CN108958215A - A kind of engineering truck failure prediction system and its prediction technique based on data mining - Google Patents
A kind of engineering truck failure prediction system and its prediction technique based on data mining Download PDFInfo
- Publication number
- CN108958215A CN108958215A CN201810556980.1A CN201810556980A CN108958215A CN 108958215 A CN108958215 A CN 108958215A CN 201810556980 A CN201810556980 A CN 201810556980A CN 108958215 A CN108958215 A CN 108958215A
- Authority
- CN
- China
- Prior art keywords
- attribute
- frequent item
- item set
- failure
- engineering truck
- 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
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/24—Pc safety
- G05B2219/24065—Real time diagnostics
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the crossing domain of data mining and engineer application, a kind of engineering truck failure prediction system and its prediction technique based on data mining.This includes the excavation failure correlation frequent item set module being sequentially connected, establishes decision forest module and failure predication module.The basic data of engineering truck is pre-processed by excavating failure correlation frequent item set module, obtains fault car attribute；It excavates failure correlation frequent item set module and frequent item set is further obtained by fault car attribute；It establishes decision forest module and carries out data reprocessing, construct training set, and further form decision tree and corresponding decision forest；Failure predication module carries out classification prediction to the failure risk of engineering truck, and exports result.Rule relevant to failure generation can be obtained from a large amount of irregular engineering truck work information, the failure that existing engineering truck work information prediction may be found that is efficiently used, be a kind of efficiently quick failure prediction method.
Description
Technical field
The invention belongs to the crossing domain of data mining and engineer application, a kind of engineering truck failure based on data mining
Forecasting system and its prediction technique.
Background technique
In actual life, need to use a large amount of engineering truck in various mining sites, building site.Due to the originals such as work long hours
Cause, these engineering trucks are easy to appear failure, and once break down, and the factors such as consigning, repair, delaying work also can be to entire engineering
It affects greatly.In order to reduce the failure rate of engineering truck, be lost to reduce engineering, by data mining algorithm, excavate
The relationship hidden between engineering truck operating condition and failure, so that making correctly prediction for engineering car state is to have very big demand
's.
For engineering truck failure predication, there are following two major demands:
1) rule relevant to failure generation is obtained from a large amount of irregular engineering truck work information, to carry out targetedly
Processing；
2) current engineering car state can be determined in time, has found that it is likely that the failure and suspension of engineering work vehicle of appearance in time
Operation.
Effectively for two above major demands and timely failure predication can be made using the method for data mining.
Data mining refers to the process of and excavates useful information from mass data.Specifically, data mining refer to from
Largely have noise, it is incomplete, fuzzy, from the data generated in practical application, discovery and extract people cannot be direct
It obtains but there is the information of potential value and the process of knowledge.Here knowledge refers to concept, rule, mode, rule etc.,
There is actual application value in specific field, and can be understood by user.Data mining task is generally divided into two major classes: pre-
Survey task and description task.And common data mining algorithm is broadly divided into: cluster result algorithm, classification algorithm for datamining and
Correlation rule (frequent item set) mining algorithm etc..
Cluster result algorithm can be found that the accumulation phenomena in large data, and is subject to quantitative description；Classified excavation is calculated
Method refers to that data tuple according to existing disaggregated model, is divided into them different classifications；Correlation rule (frequent item set) excavates
Algorithm is used to describe potentially to contact in data attribute.
The existing patent or document closely related with the engineering truck failure prediction method based on data mining mainly have:
Patent " a kind of train groups prognostic and health management ground intelligent processing system and the method " (date of application
2017.09.12 publication number CN107697107A).Which disclose a kind of failure prediction methods for train groups, still
There are essential distinctions with failure prediction method of the invention: the patent is directed to the failure predication of train groups, and needle of the present invention
Pair be engineering truck；The patent is the model analyzed slave data access to data an of entirety, and data mining is only at this
The data analysis component of model accounts for very small part (and being the option of the model), and the present invention is one and is completely based on
The fault prediction model of data mining.Patent " System, method, and computer program product for
Fault prediction in vehicle monitoring and reporting the system " (date of application
2004.11.10, publication number US7230527B2) also disclose a kind of failure prediction method for vehicle, but the patent is public
The method opened and the data digging method that failure prediction method of the invention uses are completely different.
