Fault detection method and system

Abstract

The invention discloses a fault detection method and system, and relates to the technical field of diesel engine fault detection. The fault detection method is used for detecting whether the electromagnetic valve driving circuit of the electronic control fuel injector has a fault or not and comprises the steps that a driving current waveform is obtained based on a current feedback signal of the electromagnetic valve driving circuit; acquiring a cylinder number mark corresponding to the electronic control fuel injector; acquiring a detection characteristic quantity based on the driving current waveform and the cylinder number mark; the detection characteristic quantity is configured to include a rise time, an excitation current holding time, a follow current time and a current maximum value; inputting the detection characteristic quantity into a pre-established decision tree classification model to obtain a plurality of classification results; and outputting a fault detection result based on the classification result. Thus, the current feedback signals are classified through the decision tree classification model which is built in advance, online fault detection is conducted on the electromagnetic valve driving circuit based on the classification result, and operation safety of the electronic control fuel injector is guaranteed.

Description

Technical Field

[0001] The present application relates to the technical field of diesel engine fault detection, and in particular to a fault detection method and system. Background Art

[0002] The injector is the core component of the high-pressure common rail diesel engine. By controlling the injection amount and injection timing of the injector, the fuel can be precisely controlled. As one of the core functions of the electronic management system of the high-pressure common rail diesel engine, the solenoid valve drive circuit of the injector controls the working state of the injector by controlling the switch of the solenoid valve.

[0003] Once the solenoid valve drive circuit fails, it may lead to inaccurate injection quantity or inaccurate injection timing, affecting the combustion effect and power output of the fuel, and then causing the performance of the high-pressure common rail diesel engine to decline, fuel economy to reduce, and even cause other failures, which will have an adverse effect on the reliability and continuous operation of the high-pressure common rail diesel engine.

[0004] However, currently, online fault detection of the solenoid valve drive circuit of the electronic fuel injector is mostly performed through hardware circuits, and the detection accuracy is low. Summary of the invention

[0005] An embodiment of the present application provides a fault detection method, which classifies the current feedback signal through a pre-built decision tree classification model, and performs online fault detection based on the classification result, thereby solving the problem of low detection accuracy of the hardware circuit in the prior art for online fault detection of the electronically controlled injector solenoid valve drive circuit; an embodiment of the present application also provides a fault detection system.

[0006] The present application provides a fault detection method for detecting whether there is a fault in a solenoid valve driving circuit of an electronically controlled fuel injector. The fault detection method includes:

[0007] Acquiring a driving current waveform based on a current feedback signal of the solenoid valve driving circuit;

[0008] Obtaining the cylinder number mark corresponding to the electronically controlled fuel injector;

[0009] Based on the driving current waveform and the cylinder number mark, a detection feature quantity is obtained; the detection feature quantity is configured to include a rise time, an excitation current maintenance time, a free flow time and a current maximum value;

[0010] Inputting the detection feature quantity into a pre-built decision tree classification model to obtain multiple classification results;

[0011] Based on the classification result, a fault detection result is output.

[0012] In some embodiments, the decision tree classification model is pre-built by the following steps:

[0013] Acquire full-operation test data of the solenoid valve drive circuit, a fault simulation data set, and a fault label corresponding to each fault data sample; wherein the fault simulation data set includes a plurality of the fault data samples;

[0014] Based on the full-operating condition test data, the fault data samples and the fault labels, an initial data set is acquired;

[0015] Determining a number of judgments; the number of judgments is configured to be an odd number;

[0016] Based on the initial data set and the number of judgments, a decision tree classification model set and a fault identification accuracy rate set are obtained;

[0017] Based on the decision tree classification model set and the fault identification accuracy set, the decision tree classification model is built.

[0018] In some embodiments, the step of obtaining a decision tree classification model set and a fault identification accuracy set based on the initial data set and the number of decisions includes:

[0019] Based on the number of judgments, executing a preset number of loop steps to obtain multiple decision tree classification model subsets and multiple fault identification accuracy subsets; the preset number of times is determined based on the number of judgments;

[0020] Determine a plurality of decision tree classification model subsets as the decision tree classification model set;

[0021] Taking the average value of each component element of a plurality of the fault identification accuracy rate subsets as the fault identification accuracy rate set;

[0022] Wherein, the circulation step comprises:

[0023] Determine a training set and a test set based on the initial data set;

[0024] Establishing the decision tree classification model subset based on the training set;

[0025] Obtain a fault recognition accuracy subset of the test set.

