CN109839917A - A kind of Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting - Google Patents

Abstract

The invention discloses a kind of Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting, which is made of sensing module, fault diagnosis module, diagnostic result display instrument.Sensing module measures each failure symptom data of coalcutter, and these data are passed to fault diagnosis module；Fault diagnosis module identifies failure cause according to failure symptom data intelligence, and result transmission diagnostic result display instrument is shown.The present invention overcomes current coalcutter intelligent failure diagnosis method can not real-time update model deficiency, use gradient boosted tree GBDT as classifier, diagnostic accuracy is high；Model is automatically updated using adaptively correcting strategy, diagnostic model is adjusted once discovery diagnostic result is deteriorated, maintains the diagnostic accuracy of system.

Description

A kind of Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting

Technical field

The present invention relates to coal production field more particularly to a kind of Malfunction Diagnosis for Coal-Mining Machine systems of adaptively correcting.

Background technique

The exploitation of coal is a complicated system engineering, and the equipment being related to is also many kinds of.With comprehensive mechanical Change and mine universal, role just seems the most prominent to fully-mechanized mining equipment wherein, and in fully-mechanized mining equipment it is most important not Excessively fully mechanized mining " three machines ": hydraulic support, drag conveyor, coalcutter.Hydraulic support and drag conveyor construct phase in " three machines " To simple, the frequency of failure is also relatively fewer and failure is plain, and experienced check man can quickly find the problem, solve Problem will not generally be brought greater impact to coal work.But coalcutter is just different, is used to mine as fully-mechanized mining working " both hands ", construction is complicated, and the risk point of failure is more, and failure has concealment and diversity.If there is event Barrier will will lead to entire coal work and interrupt, and cause serious loss.The failure rate of coalcutter how is reduced, work effect is improved Rate, this is also one of the important problem for perplexing coal mining enterprise and respectively producing mine.

In the 1960s, with the development of computer technology and electronic technology, it is external some national by fault diagnosis skill Art is applied on coalcutter, and fault diagnosis technology also from the sensible judgments technologies such as the sense of hearing, tactile are used merely, develops to portable Diagnostic instrments, then arrive present self diagnosis.It is more accurate, quick that this makes the fault diagnosis of coalcutter, and greatly improves Coal production benefit.But at present coalcutter method for diagnosing faults often once foundation once can not online updating, affect The validity of actual use and the accuracy of diagnostic result.

Summary of the invention

In order to overcome the shortcomings of current coalcutter intelligent failure diagnosis method can not real-time update model, the purpose of the present invention It is to provide a kind of Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting, which is classified based on gradient boosted tree GBDT Device, and introduce the Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting strategy.

The purpose of the present invention is achieved through the following technical solutions: a kind of Malfunction Diagnosis for Coal-Mining Machine of adaptively correcting System, the system are made of sensing module, fault diagnosis module, diagnostic result display instrument.The effect and connection type of each section Are as follows: sensing module measures each failure symptom data of coalcutter, and these data are passed to fault diagnosis module；Fault diagnosis mould Root tuber identifies failure cause according to failure symptom data intelligence, and result transmission diagnostic result display instrument is shown.

Further, the failure symptom data X=(X1, X2, X3 ..., X9) of sensing module measurement coalcutter, and by data It is transferred to fault diagnosis module.Wherein: X1 indicates oil compensation pressure when coalcutter zero load；X2 indicates repairing when coalcutter load Pressure；X3 indicates auxiliary system pressure；X4 indicates the total feed liquor flow of hydraulic motor and total difference for returning flow quantity；X5 indicates rocker arm liter Play the time；X6 indicates current of electric；X7 indicates motor temperature；X8 indicates cooling water pressure；X9 indicates cooling water flow.Failure is former Because integrating as Y=(Y1, Y2, Y3 ..., Y7), wherein Y1 indicates main pump failure；Y2 indicates repairing failure of pump；Y3 indicates oil filter event Barrier；Y4 indicates auxiliary failure of pump；Y5 indicates hydraulic motor failure；Y6 indicates motor overload；Y7 indicates cooling system failure.It is each One or more failures in corresponding seven failure causes of Y1 to Y7 of a failure symptom data X=(X1, X2, X3 ..., X9) are former Cause.