Paper " Hu C, He R, Wang R, et al. Fault Prediction and Fault-Tolerant
of Lithium-ion Batteries Temperature Failure for Electric Vehicle[C]// Third
International Conference on Digital Manufacturing & Automation. IEEE Computer
Society, 2012:410-413. " devise a kind of side of lithium battery failure for predicting electric car using machine learning algorithm
Method；Paper " Ferreira J, Monteiro V, Afonso J L. Data Mining Approach for Range
Prediction of Electric Vehicle [J] Mar-2012,2012. " using regression model for electric car away from
From being predicted.Although method that above-mentioned paper all employs data mining predicts vehicle, but with the present invention
Failure prediction method application scenarios have very big difference.
Summary of the invention
The purpose of the present invention is provide a kind of engineering truck failure predication based on data mining in view of the above shortcomings
System and its prediction technique, in order to carry out effectively failure predication to engineering truck, the present invention utilizes the association in data mining
Regular (frequent item set) excavates and classification algorithm for datamining, devises a kind of engineering truck failure predication side based on data mining
Method is excavated the rule to break down from a large amount of irregular work information and is carried out according to the working condition of engineering truck
Failure predication.
The present invention adopts the following technical solutions to achieve:
A kind of engineering truck failure prediction system based on data mining, including the excavation failure correlation frequent item set being sequentially connected
Module establishes decision forest module and failure predication module；
The excavation failure correlation frequent item set module, for engineering truck fault data being pre-processed and being obtained frequent episode
Collection；
It is described to establish decision forest module, for being carried out again to the frequent item set obtained in excavation failure correlation frequent item set module
Processing, and form decision forest；
The failure predication module for carrying out failure predication to the decision forest for establishing the acquisition of decision forest module, and exports
As a result.
Excavating failure correlation frequent item set module includes data prediction part and acquisition frequent item set part；
The data prediction part is used to be directed to each engineering truck fault data, and corresponding failure is obtained in floor data
The engineering truck operating condition at moment records non-zero and non-empty attribute-name in operating condition, will be non-in the operating condition of each engineering truck fault moment
Zero is considered as a record with non-empty attribute set, carries out subsequent frequent item set mining for obtaining frequent item set part；It is described non-
Zero and non-empty attribute-name include engine speed, fuel oil temperature etc.；
The acquisition frequent item set part is used to record pretreated fault condition as input, obtains frequent item set；Tool
Body is to establish frequent pattern tree (fp tree) using FP-Growth algorithm；And minimum support is set, obtain the item for reaching minimum support
Collect (operating condition attribute) set, later by setting frequent item set minimum attribute number threshold value, deletes the very few item collection of element number, most
The frequent item set for deleting content redundancy afterwards, obtains multiple groups frequent item set.
Decision forest module is established to include data reprocessing part and establish decision tree part；
Frequent item set of the data reprocessing part for by obtaining in excavating failure correlation frequent item set module obtains
Whole training set, and according to the different attribute in frequent item set, obtained from whole training set the frequent item set correspond to attribute with
The correspondence training set of property trends；
The frequent item set obtained in excavating failure correlation frequent item set module is specifically done into union, obtains the relevant category of institute
Property, for each relating attribute, processing obtains failure engineering truck from engineering truck floor data and engineering truck fault data
Variation tendency before the attribute value of sliding-model control and each attribute break down when breaking down marks " can break down ",
The engineering truck floor data that multiple groups (total 5 ~ 10 times for the vehicle fleet that breaks down) do not break down is randomly selected simultaneously, is obtained
Corresponding attribute value and attribute change trend are taken, is labeled as " will not break down ", obtains whole training set.Further according to every group of frequency
Different attribute in numerous item collection obtains the corresponding training that the frequent item set corresponds to attribute (with property trends) from whole training set
Collection；
The decision tree part of establishing is used to establish decision tree and decision forest；
Particularly minimum purity threshold value (0.5 ~ 1.0) is set, attribute corresponding for each group of frequent item set and attribute become
Gesture chooses the related training set that data reprocessing obtains, establishes C4.5 decision tree, in order to ensure C4.5 decision tree height is unlikely to
It is excessively high, when the record purity (" can break down " or " will not break down ") in decision tree nodes be greater than minimum purity threshold value,
Then no longer divide.The corresponding decision tree of every group of frequent item set, all decision trees form decision forest.
Failure predication module includes ballot classifier part and result output par, c, and ballot classifier part is used for input
Engineering car state carry out classification prediction；As a result output par, c is used to export ballot classifier part.
