sending fault diagnosis early warning information if matching is successful, or only carrying out warning aiming at the abnormal state; and S03, if the system sends out diagnosis early warning information, corresponding operation and maintenance suggestions are given according to different fault types deduced in the step S02.
2. The energy storage power station key equipment fault diagnosis and early warning method based on data mining as claimed in claim 1, characterized in that: step S01 is to establish the diagnostic fault rule base and the operation and maintenance suggestion base as follows: s11, according to engineering experience, preliminarily screening influence parameters which may cause equipment failure from the operation and maintenance parameters of the original energy storage power station, and then selecting corresponding influence parameter data in a complete period from a historical database and preprocessing the data; s12, analyzing the correlation between the preliminarily screened influence parameters and the faults according to historical data, and determining the parameters with the correlation meeting the requirements as main influence factors; s13), discretizing the historical data of the main influence factors according to a set threshold value, and establishing an equipment fault diagnosis early warning rule base by adopting a data mining algorithm; s14), making operation and maintenance suggestions according to the fault types corresponding to different rules in the rule base, and establishing an operation and maintenance suggestion base.
3. The energy storage power station key equipment fault diagnosis and early warning method based on data mining as claimed in claim 2, characterized in that: the correlation between a certain parameter and a fault is analyzed by adopting a K-neighbor mutual information algorithm, and the specific method comprises the following steps: firstly, setting a correlation threshold value theta, then calculating mutual information MI between a certain fault and the fault by using a K-neighbor mutual information algorithm, comparing the correlation threshold value theta with the mutual information MI, if MI is less than theta, abandoning the parameter, otherwise, recording the parameter.
4. The energy storage power station key equipment fault diagnosis and early warning method based on data mining as claimed in claim 2, characterized in that: determining a group of influence parameters related to the fault by adopting an Apriori association rule algorithm, thereby establishing an equipment fault diagnosis early warning rule base, which comprises the following specific steps: a. setting a minimum support degree and a minimum confidence degree; b. scanning the discretized historical data of the main influence factors, and finding out all items meeting the minimum support degree, wherein the items are called a frequent 1 item set; c. scanning the history data of the discretized main influence factors on the basis of the frequent 1 item set to find a frequent 2 item set, and circulating the process until no new frequent k +1 item set exists, wherein k is the number of the types of the main influence factors; d. and calculating whether the confidence coefficient among the frequent item sets is not less than the minimum confidence coefficient, and determining strong correlation.
5. The energy storage power station key equipment fault diagnosis and early warning method based on data mining as claimed in claim 4, characterized in that: adopt in the process that the frequent k item set generates the frequent k +1 item setThe efficiency is improved by a connecting step and a pruning step, wherein the connecting step is to find out a frequent k +1 item set Lk+1From L to LkConnecting the medium-frequency k item sets to generate a candidate set C k+1To ensure that the resulting item sets are not related, the top k-1 items after ordering of the 2 frequent k item sets are the same and can be connected; the pruning step is to find out Ck+1The item set meeting the minimum support degree, the support degree of each item set is calculated by directly scanning the history data of the discretized main influence factors, and when C isk+1When the value is larger than a set threshold value, the priori knowledge is adopted to compress Ck+1。
6. The energy storage power station key equipment fault diagnosis and early warning method based on data mining as claimed in claim 1, characterized in that: step S02 specifically includes: s21, collecting the real-time data of the parameters screened in the step S01, and recording the signal data if an abnormal signal occurs; s22, timing is started by collecting first data from S21, abnormal signals occurring in a period of time are counted, and data of the abnormal signals are recorded; s23, discretizing the data recorded in the steps S21 and S22 according to a set threshold value, and fuzzy matching the processed parameter real-time data with a diagnosis and early warning rule base; and S24, if the matching is successful, sending out diagnosis and early warning information aiming at different fault types and levels, otherwise, only vibrating to warn the abnormal state.