Further, fault diagnosis module uses gradient boosted tree GBDT as classifier, and uses adaptively correcting plan Slightly automatically update model.The foundation of fault diagnosis module and operating procedure are as follows:

(1) input of GBDT classifier is X=(X1, X2, X3 ..., X9), is exported as corresponding failure cause.To own Sample with complete inputoutput pair is divided into training set and verifying collects, and is trained, obtains in training set input GBDT classifier The fault diagnosis model completed to training, and calculate the classification accuracy accuracy_val of verifying collection.

(2) the data X=(X1, X2, X3 ..., X9) obtained the sensing module measurement of unknown classification results is input to instruction Practice the fault diagnosis model completed, analysis obtains specific failure cause, then result is passed to diagnostic result display instrument and is shown Show.

(3) it checks coalcutter failure on the spot according to diagnostic result, looks for out of order true cause.This result is added to survey Examination collection.When the quantity of test set is more than the half of verifying collection quantity, the classification accuracy accuracy_ of test set is calculated test。

(4) if accuracy_test < 0.95 × accuracy_val, test set sample is added to training set, again Training pattern.

(5) step (1)-(4) are repeated, the adaptively correcting of Malfunction Diagnosis for Coal-Mining Machine system is realized.

Beneficial effects of the present invention are mainly manifested in: the present invention uses gradient boosted tree GBDT as classifier, and diagnosis is quasi- True property is high；Model is automatically updated using adaptively correcting strategy, diagnostic model is adjusted once discovery diagnostic result is deteriorated, maintains The diagnostic accuracy of system.

Detailed description of the invention

Fig. 1 is structural schematic diagram of the invention.

Fig. 2 is foundation and the operational process of fault diagnosis module of the present invention.

Specific embodiment

The present invention is illustrated below according to attached drawing.

Referring to Fig.1, the Malfunction Diagnosis for Coal-Mining Machine system of a kind of adaptively correcting, the system is by sensing module 2, fault diagnosis Module 3, diagnostic result display instrument 6 form.The effect and connection type of each section are as follows: sensing module 2 measures each event of coalcutter 1 Hinder sign variable, and these data are passed into fault diagnosis module 3；Fault diagnosis module 3 is according to failure symptom data intelligence It identifies failure cause, and result transmission diagnostic result display instrument 6 is shown.

Sensing module 2 measures the failure symptom data X=(X1, X2, X3 ..., X9) of coalcutter 1, and transfers data to Fault diagnosis module 3.Wherein: X1 indicates oil compensation pressure when coalcutter zero load；X2 indicates oil compensation pressure when coalcutter load； X3 indicates auxiliary system pressure；X4 indicates the total feed liquor flow of hydraulic motor and total difference for returning flow quantity；When X5 indicates that rocker arm rises Between；X6 indicates current of electric；X7 indicates motor temperature；X8 indicates cooling water pressure；X9 indicates cooling water flow.Failure cause collection For Y=(Y1, Y2, Y3 ..., Y7), wherein Y1 indicates main pump failure；Y2 indicates repairing failure of pump；Y3 indicates oil filter failure；Y4 Indicate auxiliary failure of pump；Y5 indicates hydraulic motor failure；Y6 indicates motor overload；Y7 indicates cooling system failure.Each event Hinder one or more failure causes in corresponding seven failure causes of Y1 to Y7 of sign variable X=(X1, X2, X3 ..., X9).