Particularly using the method for ballot, for the engineering car state of input, using the decision tree of random forest, to work
The failure risk of journey vehicle carries out classification prediction, if " can break down " occupies the majority, then it is assumed that engineering truck, which has, at this time breaks down
Risk, otherwise it is assumed that engineering truck works in this state to break down.
If it is the state that will not break down that ballot classifier result, which thinks engineering truck currently, result is directly exported；If recognizing
There is the risk to break down for engineering truck, while exporting result, notifies the use of engineering truck operator suspension of engineering work vehicle.
A kind of engineering truck failure prediction method based on data mining, includes the following steps:
1) basic data of engineering truck is pre-processed by excavating failure correlation frequent item set module, obtains fault car
Attribute；The basic data includes engineering truck floor data and fault data；
2) it excavates the fault car attribute that failure correlation frequent item set module is obtained by step 1) and further obtains frequent item set；
3) it establishes decision forest module and data reprocessing is carried out to the frequent item set obtained in step 2, construct training set, go forward side by side
One step forms decision tree and corresponding decision forest；
4) decision tree that failure predication module is obtained using step 3) carries out classification prediction to the failure risk of engineering truck, and
Export result.
The engineering truck failure prediction method based on data mining that the present invention designs has following 3 advantages:
1, failure correlation frequent item set module is excavated, can be obtained from a large amount of irregular engineering truck work information and failure
Relevant rule occurs.
One very big challenge of engineering truck failure predication is to be difficult to effectively utilize magnanimity engineering truck regime history number
According to how efficiently extracting these correlations and belong to since floor data dimension is very high but relevant to failure attribute is relatively fewer
Property becomes a urgent problem needed to be solved.The excavation failure correlation frequent item set module that the present invention designs passes through frequent item set
(correlation rule) mining algorithm can successfully excavate attribute relevant to failure from the floor data of fault car, make
For the decision attribute of follow-up decision tree prediction.
2, the failure that existing engineering truck work information prediction may be found that can be efficiently used.
The each group of failure association attributes set excavated all can establish a decision tree as one group of decision attribute, i.e.,
Each decision tree is all related to breaking down, and this method carries out failure predication using the decision forest that all decision trees form,
Final result is obtained in the form for classification of voting and is returned in time, if ballot classifier result thinks that engineering truck breaks down
Risk, while exporting result, notify engineering truck operator suspension of engineering work vehicle use, engineering truck can be effectively reduced
The generation of failure.
3, efficient quick failure prediction method.
The present invention is made of three modules: being excavated failure correlation frequent item set module, is established decision forest module and failure
Prediction module.Wherein, the first two module need to only execute once, not need the execution in prediction.In failure predication module, it is based on
Established decision forest inputs the current working data of engineering truck, obtains the prediction result of every decision tree, passes through ballot point
The method of class obtains final result；And all decision trees in decision forest are mutually indepedent, can execute every decision tree parallel
Prediction, while setting minimum purity threshold value for decision tree, define the size of decision tree, reduce holding for every decision tree
The row time.So that each failure predication can achieve target rapidly and efficiently.
Detailed description of the invention
Below with reference to attached drawing, the invention will be further described:
Fig. 1 is the comprising modules schematic diagram of the engineering truck failure prediction system the present invention is based on data mining；
Fig. 2 is the implementation flow chart of the acquisition frequent item set part of present system in embodiment；
Fig. 3 is that present system establishes decision tree part implementation flow chart in embodiment；
Fig. 4 is the implementation flow chart of the failure predication module of present system in embodiment.
Specific embodiment
Below will be by embodiment, in conjunction with attached drawing, the invention will be further described.
Referring to Fig.1, the engineering truck failure prediction system based on data mining, the excavation failure including being sequentially connected are related
Frequent item set module establishes decision forest module and failure predication module；Excavating failure correlation frequent item set module includes data
Preprocessing part and acquisition frequent item set part；Decision forest module is established to include data reprocessing part and establish decision tree portion
Point；Failure predication module includes ballot classifier part and result output par, c.