[0026] In some embodiments, the full-operating condition test data is obtained through test simulation, and the fault data samples are obtained through simulation tests.

[0027] In some embodiments, the step of acquiring the driving current waveform based on the current feedback signal of the solenoid valve driving circuit includes:

[0028] The current feedback signal is collected in real time through an analog input circuit;

[0029] The current feedback signal is converted into a digital signal by a fast analog-to-digital converter to obtain the driving current waveform.

[0030] In some embodiments, the step of obtaining the cylinder number mark corresponding to the electronically controlled fuel injector includes:

[0031] Obtaining the gate cylinder number currently driven by the solenoid valve drive circuit;

[0032] The cylinder number flag is determined based on the current feedback signal and the gated cylinder number.

[0033] In some embodiments, the step of acquiring the detection feature quantity based on the driving current waveform and the cylinder number mark includes:

[0034] Performing smoothing and denoising processing on the driving current waveform;

[0035] Acquire key characteristic points in the driving current waveform; the key characteristic points include a current starting point, an excitation process starting point, a current rise end point, an excitation current end point, an excitation process end point, a current end point, and a maximum value point of a current feedback signal;

[0036] The detection feature quantity is acquired based on the key feature point.

[0037] In some embodiments, the key feature points are obtained based on the following steps:

[0038] Obtaining the maximum value point from the driving current waveform;

[0039] Traversing each sampling point between the sampling starting point and the maximum value point of the current feedback signal, obtaining a first sampling point set whose absolute value of the sampling value is less than a first preset range, and determining the last point in the first sampling point set as the current starting point; the first preset range is determined according to the fluctuation of the current feedback signal in the no-current segment;

[0040] Traversing each sampling point between the maximum value point and the sampling end point, obtaining a second sampling point set whose absolute value is less than the first preset range, and determining the first point in the second sampling point set as the current end point;

[0041] Obtaining a third sampling point set in which the absolute value of the deviation between the sampling value and the maximum value is less than a second preset range, taking the first point in the third sampling point set as the current rise end point, and taking the last point in the third sampling point set as the excitation current end point; the second preset range is determined according to the drive current waveform;

[0042] A minimum point is obtained between the end point of the current rise and the end point of the excitation current, and a fourth sampling point set is obtained in which the absolute value of the deviation between the sampling value and the minimum point is less than a third preset range, and the first point in the fourth sampling point set is used as the starting point of the excitation process, and the last point in the fourth sampling point set is used as the end point of the excitation process; the third preset range is determined according to the drive current waveform.

[0043] In some embodiments, the step of outputting a fault detection result based on the classification result includes:

[0044] In response to the number of non-fault results being greater than the number of faulty results in the plurality of classification results, not issuing a fault alarm signal;

[0045] In response to the number of the faulty results being greater than the number of the non-faulty results, obtaining the occurrence frequency of the faulty results in the classification results;

[0046] The fault mode corresponding to the fault result with the highest occurrence frequency is obtained, and a fault alarm signal is issued based on the fault mode.

[0047] In some embodiments, the step of outputting a fault detection result based on the classification result further includes:

[0048] In response to the number of the faulty results with the highest occurrence frequency being multiple, obtaining multiple fault modes corresponding to the multiple faulty results;

[0049] A fault warning signal is issued based on a plurality of the fault modes.

[0050] In some embodiments, it also includes:

[0051] Based on the classification result and the detection feature quantity, the fault simulation data set is updated.

[0052] In some embodiments, the step of updating the fault simulation data set based on the classification result and the detection feature quantity includes:

[0053] Obtaining the latest fault simulation data set;

[0054] Acquire a current decision tree classification model and a current fault identification accuracy set based on the fault simulation data set;

[0055] In response to the current fault identification accuracy set being less than the fault identification accuracy set, the current decision tree classification model is updated to a new decision tree classification model, and the current fault identification accuracy set is updated to a new fault identification accuracy set.