Fault diagnosis module 3 carries out fault diagnosis using gradient boosted tree GBDT classifier 5, and uses adaptively correcting plan Slightly 4 automatically update model.Referring to Fig. 2, the foundation of fault diagnosis module 3 and operating procedure are as follows:

(1) input of GBDT classifier 5 is X=(X1, X2, X3 ..., X9), is exported as corresponding failure cause.To own Sample with complete inputoutput pair is divided into training set and verifying collects, and is trained, obtains in training set input GBDT classifier The fault diagnosis model completed to training, and calculate the classification accuracy accuracy_val of verifying collection.

(2) sensing modules 2 of unknown classification results is measured to obtained data X=(X1, X2, X3 ..., X9) and is input to instruction Practice the fault diagnosis model completed, analysis obtains specific failure cause, then result is passed to diagnostic result display instrument 6 and is shown Show.

(3) it checks coalcutter failure on the spot according to diagnostic result, looks for out of order true cause.This result is added to survey Examination collection.When the quantity of test set is more than the half of verifying collection quantity, the classification accuracy accuracy_ of test set is calculated test。

(4) if accuracy_test < 0.95 × accuracy_val, test set sample is added to training set, weight New training pattern.

(5) step (1)-(4) are repeated, the adaptively correcting of Malfunction Diagnosis for Coal-Mining Machine system is realized.

Above-described embodiment is used to illustrate the present invention, rather than limits the invention, in spirit of the invention and In scope of protection of the claims, to any modifications and changes that the present invention makes, protection scope of the present invention is both fallen within.

Claims (3)

1. a kind of Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting, it is characterised in that: the system is by sensing module, fault diagnosis Module, diagnostic result display instrument composition.

2. the Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting according to claim 1, it is characterised in that: the sensing module The failure symptom data X=(X1, X2, X3 ..., X9) of coalcutter is measured, and transfers data to fault diagnosis module.Wherein: X1 indicates oil compensation pressure when coalcutter zero load；X2 indicates oil compensation pressure when coalcutter load；X3 indicates auxiliary system pressure； X4 indicates the total feed liquor flow of hydraulic motor and total difference for returning flow quantity；X5 indicates rocker arm rise time；X6 indicates current of electric；X7 Indicate motor temperature；X8 indicates cooling water pressure；X9 indicates cooling water flow.Failure cause integrate as Y=(Y1, Y2, Y3 ..., Y7), wherein Y1 indicates main pump failure；Y2 indicates repairing failure of pump；Y3 indicates oil filter failure；Y4 indicates auxiliary failure of pump；Y5 Indicate hydraulic motor failure；Y6 indicates motor overload；Y7 indicates cooling system failure.Each failure symptom data X=(X1, X2, X3 ..., X9) one or more failure causes in corresponding seven failure causes of Y1 to Y7.

3. the Malfunction Diagnosis for Coal-Mining Machine system of adaptively correcting according to claim 1, it is characterised in that: the fault diagnosis Module uses gradient boosted tree GBDT as classifier, and automatically updates model using adaptively correcting strategy.Fault diagnosis mould The foundation of block and operating procedure are as follows:

(1) input of GBDT classifier is X=(X1, X2, X3 ..., X9), is exported as corresponding failure cause.Have all The sample of complete inputoutput pair is divided into training set and verifying collects, and is trained, is instructed in training set input GBDT classifier Practice the fault diagnosis model completed, and calculates the classification accuracy accuracy_val of verifying collection.

(2) the data X=(X1, X2, X3 ..., X9) obtained the sensing module measurement of unknown classification results is input to and has trained At fault diagnosis model, analysis obtains specific failure cause, then result passed to diagnostic result display instrument and is shown.

(3) it checks coalcutter failure on the spot according to diagnostic result, looks for out of order true cause.This result is added to test Collection.When the quantity of test set is more than the half of verifying collection quantity, the classification accuracy accuracy_test of test set is calculated.

(4) if accuracy_test < 0.95 × accuracy_val, test set sample is added to training set, re -training Model.

(5) step (1)-(4) are repeated, the adaptively correcting of Malfunction Diagnosis for Coal-Mining Machine system is realized.