7. The energy storage power station key equipment fault diagnosis and early warning method based on data mining as claimed in claim 1, characterized in that: step S03 includes 2 steps, specifically: s31, if the matching in the step S02 is successful, the diagnosis and early warning information is sent out, and meanwhile, corresponding operation and maintenance suggestions are accurately matched in the operation and maintenance library according to different fault types; and S32, pushing the operation and maintenance suggestions matched in the S31 through the picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210331092.6A CN114755515A (en) | 2022-03-31 | 2022-03-31 | Energy storage power station key equipment fault diagnosis and early warning method based on data mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210331092.6A CN114755515A (en) | 2022-03-31 | 2022-03-31 | Energy storage power station key equipment fault diagnosis and early warning method based on data mining |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114755515A true CN114755515A (en) | 2022-07-15 |
Family
ID=82329355
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210331092.6A Pending CN114755515A (en) | 2022-03-31 | 2022-03-31 | Energy storage power station key equipment fault diagnosis and early warning method based on data mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114755515A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115825756A (en) * | 2023-02-16 | 2023-03-21 | 中国华能集团清洁能源技术研究院有限公司 | Distributed energy storage power station fault multi-stage diagnosis method and system |
| CN116577673A (en) * | 2023-07-12 | 2023-08-11 | 深圳先进储能材料国家工程研究中心有限公司 | Distributed neural network-based energy storage power station fault diagnosis method and system |
| CN116720324A (en) * | 2023-05-15 | 2023-09-08 | 中铁第四勘察设计院集团有限公司 | Traction substation key equipment fault early warning method and system based on prediction model |
-
2022
- 2022-03-31 CN CN202210331092.6A patent/CN114755515A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115825756A (en) * | 2023-02-16 | 2023-03-21 | 中国华能集团清洁能源技术研究院有限公司 | Distributed energy storage power station fault multi-stage diagnosis method and system |
| CN116720324A (en) * | 2023-05-15 | 2023-09-08 | 中铁第四勘察设计院集团有限公司 | Traction substation key equipment fault early warning method and system based on prediction model |
| CN116577673A (en) * | 2023-07-12 | 2023-08-11 | 深圳先进储能材料国家工程研究中心有限公司 | Distributed neural network-based energy storage power station fault diagnosis method and system |
| CN116577673B (en) * | 2023-07-12 | 2023-09-12 | 深圳先进储能材料国家工程研究中心有限公司 | Distributed neural network-based energy storage power station fault diagnosis method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114755515A (en) | Energy storage power station key equipment fault diagnosis and early warning method based on data mining | |
| CN112817280A (en) | Implementation method for intelligent monitoring alarm system of thermal power plant | |
| CN113240289B (en) | Power grid dispatching control system operation state evaluation method and system | |
| CN106209405B (en) | Method for diagnosing faults and device | |
| CN106054104A (en) | Intelligent ammeter fault real time prediction method based on decision-making tree | |
| CN112072642B (en) | Method for estimating abnormal operation state of power system | |
| CN116433045A (en) | Accurate supply management method and system for electric power emergency materials | |
| CN109766448A (en) | A kind of long-range preventative frequency conversion management system | |
| CN113644741B (en) | Massive fault recording data processing method based on edge calculation | |
| CN114202304A (en) | Intelligent monitoring processing method and system for power grid faults | |
| CN117374978B (en) | Grid-connected scheduling management method and system constructed by combining knowledge graph | |
| CN112054510B (en) | Method for estimating abnormal operation state of power system | |
| CN116468426A (en) | Power grid fault determining method and storage medium based on Apriori algorithm | |
| CN116094174A (en) | Knowledge graph-based power grid operation and maintenance monitoring method, system, equipment and medium | |
| Devitt et al. | Topographical proximity for mining network alarm data | |
| CN111539642B (en) | Object-oriented power plant data acquisition and processing system and method thereof | |
| CN112101422B (en) | Typical case self-learning method for power system fault case | |
| CN115333113A (en) | Electric power allotment system based on machine learning | |
| Wang et al. | LSTM-based alarm prediction in the mobile communication network | |
| Dai et al. | Automatic generation of power grid dispatching and control scheme based on heterogeneous information network | |
| CN111932097A (en) | Data quality monitoring method and device based on electric vehicle service platform | |
| CN113129160A (en) | Electric power communication network inspection method based on equipment state perception and intellectualization | |
| CN111401577A (en) | Device management method, device and storage medium | |
| Pinzón et al. | Data analytics for power grid control centers management | |
| CN115963408B (en) | Energy storage power station single battery fault early warning system and method |
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 |