[0056] Accordingly, the embodiment of the present application further provides a fault detection system for executing the fault detection method as described in any one of the above embodiments to detect whether there is a fault in the electronically controlled injector solenoid valve drive circuit, the fault detection system comprising:

[0057] A first acquisition module is configured to acquire a driving current waveform based on a current feedback signal of the solenoid valve driving circuit;

[0058] A second acquisition module is configured to acquire a cylinder number mark corresponding to the electronically controlled fuel injector;

[0059] A third acquisition module is configured to acquire a detection feature quantity based on the drive current waveform and the cylinder number mark; the detection feature quantity is configured to include a rise time, an excitation current maintenance time, a freewheeling time, and a current maximum value;

[0060] A fourth acquisition module is configured to input the detection feature quantity into a pre-built decision tree classification model to obtain multiple classification results;

[0061] The result output module is configured to output a fault detection result based on the classification result.

[0062] Compared with the prior art, a fault detection method in an embodiment of the present application is used to detect whether there is a fault in the solenoid valve drive circuit of an electronically controlled fuel injector. The fault detection method includes: obtaining a drive current waveform based on the current feedback signal of the solenoid valve drive circuit; obtaining the cylinder number mark corresponding to the electronically controlled fuel injector; obtaining a detection feature quantity based on the drive current waveform and the cylinder number mark; the detection feature quantity is configured to include rise time, excitation current maintenance time, continuous flow time and current maximum value; the detection feature quantity is input into a pre-built decision tree classification model to obtain multiple classification results; based on the classification results, the fault detection result is output. In this way, the current feedback signal is classified by the pre-built decision tree classification model, and the solenoid valve drive circuit is online fault detection based on the classification results to ensure the safe operation of the electronically controlled fuel injector.

[0063] It can be understood that, compared with the prior art, the fault detection system provided in the embodiment of the present application has all the technical features and beneficial effects of the above-mentioned fault detection method, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1A first flow chart of a fault detection method provided in an embodiment of the present application;

[0066] Figure 2 A schematic diagram of a driving current waveform in a fault detection method provided in an embodiment of the present application;

[0067] Figure 3 A second flow chart of a fault detection method provided in an embodiment of the present application;

[0068] Figure 4 A schematic diagram of the structure of a decision tree classification model in a fault detection method provided in an embodiment of the present application;

[0069] Figure 5 A schematic diagram of the structure of a fault detection system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0071] See also Figure 1 and Figure 3 , Figure 1 A schematic diagram of a flow chart of a fault detection method provided in an embodiment of the present application is illustrated; Figure 3 The second flow chart of a fault detection method provided by the embodiment of the present application is shown. The embodiment of the present application provides a fault detection method for detecting whether there is a fault in the solenoid valve driving circuit of an electronically controlled fuel injector. The fault detection method includes the following steps 101 to 105.

[0072] Step 101, based on the current feedback signal of the solenoid valve drive circuit, obtain the drive current waveform. Specifically, the current feedback signal directly reflects whether the working state of the solenoid valve is normal. After converting the current feedback signal into a current waveform, it can more intuitively reflect the working state of the solenoid valve drive circuit. Therefore, the present application obtains the drive current waveform to monitor the working state of the solenoid valve drive circuit in real time, so that when an abnormal current waveform is found, it can be repaired or replaced as soon as possible.

[0073] In some embodiments, the step of obtaining the driving current waveform based on the current feedback signal of the solenoid valve drive circuit includes: collecting the current feedback signal in real time through the analog input circuit; converting the current feedback signal into a digital signal through a fast analog-to-digital converter to obtain the driving current waveform. In the embodiment of the present application, by collecting the current feedback signal in real time and converting it into a digital signal, the waveform of the driving current can be obtained, thereby providing detailed information about the working process of the solenoid valve. Further, by analyzing the waveform characteristics of the driving current waveform, such as the rise time, the continuous current time and the maximum current value, it can be detected whether the solenoid valve drive circuit has a fault or abnormality, so as to detect the problem of the solenoid valve drive circuit early and take appropriate maintenance measures to avoid further damage or failure of the solenoid valve and the electronically controlled injector.