The failure prediction method of this system includes the following steps:
1) basic data of engineering truck is pre-processed by excavating failure correlation frequent item set module, obtains fault car
Attribute；The basic data includes engineering truck floor data and fault data；Particularly according to engineering truck ID and generation event
Downtime obtains operating condition when engineering truck breaks down, and deletes the attribute in work information and is 0 or is empty attribute, and deletes
Except the attribute (such as time, longitude and latitude) unrelated with fault message is determined, to each fault condition information, related belong to is extracted
Property, the input as frequent item set mining in step 2；
2) it excavates the fault car attribute that failure correlation frequent item set module is obtained by step 1) and further obtains frequent item set；
Frequent item set part specific embodiment process is obtained referring to attached drawing 2, comprising the following steps:
Frequent pattern tree (fp tree) FP-Tree 2-1) is established according to the fault car attribute that step 1) data prediction obtains；
The minimum support of frequent item set 2-2) is set, obtains item collection (operating condition attribute) set for reaching minimum support, later
By the way that frequent item set minimum attribute number threshold value is arranged；
The frequent item set (attribute set relevant to failure) for reaching support 2-3) is obtained using FP-Growth algorithm；
2-4) delete the very few item collection of frequent item set interior element number, finally delete content redundancy frequent item set (such as item collection a,
B, c } with item collection { a, b, c, d } meet the requirement of frequent item set, delete item collection { a, b, c }), obtain multiple groups frequent item set.
3) it establishes decision forest module and data reprocessing is carried out to the frequent item set obtained in step 2；
3-1) construct training set；
Constructing training set, specific step is as follows:
The frequent item set obtained in excavating failure correlation frequent item set module 3-1-1) is done into union, obtains the relevant category of institute
Property；
Meaningless attribute in above-mentioned relating attribute 3-1-2) is deleted, and obtains all important fault attributes.Wherein:
Meaningless attribute refers in all engineering truck operating conditions, a kind of attribute of attribute value only occurs.Such as operating status → { fortune
Row stops }, when engineering truck work, state must be " RUN ", which is meaningless for failure excavation.
Important fault attribute: on the basis of all relating attributes, the attribute set of meaningless attribute is removed.
Attribute (such as oil temperature, coolant temperature with variation tendency 3-1-3) are chosen in above-mentioned important fault attribute
Deng), to the engineering truck that each breaks down, the variation that they correspond to attribute before breaking down is obtained according to floor data
Trend, and to attribute value sliding-model control when breaking down, it is labeled as " can break down "；It is (total that multiple groups are randomly selected simultaneously
Number is to break down 5 ~ 10 times of vehicle fleet) the engineering truck floor data that does not break down, it obtains corresponding attribute value and belongs to
Property variation tendency, be labeled as " will not break down ", obtain whole training set；
3-1-4) according to the different attribute in every group of frequent item set, the correspondence of the frequent item set is obtained from above-mentioned whole training set
Attribute (for changing the attribute of trend, obtains corresponding attribute value and variation tendency；Remaining attribute obtains attribute value), it establishes
Corresponding training set.
It 3-2) establishes decision forest module and further forms decision tree and corresponding decision forest；
Decision tree (decision tree) algorithm is based on characteristic attribute and classifies, main advantage: model has readable
Property, calculation amount is small, and classification speed is fast.Decision Tree algorithms include the ID3 and C4.5 proposed by Quinlan, and Breiman etc. is proposed
CART.Wherein, C4.5 is to be made that improvement to the objective function of Split Attribute based on ID3.
The specific steps of decision tree and corresponding decision forest are formed referring to attached drawing 3, are specifically included that
Minimum purity threshold value 3-2-1) is set, and minimum purity threshold value is between 0.5 ~ 1.0；
3-2-2) such as there is untreated frequent item set, then chooses the corresponding trend attribute of one group of untreated frequent item set
Decision attribute as decision tree；Untreated frequent item set is such as not present, then enters step 3-2-4)；
3-2-3) input the corresponding decision attribute of above-mentioned frequent item set and corresponding training dataset；And it is corresponding to establish frequent item set
C4.5 decision tree, in order to ensure C4.5 decision tree height is unlikely to excessively high, when the record purity in decision tree nodes (" can be sent out
Raw failure " or " will not break down ") be greater than step 3-2-1) the minimum purity threshold value of setting when, then no longer divide, decision tree
Sum increase by 1；The corresponding decision tree of every group of frequent item set, all decision trees form decision forest；
3-2-4) record decision tree sum, the decision forest that all decision trees are formed a whole；The step terminates.
