Knock signal noise correction method and device and medium

Abstract

The invention discloses a knock signal noise correction method and device and a medium, and the method comprises the steps: collecting a knock sensor signal of an engine at a preset moment, and carrying out the high-pass filtering of the knock sensor signal according to a target filtering coefficient, and obtaining a filtering signal; wherein the target filtering coefficient is determined according to the rotating speed and the load of the engine; correcting the filtering signal according to the target learning value to obtain a knock signal value; wherein the target learning value is calculated according to learning values accumulated by the engine under the dynamic working condition and the steady working condition. According to the knock signal noise correction method and device and the medium, high-pass filtering is carried out on the knock sensor signal according to the target filtering coefficient, more signal components irrelevant to knock can be filtered out, the filtering signal is corrected according to the target learning value, the knock signal recognition accuracy can be improved, and the knock signal noise correction efficiency is improved. The problem that knock signal noise is difficult to correct accurately according to the current working condition of the engine can be solved.

Description

Technical Field

[0001] The invention relates to the field of engine technology, and in particular to a knock signal noise correction method, device and medium. Background Art

[0002] The engine can efficiently drive the car to run smoothly with the help of the combustion process of fuels such as gasoline or diesel; due to unfavorable factors such as early ignition timing, excessive carbon deposits inside the engine, and excessive engine temperature, abnormal combustion of fuel may be caused, which in turn triggers engine knock signal noise as a warning of abnormal working conditions; since the engine knock signal noise may contain various noise components, these noises will interfere with the accurate identification of the knock signal, so the knock signal noise needs to be corrected. When processing the signal generated by engine knock, a high-pass filter algorithm is usually used, which can effectively filter out low-frequency noise and identify knock based on high-frequency amplitude; among them, the filter coefficient is obtained by collecting multi-condition knock signals with an oscilloscope and performing weighted averaging; and there is only one knock self-learning value in traditional software, and the compensation value needs to be calculated through each knock value to be stored in the learning value.

[0003] However, since the traditional knock signal noise correction method uses the same filter coefficient under all operating conditions, it has poor adaptability to the differences in knock signals under different operating conditions, which can easily lead to misjudgment or missed judgment, reduce the accuracy of knock prediction, and affect the accuracy of engine performance monitoring and fault diagnosis. Summary of the invention

[0004] The present invention provides a knock signal noise correction method, device and medium to solve the problem that it is difficult to accurately correct the knock signal noise according to the current working condition of the engine.

[0005] Collecting a knock sensor signal of the engine at a preset time, and performing high-pass filtering on the knock sensor signal according to a target filter coefficient to obtain a filtered signal; wherein the target filter coefficient is determined according to the speed and load of the engine;

[0006] The filtered signal is corrected according to a target learning value to obtain a knock signal value; wherein the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions.

[0007] The present invention determines the target filter coefficient according to the engine speed and load, and can optimize the frequency response characteristics of the filter so that it better matches the spectrum characteristics of the knock signal. This method can not only improve the filtering effect of the filter, but also reduce the signal distortion and delay in the filtering process, and further improve the real-time and accuracy of knock signal recognition. The target learning value is calculated based on the accumulated learning value of the engine under dynamic conditions and steady-state conditions, which means that the knock signal characteristics of the engine under different conditions have been considered, and correction experience has been gradually accumulated through learning; therefore, when correcting the filter signal, using the target learning value can make the correction more effective and closer to the knock signal characteristics of the engine under the current working condition, ensuring that an accurate knock signal value can be obtained in the end.

[0008] Compared with the prior art, the present invention performs high-pass filtering on the knock sensor signal according to the target filter coefficient, which helps to filter out more signal components that are not related to knock. Correcting the filtered signal according to the target learning value can improve the accuracy of knock signal recognition, thereby solving the problem of difficulty in accurately correcting the knock signal noise according to the current operating conditions of the engine.

[0009] As a preferred solution, the target filter coefficient is determined according to the speed and load of the engine, specifically:

[0010] Determining the target range of the speed and load to be calibrated according to the performance parameters of the engine under different working conditions;

[0011] Selecting a target speed and a target load point within the target range as coordinate points to establish a two-dimensional mapping diagram;

[0012] Starting the engine according to the coordinate point set in the two-dimensional map, and making the engine run at a speed and load combination corresponding to the coordinate point set, to obtain a knock signal set;

[0013] Performing filtering processing on the knock signal set according to a number of filtering coefficients to obtain a filtered signal set;

[0014] An optimal filter coefficient for a corresponding signal in the filter signal set is selected based on the rotation speed and load of the engine as the target filter coefficient.