[0074] Step 102, obtain the cylinder number mark corresponding to the electronically controlled fuel injector. Specifically, a diesel engine usually has multiple cylinders (mostly 6 cylinders, 8 cylinders, 12 cylinders, 16 cylinders or 20 cylinders), each cylinder is equipped with an electronically controlled fuel injector, each injector has a driving current feedback signal, and the signals of each cylinder may be different, so it is necessary to know which cylinder the currently collected current feedback signal is the signal of, and when a fault alarm is issued, it is also necessary to report which cylinder has a driving fault. By obtaining the cylinder number mark corresponding to the electronically controlled fuel injector, the present application can accurately determine the cylinder corresponding to the currently collected current feedback signal, thereby improving the accuracy of subsequent fault detection and analysis.

[0075] In some embodiments, the step of obtaining the cylinder number mark corresponding to the electronically controlled fuel injector includes: obtaining the gated cylinder number of the solenoid valve drive circuit in the current drive; and determining the cylinder number mark based on the current feedback signal and the gated cylinder number. Specifically, by obtaining the gated cylinder number of the solenoid valve drive circuit in the current drive, the cylinder corresponding to each electronically controlled fuel injector can be accurately determined, thereby associating the working state of the fuel injector with a specific cylinder, and more accurately performing fault diagnosis and analysis. In addition, after obtaining the cylinder number mark corresponding to the electronically controlled fuel injector, it can be recorded for subsequent analysis and comparison, thereby establishing a historical data record for long-term trend analysis and fault diagnosis, and improving design and maintenance strategies.

[0076] Step 103, based on the drive current waveform and the cylinder number mark, the detection feature quantity is obtained; the detection feature quantity is configured to include the rise time, the excitation current maintenance time, the continuous flow time and the current maximum value. Specifically, by extracting the detection feature quantity of the drive current waveform and the cylinder number mark, detailed information about the working process of the injector can be obtained to determine whether there is an abnormality in the injector.

[0077] In some embodiments, based on the drive current waveform and the cylinder number mark, the step of obtaining the detection feature quantity includes: smoothing and denoising the drive current waveform; obtaining key feature points in the drive current waveform; the key feature points include the current starting point I1, the excitation process starting point I2, the current rise end point I3, the excitation current end point I4, the excitation process end point I5, the current end point I6 and the maximum value point Imax of the current feedback signal; and obtaining the detection feature quantity based on the key feature points. Specifically, the present application first performs smoothing and denoising on the drive current waveform to reduce the interference of noise on the positioning of the key feature points, and at the same time obtains the pulse width of the current gate cylinder number to extract more accurate key feature points, thereby reducing errors and improving the accuracy of key feature point extraction, so as to more reliably analyze the drive current waveform. Among them, the calculation method of k-order smoothing and denoising is to take the average value of a total of k sampling points before and after the current sampling point as the value of the current sampling point. Secondly, the present application accurately describes the change process of the solenoid valve drive current by obtaining the key feature points in the drive current waveform, thereby intuitively reflecting the working state of the drive circuit. In addition, various detection feature quantities are calculated and extracted based on key feature points for subsequent fault diagnosis, performance evaluation and comparative analysis, which helps to quickly and accurately detect whether there are abnormalities in the solenoid valve drive circuit.

[0078] See also Figure 2 , Figure 2A schematic diagram of a driving current waveform in a fault detection method provided in an embodiment of the present application is illustrated. In some embodiments, the key feature points are obtained based on the following steps: obtaining a maximum value point from the driving current waveform; traversing each sampling point between the sampling starting point and the maximum value point of the current feedback signal, obtaining a first sampling point set whose absolute value of the sampling value is less than a first preset range, and determining the last point in the first sampling point set as the current starting point; the first preset range is determined according to the fluctuation of the current feedback signal in the no-current segment, for example, the first preset range is -0.05 to 0.05; traversing each sampling point between the maximum value point and the sampling end point, obtaining a second sampling point set whose absolute value is less than the first preset range, and determining the first point in the second sampling point set as the current ending point; obtaining a sampling value whose absolute value of the deviation from the maximum value is less than the second preset range; determining the first point in the second sampling point set as the current ending point ... A third sampling point set of a range is set, the first point in the third sampling point set is used as the end point of the current rise, and the last point in the third sampling point set is used as the end point of the excitation current; the second preset range is determined according to the driving current waveform, for example, the second preset range is -0.01 to 0.01; between the end point of the current rise and the end point of the excitation current, the minimum value point is obtained, and the fourth sampling point set whose absolute value of the deviation between the sampling value and the minimum value point is less than the third preset range is obtained, the first point in the fourth sampling point set is used as the starting point of the excitation process, and the last point in the fourth sampling point set is used as the end point of the excitation process; the third preset range is determined according to the driving current waveform, for example, the third preset range is -0.01 to 0.01. In this way, the present application can accurately determine each key feature point by traversing the sampling points of the driving current waveform and the current feedback signal, and obtaining a specific sampling point set according to the preset range, thereby accurately describing the change process of the driving current and providing a reliable data basis for subsequent fault diagnosis and analysis.