4) decision tree and decision forest that failure predication module is obtained using step 3), to the failure risk of engineering truck into
Row classification prediction, and export result；
Referring to attached drawing 4, detailed process step includes: the embodiment process of failure predication module
4-1) work condition state of input engineering truck and each attribute change trend；
4-2) processing step 4-1) in engineering truck work condition state and each attribute change trend data information, be allowed to meet every
Decision tree input requirements namely engineering truck select the corresponding attribute value of engineering truck for each decision tree in decision forest；
4-3) each decision tree exports the prediction result of this decision tree to ballot classifier；
4-4) when all decision trees prediction finish, to " can break down " in decision tree result with the sum of " will not break down "
It is counted, the result to occupy the majority is the final result of decision forest；
If 4-4-1) final result is " will not break down ", then it is assumed that engineering truck is currently without breaking down state, and directly
It returns the result；
If 4-4-2) final result is " can break down ", then it is assumed that engineering truck has the risk to break down, what is returned the result
Meanwhile notifying the use of engineering truck operator suspension of engineering work vehicle.
Pretreated process includes in step 1), according to engineering truck identifier ID and down time, obtains engineering
Operating condition when vehicle breaks down deletes the attribute in work information and is 0 or is empty attribute, and deletes determining and fault message
Unrelated attribute (such as time, longitude and latitude) extracts dependent failure vehicle attribute, as step to each fault condition information
It is rapid 2) in frequent item set mining input.
Claims (10)
1. a kind of engineering truck failure prediction system based on data mining, it is characterised in that: including the excavation event being sequentially connected
Hinder related frequent item set module, establish decision forest module and failure predication module；
The excavation failure correlation frequent item set module, for engineering truck fault data being pre-processed and being obtained frequent episode
Collection；
It is described to establish decision forest module, for being carried out again to the frequent item set obtained in excavation failure correlation frequent item set module
Processing, and form decision forest；
The failure predication module for carrying out failure predication to the decision forest for establishing the acquisition of decision forest module, and exports
As a result.
2. the engineering truck failure prediction system according to claim 1 based on data mining, it is characterised in that: excavate event
Hindering related frequent item set module includes data prediction part and acquisition frequent item set part；
The data prediction part is used to be directed to each engineering truck fault data, and corresponding failure is obtained in floor data
The engineering truck operating condition at moment records non-zero and non-empty attribute-name in operating condition, will be non-in the operating condition of each engineering truck fault moment
Zero is considered as a record with non-empty attribute set, carries out subsequent frequent item set mining for obtaining frequent item set part；
The acquisition frequent item set part is used to record pretreated fault condition as input, obtains frequent item set.
3. the engineering truck failure prediction system according to claim 1 based on data mining, it is characterised in that: foundation is determined
Plan forest module includes data reprocessing part and establishes decision tree part；
Frequent item set of the data reprocessing part for by obtaining in excavating failure correlation frequent item set module obtains
Whole training set, and according to the different attribute in frequent item set, obtained from whole training set the frequent item set correspond to attribute with
The correspondence training set of property trends；
The decision tree part of establishing is used to establish decision tree and decision forest.
4. the engineering truck failure prediction system according to claim 1 based on data mining, it is characterised in that: failure is pre-
Surveying module includes ballot classifier part and result output par, c, ballot classifier part be used for the engineering car state of input into
Row classification prediction；As a result output par, c is used to export ballot classifier part.
5. a kind of engineering truck failure prediction method based on data mining, which comprises the steps of:
1) basic data of engineering truck is pre-processed by excavating failure correlation frequent item set module, obtains fault car
Attribute；The basic data includes engineering truck floor data and fault data；
2) it excavates the fault car attribute that failure correlation frequent item set module is obtained by step 1) and further obtains frequent item set；
3) it establishes decision forest module and data reprocessing is carried out to the frequent item set obtained in step 2, construct training set, go forward side by side
One step forms decision tree and corresponding decision forest；
4) decision tree that failure predication module is obtained using step 3) carries out classification prediction to the failure risk of engineering truck, and
Export result.
6. a kind of engineering truck failure prediction method based on data mining according to claim 5, it is characterised in that: step
It is rapid 1) in pretreated process include, according to engineering truck identifier ID and down time, obtaining engineering truck and breaking down
When operating condition, delete the attribute in work information and be 0 or be empty attribute, and delete and determine the category unrelated with fault message
Property, to each fault condition information, extract dependent failure vehicle attribute, the input as frequent item set mining in step 2.