[0015] This preferred solution determines the target range of speed and load to be calibrated based on the performance parameters of the engine under different working conditions, and can more accurately select representative speed and load points for calibration. The establishment of a two-dimensional mapping diagram makes the selection of coordinate points more intuitive and accurate, which helps to ensure that the calibration process covers the main working area of ​​the engine. By using several filter coefficients to filter the knock signal set, the optimal filter coefficient can be selected according to actual conditions, thereby more effectively removing noise and interference and retaining useful signals related to knock; and the selection of the optimal filter coefficient is based on the engine speed and load, which further improves the pertinence and accuracy of signal processing.

[0016] As a preferred solution, the optimal filter coefficient of the corresponding signal in the filter signal set is selected based on the speed and load of the engine as the target filter coefficient, specifically:

[0017] According to the knock recognition rate of the filter signal set, the filter coefficient of each coordinate point in the coordinate point set that performs best during filtering is calculated to obtain a filter coefficient set corresponding to the coordinate point set;

[0018] In the filter coefficient set, a first coordinate point is determined according to the rotation speed and load of the engine when the knock sensor signal is generated, and the filter coefficient corresponding to the first coordinate point is defined as the target filter coefficient.

[0019] This preferred solution can ensure that the knock signal can be optimally processed under each specific speed and load combination by calculating the best filter coefficient for each coordinate point during filtering. This method helps to improve the accuracy of knock identification and reduce false alarms and missed alarms. In addition, for different speed and load combinations, selecting the best filter coefficient can more effectively remove noise and interference while retaining the important features of the knock signal, which can improve the performance of the filter and make the filtered signal clearer and more accurate.

[0020] As a preferred solution, the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions, specifically:

[0021] A first learning value is calculated based on the engine state information corresponding to the filtered signal, and the first learning value is stored in a learning value storage interval; wherein the learning value storage interval is divided according to the speed and the load, and is established by calculating the knock signal learning value of the engine under the dynamic working condition and the steady-state working condition;

[0022] The target learning value is obtained by iteratively calculating a plurality of learning values ​​in the learning value storage interval according to a preset algorithm.

[0023] In this preferred solution, the learning value is calculated based on the knock signal and state information of the engine under real working conditions, so it can more accurately reflect the knock characteristics of the engine under different speeds and loads. Through iterative calculation, the error and uncertainty in the learning value can be further reduced, and its accuracy and reliability can be improved.

[0024] As a preferred solution, the learning value storage interval is divided according to the speed and load, and is established by calculating the knock signal learning value of the engine under dynamic working conditions and steady-state working conditions, specifically:

[0025] An initial learning value storage interval including a speed-related interval and a load-related interval is established based on the coordinate system;

[0026] Acquire the knock sensor signal of the engine under dynamic working condition and steady-state working condition to obtain a knock sensor signal set; wherein the dynamic working condition refers to the working condition of the engine when it is accelerating, decelerating and the load exceeds a preset number of changes within a preset time, and the steady-state working condition refers to the working condition of the engine when it is running at a constant speed and a constant load;

[0027] Calculating a number of learning values ​​according to the engine state information corresponding to the knock sensor signal set;

[0028] The plurality of learning values ​​are stored in the initial learning value storage interval according to the rotation speed and the load to obtain the learning value storage interval.

[0029] This preferred solution establishes an initial learning value storage interval containing speed-related intervals and load-related intervals based on the coordinate system, realizing structured storage of learning values. This structured storage method helps to quickly find and update the corresponding learning value when the engine working state changes. The clear division of dynamic working conditions and steady-state working conditions helps to more accurately obtain the knock sensor signal set of the engine under different working conditions; and by considering the dynamic working conditions of the engine when accelerating, decelerating, and the load exceeds the preset number of changes within a preset time, as well as the steady-state working conditions of uniform speed driving and constant load operation, it is ensured that the learning value can cover the main working area of ​​the engine.

[0030] The present application also provides a knock signal noise correction device, comprising a filtering module and a correction module;

[0031] The filtering module is used to collect the knock sensor signal of the engine at a preset time, and perform high-pass filtering on the knock sensor signal according to a target filtering coefficient to obtain a filtered signal; wherein the target filtering coefficient is determined according to the speed and load of the engine;

[0032] The correction module is used to correct the filtered signal according to a target learning value to obtain a knock signal value; wherein the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions.