[0079] Step 104: Input the detected feature quantity into the pre-built decision tree classification model to obtain multiple classification results. The present application uses the pre-built decision tree classification model to achieve rapid classification of the detected feature quantity, thereby quickly and accurately determining whether the solenoid valve drive circuit is working normally, and whether there is a fault or abnormality.

[0080] See also Figure 4 , Figure 4The structure diagram of the decision tree classification model in a fault detection method provided by the embodiment of the present application is illustrated; in some embodiments, the decision tree classification model is pre-built by the following steps: obtaining the full-operating test data of the solenoid valve drive circuit, the fault simulation data set and the fault label corresponding to each fault data sample; wherein the fault simulation data set includes multiple fault data samples; based on the full-operating test data, the fault data sample and the fault label, an initial data set is obtained; determining the number of judgments; the number of judgments is configured as an odd number; based on the initial data set and the number of judgments, a decision tree classification model set and a fault identification accuracy set are obtained; based on the decision tree classification model set and the fault identification accuracy set, a decision tree classification model is built. Specifically, the present application can establish a comprehensive initial data set by obtaining the full-operating test data of the solenoid valve drive circuit, the fault simulation data set and the fault label corresponding to each fault data sample. It should be noted that the initial data set contains normal samples and fault samples under different working conditions, thereby providing rich samples and labels for the training and evaluation of the decision tree classification model. Secondly, the number of judgments is configured as an odd number to avoid ties, so that the decision results of the decision tree classification model are more certain. At the same time, based on the initial data set and the number of decisions, multiple decision tree classification models can be trained and the fault recognition accuracy of each model can be calculated to select the best decision tree classification model for fault detection.

[0081] In some embodiments, based on the initial data set and the number of judgments, the step of obtaining a decision tree classification model set and a fault recognition accuracy set includes: based on the number of judgments, executing a preset number of loop steps to obtain multiple decision tree classification model subsets and multiple fault recognition accuracy subsets; the preset number is determined based on the number of judgments; multiple decision tree classification model subsets are determined as decision tree classification model sets; the average value of each component element of multiple fault recognition accuracy subsets is used as the fault recognition accuracy set; wherein the loop step includes: determining a training set and a test set based on the initial data set; establishing a decision tree classification model subset based on the training set; and obtaining a fault recognition accuracy subset of the test set. Specifically, the ratio of the training set to the test set can be determined according to the specific use. In the embodiment of the present application, by executing the preset number of loop steps, multiple decision tree classification model subsets can be obtained, and each decision tree classification model subset is a decision tree classification model obtained by training on the initial data set. That is to say, the decision tree classification model subset is based on the data of the training set, and the mapping relationship between the detection feature quantity and the normal label or the fault label is established, so that the diversity and robustness of the decision tree classification model are increased, and the accuracy of fault recognition is improved. In addition, the fault identification accuracy subset of the present application is an array of n accuracy values, such as {99.1%, 99%, 98.6%, 99.2%, ...}, and one decision tree classification model subset corresponds to one fault identification accuracy subset. The final decision tree classification model set obtained is M∈{M1, M2, ..., Mn}, and the fault identification accuracy set is R∈{R1, R2, ..., Rn}, where M1, M2, ..., Mn are decision tree classification model subsets, and R1, R2, ..., Rn are fault identification accuracy subsets.

[0082] In some embodiments, full-operating condition test data is obtained through test simulation, and fault data samples are obtained through simulation tests. Specifically, test simulation can cover data under various normal operating conditions, and simulation tests can simulate different types of fault conditions, thereby ensuring that the data set includes various possible conditions and improving the generalization ability and adaptability of the decision tree classification model.