7. a kind of engineering truck failure prediction method based on data mining according to claim 5, which is characterized in that obtain
Take the specific process in frequent item set part the following steps are included:
Frequent pattern tree (fp tree) FP-Tree 2-1) is established according to the fault car attribute that step 1) data prediction obtains；
The minimum support of frequent item set 2-2) is set, the item collection set for reaching minimum support is obtained, passes through setting frequency later
Numerous item collection minimum attribute number threshold value；
The frequent item set for reaching support 2-3) is obtained using FP-Growth algorithm；
The very few item collection of frequent item set interior element number 2-4) is deleted, the frequent item set of content redundancy is finally deleted, obtains multiple groups
Frequent item set.
8. a kind of engineering truck failure prediction method based on data mining according to claim 5, which is characterized in that step
Rapid detailed process 3) are as follows:
3-1) construct training set；
Constructing training set, specific step is as follows:
The frequent item set obtained in excavating failure correlation frequent item set module 3-1-1) is done into union, obtains the relevant category of institute
Property；
Meaningless attribute in above-mentioned relating attribute 3-1-2) is deleted, and obtains all important fault attributes；
The attribute with variation tendency, the engineering break down to each 3-1-3) are chosen in above-mentioned important fault attribute
Vehicle obtains the variation tendency that they correspond to attribute before breaking down according to floor data, and to attribute when breaking down
It is worth sliding-model control, is labeled as " can break down "；Multiple groups are randomly selected simultaneously, and group number sum is the vehicle fleet that breaks down
5 ~ 10 times, the engineering truck floor data not broken down obtains corresponding attribute value and attribute change trend, and being labeled as " will not
Break down ", obtain whole training set；
3-1-4) according to the different attribute in every group of frequent item set, the corresponding category of the frequent item set is obtained from above-mentioned whole training set
Property, establish corresponding training set；
It 3-2) establishes decision forest module and further forms decision tree and corresponding decision forest；
Composition decision tree and the specific steps of corresponding decision forest include,
Minimum purity threshold value 3-2-1) is set；
3-2-2) such as there is untreated frequent item set, then chooses the corresponding trend attribute of one group of untreated frequent item set
Decision attribute as decision tree；Untreated frequent item set is such as not present, then enters step 3-2-4)；
3-2-3) input the corresponding decision attribute of above-mentioned frequent item set and corresponding training dataset；And it is corresponding to establish frequent item set
C4.5 decision tree, in order to ensure C4.5 decision tree height is unlikely to excessively high, when the record purity in decision tree nodes (" can be sent out
Raw failure " or " will not break down ") be greater than step 3-2-1) the minimum purity threshold value of setting when, then no longer divide, decision tree
Sum increase by 1；The corresponding decision tree of every group of frequent item set, all decision trees form decision forest；
3-2-4) record decision tree sum, the decision forest that all decision trees are formed a whole；The step terminates.
9. a kind of engineering truck failure prediction method based on data mining according to claim 8, which is characterized in that step
Rapid 3-1-2) described in meaningless attribute refer in all engineering truck operating conditions a kind of attribute of attribute value only occur；It is important
Fault attribute refers on the basis of all relating attributes, removes the attribute set of meaningless attribute.
10. a kind of engineering truck failure prediction method based on data mining according to claim 5, which is characterized in that
The detailed process step of failure predication module includes:
4-1) work condition state of input engineering truck and each attribute change trend；
4-2) processing step 4-1) in engineering truck work condition state and each attribute change trend data information, be allowed to meet every
Decision tree input requirements namely engineering truck select the corresponding attribute value of engineering truck for each decision tree in decision forest；
4-3) each decision tree exports the prediction result of this decision tree to ballot classifier；
4-4) when all decision trees prediction finish, to " can break down " in decision tree result with the sum of " will not break down "