[0083] Step 105: output the fault detection result based on the classification result. Specifically, the present application outputs the fault detection result based on the classification result, so as to timely discover and report the fault of the solenoid valve drive circuit, improve the reliability and safety of the electronically controlled fuel injector, and thus ensure the safe operation of the high-pressure common rail diesel engine.

[0084] In some embodiments, based on the classification results, the step of outputting the fault detection results includes: in response to the number of non-fault results being greater than the number of faulty results in multiple classification results, not issuing a fault alarm signal; in response to the number of faulty results being greater than the number of non-faulty results, obtaining the frequency of occurrence of faulty results in the classification results; obtaining the fault mode corresponding to the faulty result with the highest frequency of occurrence, and issuing a fault alarm signal based on the fault mode. Specifically, the present application determines whether to issue a fault alarm signal by comparing the number of non-faulty results and faulty results, thereby reducing the possibility of false alarms and improving the reliability of fault detection. Secondly, in the case of faulty results, the fault mode with the highest frequency of occurrence is obtained to issue an alarm signal to better perform fault diagnosis and processing.

[0085] In some embodiments, based on the classification results, the step of outputting the fault detection results also includes: in response to the number of the most frequently occurring faulty results being multiple, obtaining multiple fault modes corresponding to the multiple faulty results; and issuing a fault alarm signal based on the multiple fault modes. Specifically, when the number of the most frequently occurring faulty results is multiple, there are also multiple corresponding fault modes, and each fault mode describes different types of fault characteristics and manifestations. At this time, issuing a fault alarm signal for all the above fault modes helps to comprehensively consider multiple fault modes, provide more comprehensive and accurate fault alarm information, reduce false positives and false negatives, and improve the accuracy and reliability of fault detection.

[0086] In some embodiments, the fault detection method further includes: updating the fault simulation data set based on the classification results and the detection feature quantity. Specifically, the present application also updates the fault simulation data set based on the classification results and the detection feature quantity to continuously improve and perfect the fault simulation data set to make it more accurate and comprehensive, thereby further improving the training effect of the decision tree classification model and the accuracy of fault detection.

[0087] In some embodiments, the step of updating the fault simulation data set based on the classification result and the detection feature quantity includes: obtaining the latest fault simulation data set; obtaining the current decision tree classification model and the current fault recognition accuracy set based on the fault simulation data set; in response to the current fault recognition accuracy set being less than the fault recognition accuracy set, updating the current decision tree classification model to a new decision tree classification model, and updating the current fault recognition accuracy set to a new fault recognition accuracy set. In this way, by obtaining the latest fault simulation data set, and obtaining the current decision tree classification model and the current fault recognition accuracy set based on the fault simulation data set, and comparing the current fault recognition accuracy set with the fault recognition accuracy set, the decision tree classification model can be iterated and optimized to improve the performance of the decision tree classification model. In addition, by continuously updating the fault simulation data set and iterating the decision tree classification model, the decision tree classification model can be continuously improved and optimized. With the passage of time and feedback from actual applications, the effect and efficiency of the fault detection method are continuously improved, and the detection and diagnosis capabilities of the solenoid valve drive circuit fault are continuously improved.

[0088] A fault detection method in an embodiment of the present application is used to detect whether there is a fault in the solenoid valve drive circuit of an electronically controlled fuel injector. The fault detection method includes: obtaining a drive current waveform based on a current feedback signal of the solenoid valve drive circuit; obtaining a cylinder number mark corresponding to the electronically controlled fuel injector; obtaining a detection feature quantity based on the drive current waveform and the cylinder number mark; the detection feature quantity is configured to include a rise time, an excitation current maintenance time, a continuous current time and a current maximum value; the detection feature quantity is input into a pre-built decision tree classification model to obtain multiple classification results; and based on the classification result, a fault detection result is output. In this way, the current feedback signal is classified through the pre-built decision tree classification model, and the solenoid valve drive circuit is detected online based on the classification result. It can realize online diagnosis of ten typical faults of the solenoid valve drive circuit, including high-end high-voltage MOS tube open circuit, high-end high-voltage MOS tube short circuit, high-end low-voltage MOS tube open circuit, high-end low-voltage MOS tube short circuit, low-end MOS tube short circuit, C3 short circuit, C4 short circuit, operational amplifier feedback resistor R16 short circuit, operational amplifier feedback resistor R16 open circuit, and solenoid valve short circuit, so as to ensure the safe operation of the electronically controlled injector and improve the reliability of the ship diesel engine.

[0089] Accordingly, the present application embodiment provides a fault detection system, see Figure 5 , Figure 5The structure diagram of a fault detection system provided by an embodiment of the present application is illustrated. The fault detection system is used to execute a fault detection method such as any one of the above embodiments to detect whether there is a fault in the solenoid valve drive circuit of the electronically controlled fuel injector. The fault detection system includes a first acquisition module, a second acquisition module, a third acquisition module, a fourth acquisition module and a result output module. Among them, the first acquisition module is configured to acquire a drive current waveform based on a current feedback signal of the solenoid valve drive circuit; the second acquisition module is configured to acquire a cylinder number mark corresponding to the electronically controlled fuel injector; the third acquisition module is configured to acquire a detection feature quantity based on the drive current waveform and the cylinder number mark; the detection feature quantity is configured to include a rise time, an excitation current maintenance time, a continuous flow time and a current maximum value; the fourth acquisition module is configured to input the detection feature quantity into a pre-built decision tree classification model to obtain multiple classification results; the result output module is configured to output a fault detection result based on the classification result.

[0090] It can be understood that, compared with the prior art, the fault detection system provided in the embodiment of the present application has all the technical features and beneficial effects of the above-mentioned fault detection method, which will not be repeated here.

[0091] The above is a detailed introduction to a fault detection method and system provided in an embodiment of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the technical solution and its core idea of ​​the present application. Ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some of the technical features therein with equivalents; and these modifications or replacements do not make the essence of the corresponding technical solution deviate from the scope of the technical solution of the embodiments of the present application.

Claims

1. A fault detection method, characterized in that: Used to detect whether there is a fault in the electronically controlled fuel injector solenoid valve drive circuit, the fault detection method includes: Acquiring a driving current waveform based on a current feedback signal of the solenoid valve driving circuit; Obtaining the cylinder number mark corresponding to the electronically controlled fuel injector; Based on the driving current waveform and the cylinder number mark, a detection feature quantity is obtained; the detection feature quantity is configured to include a rise time, an excitation current maintenance time, a free flow time and a current maximum value; Inputting the detection feature quantity into a pre-built decision tree classification model to obtain multiple classification results; Based on the classification result, a fault detection result is output.

2. The fault detection method according to claim 1, characterized in that: The decision tree classification model is pre-built through the following steps: Acquire full-operation test data of the solenoid valve drive circuit, a fault simulation data set, and a fault label corresponding to each fault data sample; wherein the fault simulation data set includes a plurality of the fault data samples; Based on the full-operating condition test data, the fault data samples and the fault labels, an initial data set is acquired; Determining a number of judgments; the number of judgments is configured to be an odd number; Based on the initial data set and the number of judgments, obtaining a decision tree classification model set and a fault identification accuracy set; Based on the decision tree classification model set and the fault identification accuracy set, the decision tree classification model is built.

3. The fault detection method according to claim 2, characterized in that: The step of obtaining a decision tree classification model set and a fault identification accuracy set based on the initial data set and the number of decisions includes: Based on the number of judgments, executing a preset number of loop steps to obtain multiple decision tree classification model subsets and multiple fault identification accuracy subsets; the preset number of times is determined based on the number of judgments; Determine a plurality of decision tree classification model subsets as the decision tree classification model set; Taking the average value of each component element of a plurality of the fault identification accuracy rate subsets as the fault identification accuracy rate set; Wherein, the circulation step comprises: Determine a training set and a test set based on the initial data set; Establishing the decision tree classification model subset based on the training set; Obtain a fault recognition accuracy subset of the test set.

4. The fault detection method according to claim 2, characterized in that: The full-operating condition test data is obtained through test simulation, and the fault data samples are obtained through simulation test.

5. The fault detection method according to claim 1, characterized in that: The step of obtaining a driving current waveform based on the current feedback signal of the solenoid valve driving circuit comprises: The current feedback signal is collected in real time through an analog input circuit; The current feedback signal is converted into a digital signal by a fast analog-to-digital converter to obtain the driving current waveform.