It is counted, the result to occupy the majority is the final result of decision forest；
If 4-4-1) final result is " will not break down ", then it is assumed that engineering truck is currently without breaking down state, and directly
It returns the result；
If 4-4-2) final result is " can break down ", then it is assumed that engineering truck has the risk to break down, what is returned the result
Meanwhile notifying the use of engineering truck operator suspension of engineering work vehicle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810556980.1A CN108958215A (en) | 2018-06-01 | 2018-06-01 | A kind of engineering truck failure prediction system and its prediction technique based on data mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810556980.1A CN108958215A (en) | 2018-06-01 | 2018-06-01 | A kind of engineering truck failure prediction system and its prediction technique based on data mining |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108958215A true CN108958215A (en) | 2018-12-07 |
Family
ID=64492974
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810556980.1A Pending CN108958215A (en) | 2018-06-01 | 2018-06-01 | A kind of engineering truck failure prediction system and its prediction technique based on data mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108958215A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109753800A (en) * | 2019-01-02 | 2019-05-14 | 重庆邮电大学 | Merge the Android malicious application detection method and system of frequent item set and random forests algorithm |
| CN112213571A (en) * | 2019-07-11 | 2021-01-12 | 华北电力大学（保定） | Method for constructing fault feature library of electric vehicle charging facility |
| CN112380274A (en) * | 2020-11-16 | 2021-02-19 | 北京航空航天大学 | Control process-oriented anomaly detection system |
| CN112598287A (en) * | 2020-12-25 | 2021-04-02 | 武汉烽火信息集成技术有限公司 | Rail transit health assessment method and system |
| CN112884179A (en) * | 2021-03-30 | 2021-06-01 | 北京交通大学 | Urban rail turn-back fault diagnosis method based on machine fault and text topic analysis |
| CN112990592A (en) * | 2021-03-26 | 2021-06-18 | 广东工业大学 | Shared vehicle fault prediction method and system |
| CN113515560A (en) * | 2021-07-19 | 2021-10-19 | 彩虹无线（北京）新技术有限公司 | Vehicle fault analysis method and device, electronic equipment and storage medium |
| CN113642893A (en) * | 2021-08-16 | 2021-11-12 | 公安部交通管理科学研究所 | New energy automobile operation risk assessment method |
| CN115878421A (en) * | 2022-12-09 | 2023-03-31 | 国网湖北省电力有限公司信息通信公司 | Data center equipment-level fault prediction method, system and medium based on log time sequence correlation characteristic mining |
| CN116974265A (en) * | 2023-07-11 | 2023-10-31 | 武汉理工大学 | Underground mine car fault diagnosis method and system under no-signal scene |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103257921A (en) * | 2013-04-16 | 2013-08-21 | 西安电子科技大学 | Improved random forest algorithm based system and method for software fault prediction |
| CN103310051A (en) * | 2013-06-05 | 2013-09-18 | 天泽信息产业股份有限公司 | Method for predicting failure rate of vehicle-mounted information terminal in entire life cycle |
| CN106066184A (en) * | 2015-04-24 | 2016-11-02 | 波音公司 | For detecting the system and method for communication tool system fault |
| CN106168799A (en) * | 2016-06-30 | 2016-11-30 | 常伟 | A kind of method carrying out batteries of electric automobile predictive maintenance based on big data machine learning |
| US9971346B2 (en) * | 2010-05-14 | 2018-05-15 | Harnischfeger Technologies, Inc. | Remote monitoring of machine alarms |
| CN108062402A (en) * | 2017-12-27 | 2018-05-22 | 广州市云润大数据服务有限公司 | A kind of event time axis method for digging and system |
-
2018
- 2018-06-01 CN CN201810556980.1A patent/CN108958215A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9971346B2 (en) * | 2010-05-14 | 2018-05-15 | Harnischfeger Technologies, Inc. | Remote monitoring of machine alarms |
| CN103257921A (en) * | 2013-04-16 | 2013-08-21 | 西安电子科技大学 | Improved random forest algorithm based system and method for software fault prediction |
| CN103310051A (en) * | 2013-06-05 | 2013-09-18 | 天泽信息产业股份有限公司 | Method for predicting failure rate of vehicle-mounted information terminal in entire life cycle |
| CN106066184A (en) * | 2015-04-24 | 2016-11-02 | 波音公司 | For detecting the system and method for communication tool system fault |
| CN106168799A (en) * | 2016-06-30 | 2016-11-30 | 常伟 | A kind of method carrying out batteries of electric automobile predictive maintenance based on big data machine learning |
| CN108062402A (en) * | 2017-12-27 | 2018-05-22 | 广州市云润大数据服务有限公司 | A kind of event time axis method for digging and system |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109753800A (en) * | 2019-01-02 | 2019-05-14 | 重庆邮电大学 | Merge the Android malicious application detection method and system of frequent item set and random forests algorithm |