6. The fault detection method according to claim 1, characterized in that: The step of obtaining the cylinder number mark corresponding to the electronically controlled fuel injector comprises: Obtaining the gate cylinder number currently driven by the solenoid valve drive circuit; The cylinder number flag is determined based on the current feedback signal and the gated cylinder number.

7. The fault detection method according to claim 1, characterized in that: The step of acquiring the detection feature quantity based on the driving current waveform and the cylinder number mark comprises: Performing smoothing and denoising processing on the driving current waveform; Acquire key characteristic points in the driving current waveform; the key characteristic points include a current starting point, an excitation process starting point, a current rise end point, an excitation current end point, an excitation process end point, a current end point, and a maximum value point of a current feedback signal; The detection feature quantity is acquired based on the key feature point.

8. The fault detection method according to claim 7, characterized in that: The key feature points are obtained based on the following steps: Obtaining the maximum value point from the driving current waveform; Traversing each sampling point between the sampling starting point and the maximum value point of the current feedback signal, obtaining a first sampling point set whose absolute value of the sampling value is less than a first preset range, and determining the last point in the first sampling point set as the current starting point; The first preset range is determined according to the fluctuation of the current feedback signal in the no-current segment; Traversing each sampling point between the maximum value point and the sampling end point, obtaining a second sampling point set whose absolute value is less than the first preset range, and determining the first point in the second sampling point set as the current end point; Acquire a third sampling point set in which the absolute value of the deviation between the sampling value and the maximum value is less than a second preset range, and use the first point in the third sampling point set as the current rise end point, and use the last point in the third sampling point set as the excitation current end point; The second preset range is determined according to the driving current waveform; Between the current rise end point and the excitation current end point, a minimum value point is obtained, and a fourth sampling point set is obtained in which the absolute value of the deviation between the sampling value and the minimum value point is less than a third preset range, and the first point in the fourth sampling point set is used as the starting point of the excitation process, and the last point in the fourth sampling point set is used as the end point of the excitation process; The third preset range is determined according to the driving current waveform.

9. The fault detection method according to claim 1, characterized in that: The step of outputting a fault detection result based on the classification result comprises: In response to the number of non-fault results being greater than the number of faulty results in the plurality of classification results, not issuing a fault alarm signal; In response to the number of the faulty results being greater than the number of the non-faulty results, obtaining the occurrence frequency of the faulty results in the classification results; The fault mode corresponding to the fault result with the highest occurrence frequency is obtained, and a fault alarm signal is issued based on the fault mode.

10. The fault detection method according to claim 9, characterized in that: The step of outputting a fault detection result based on the classification result further includes: In response to the number of the faulty results with the highest occurrence frequency being multiple, obtaining multiple fault modes corresponding to the multiple faulty results; A fault warning signal is issued based on a plurality of the fault modes.

11. The fault detection method according to claim 2, characterized in that: Also includes: Based on the classification result and the detection feature quantity, the fault simulation data set is updated.

12. The fault detection method according to claim 11, characterized in that: The step of updating the fault simulation data set based on the classification result and the detection feature quantity comprises: Obtaining the latest fault simulation data set; Acquire a current decision tree classification model and a current fault identification accuracy set based on the fault simulation data set; In response to the current fault identification accuracy set being less than the fault identification accuracy set, the current decision tree classification model is updated to a new decision tree classification model, and the current fault identification accuracy set is updated to a new fault identification accuracy set.

13. A fault detection system, characterized in that: Used to execute the fault detection method according to any one of claims 1 to 12 to detect whether there is a fault in the drive circuit of the electronically controlled fuel injector solenoid valve, the fault detection system comprising: A first acquisition module is configured to acquire a driving current waveform based on a current feedback signal of the solenoid valve driving circuit; A second acquisition module is configured to acquire a cylinder number mark corresponding to the electronically controlled fuel injector; A third acquisition module is configured to acquire a detection feature quantity based on the drive current waveform and the cylinder number mark; the detection feature quantity is configured to include a rise time, an excitation current maintenance time, a freewheeling time, and a current maximum value; A fourth acquisition module is configured to input the detection feature quantity into a pre-built decision tree classification model to obtain multiple classification results; The result output module is configured to output a fault detection result based on the classification result.

Images