| CN109753800B (en) * | 2019-01-02 | 2023-04-07 | 重庆邮电大学 | Android malicious application detection method and system fusing frequent item set and random forest algorithm |
| CN112213571B (en) * | 2019-07-11 | 2023-01-24 | 华北电力大学（保定） | Method for constructing fault feature library of electric vehicle charging facility |
| CN112213571A (en) * | 2019-07-11 | 2021-01-12 | 华北电力大学（保定） | Method for constructing fault feature library of electric vehicle charging facility |
| CN112380274A (en) * | 2020-11-16 | 2021-02-19 | 北京航空航天大学 | Control process-oriented anomaly detection system |
| CN112380274B (en) * | 2020-11-16 | 2023-08-22 | 北京航空航天大学 | Abnormality detection method for control process |
| CN112598287A (en) * | 2020-12-25 | 2021-04-02 | 武汉烽火信息集成技术有限公司 | Rail transit health assessment method and system |
| CN112990592A (en) * | 2021-03-26 | 2021-06-18 | 广东工业大学 | Shared vehicle fault prediction method and system |
| CN112884179A (en) * | 2021-03-30 | 2021-06-01 | 北京交通大学 | Urban rail turn-back fault diagnosis method based on machine fault and text topic analysis |
| CN112884179B (en) * | 2021-03-30 | 2024-02-09 | 北京交通大学 | Urban rail turn-back fault diagnosis method based on machine fault and text topic analysis |
| CN113515560A (en) * | 2021-07-19 | 2021-10-19 | 彩虹无线（北京）新技术有限公司 | Vehicle fault analysis method and device, electronic equipment and storage medium |
| CN113642893A (en) * | 2021-08-16 | 2021-11-12 | 公安部交通管理科学研究所 | New energy automobile operation risk assessment method |
| CN115878421A (en) * | 2022-12-09 | 2023-03-31 | 国网湖北省电力有限公司信息通信公司 | Data center equipment-level fault prediction method, system and medium based on log time sequence correlation characteristic mining |
| CN115878421B (en) * | 2022-12-09 | 2023-11-14 | 国网湖北省电力有限公司信息通信公司 | Data center equipment level fault prediction method, system and medium |
| CN116974265A (en) * | 2023-07-11 | 2023-10-31 | 武汉理工大学 | Underground mine car fault diagnosis method and system under no-signal scene |
| CN116974265B (en) * | 2023-07-11 | 2024-04-26 | 武汉理工大学 | Underground mine car fault diagnosis method and system under no-signal scene |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108958215A (en) | A kind of engineering truck failure prediction system and its prediction technique based on data mining | |
| CN102765643B (en) | Elevator fault diagnosis and early-warning method based on data drive | |
| CN109343995A (en) | Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot | |
| CN102902752B (en) | Method and system for monitoring log | |
| CN102289569B (en) | Method for emergency treatment of emergency events of power system | |
| CN108364124A (en) | International production capacity Cooperation Risk assessment based on big data and Decision Making Service System | |
| CN110262463A (en) | A kind of rail traffic platform door fault diagnosis system based on deep learning | |
| CN108398934B (en) | equipment fault monitoring system for rail transit | |
| CN106777703A (en) | A kind of bus passenger real-time analyzer and its construction method | |
| CN106649801A (en) | Time-space relationship analysis system based on multi-source internet-of-things position awareness | |
| CN104636751A (en) | Crowd abnormity detection and positioning system and method based on time recurrent neural network | |
| CN114389359A (en) | Intelligent operation and maintenance method of centralized control type relay protection equipment based on cloud edge cooperation | |
| CN112462696A (en) | Intelligent manufacturing workshop digital twin model construction method and system | |
| CN102439584A (en) | Method of establishing a process decision support system | |
| CN112183656B (en) | SCADA data frequent item set mining method in power grid fault | |
| CN114596061B (en) | Project data management method and system based on big data | |
| CN110427298A (en) | A kind of Automatic Feature Extraction method of distributed information log | |
| CN114267173B (en) | Multisource data fusion method, device and equipment for space-time characteristics of expressway | |
| CN106294738A (en) | A kind of Intelligent household scene collocation method | |
| CN102801548B (en) | A kind of method of intelligent early-warning, device and information system | |
| Yeh et al. | Classification knowledge discovery in mold tooling test using decision tree algorithm | |
| US20170236071A1 (en) | Alarm management system | |
| Patil | Machine Learning for Traffic Management in Large-Scale Urban Networks: A Review | |
| CN103391211A (en) | Device and method for executing configuration management automatically according to strategies | |
| Li et al. | Research on construction of crude set model of critical fault information for bus based on can-bus data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181207 |