[0033] As a preferred solution, the filtering module includes a parameter unit, a mapping unit, a signal unit, a filtering unit and a coefficient unit;

[0034] Wherein, the parameter unit is used to determine the target range of the speed and load to be calibrated according to the performance parameters of the engine under different working conditions;

[0035] The mapping unit is used to select a target speed and a target load point as a coordinate point within the target range to establish a two-dimensional mapping diagram;

[0036] The signal unit is used to start the engine according to the coordinate point set in the two-dimensional map, and make the engine run at the speed and load combination corresponding to the coordinate point set to obtain a knock signal set;

[0037] The filtering unit is used to filter the knock signal set according to a number of filtering coefficients to obtain a filtered signal set;

[0038] The coefficient unit is used to select the best filter coefficient of the corresponding signal in the filter signal set based on the speed and load of the engine as the target filter coefficient.

[0039] As a preferred solution, the coefficient unit includes a coefficient sub-unit and a coordinate sub-unit;

[0040] The coefficient subunit is used to calculate the best filter coefficient of each coordinate point in the coordinate point set when filtering according to the knock recognition rate of the filter signal set, so as to obtain the filter coefficient set corresponding to the coordinate point set;

[0041] The coordinate subunit is used to determine a first coordinate point in the filter coefficient set according to the rotation speed and load of the engine when generating the knock sensor signal, and define the filter coefficient corresponding to the first coordinate point as the target filter coefficient.

[0042] As a preferred solution, the correction module includes a storage unit and a calculation unit;

[0043] The storage unit is used to calculate a first learning value according to the engine state information corresponding to the filtered signal, and store the first learning value in a learning value storage interval; wherein the learning value storage interval is divided according to the speed and the load, and is established by calculating the knock signal learning value of the engine under the dynamic working condition and the steady-state working condition;

[0044] The calculation unit is used to iteratively calculate a number of learning values ​​in the learning value storage interval according to a preset algorithm to obtain the target learning value.

[0045] As a preferred solution, the storage unit includes an interval subunit, a data subunit, a calculation subunit and a storage subunit;

[0046] Wherein, the interval subunit is used to establish an initial learning value storage interval including a speed-related interval and a load-related interval based on a coordinate system;

[0047] The data subunit is used to obtain the knock sensor signal of the engine under dynamic working conditions and steady-state working conditions to obtain a knock sensor signal set; wherein the dynamic working condition refers to the working state of the engine when it is accelerating, decelerating, and the load exceeds a preset number of changes within a preset time period, and the steady-state working condition refers to the working state of the engine when it is running at a constant speed and a constant load;

[0048] The calculation subunit is used to calculate a number of learning values ​​according to the engine state information corresponding to the knock sensor signal set;

[0049] The storage subunit is used to store the plurality of learning values ​​in the initial learning value storage interval according to the rotation speed and the load to obtain the learning value storage interval.

[0050] The present application also provides a storage medium, on which a computer program is stored. The computer program is called and executed by a computer to implement the knock signal noise correction method as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flow chart of a knock signal noise correction method provided in an embodiment of the present application;

[0052] Figure 2 It is a structural design diagram provided by the embodiment of the present application;

[0053] Figure 3 It is a structural schematic diagram of a knock signal noise correction device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0055] In the description of this application, it should be understood that the term "first" is used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "several" means two or more.

[0056] A knock signal noise correction method provided in an embodiment of the present application is mainly used in situations where it is necessary to flexibly adapt to the processing requirements of knock information under different working conditions, effectively eliminate interference factors, and thus significantly improve the accuracy of knock detection.

[0057] Embodiment 1:

[0058] See also Figure 1 The embodiment of the present application provides a knock signal noise correction method, including S1 to S2, and the specific implementation steps are as follows:

[0059] S1. Collecting a knock sensor signal of the engine at a preset time, and performing high-pass filtering on the knock sensor signal according to a target filter coefficient to obtain a filtered signal; wherein the target filter coefficient is determined according to the engine speed and load.

[0060] Step S1 of the embodiment of the present application includes S1.1 to S1.2, specifically:

[0061] S1.1. Collecting the knock sensor signal of the engine at a preset time;

[0062] Determine the target range of speed and load to be calibrated based on the performance parameters of the engine under different working conditions;

[0063] A series of representative target speed and target load points are selected as coordinate points within the target range to establish a two-dimensional map; and in the two-dimensional map, the horizontal axis is the speed and the vertical axis is the load; wherein the "representative" target speed and target load points refer to points that can fully and accurately reflect the performance characteristics of the engine within the speed and load range;

[0064] The engine is started according to the coordinate point set (i.e., the combination set of speed and load) in the two-dimensional mapping diagram, and the engine is operated at the speed and load combination corresponding to the coordinate point set to obtain a knock signal set;

[0065] The knock signal set is filtered according to a number of filter coefficients to obtain a filtered signal set.

[0066] According to the filtering effect and knock recognition rate of the filtering signal set, the best filtering coefficient of each coordinate point in the coordinate point set is calculated as the calibration value of the point, and the best filtering coefficients of all coordinate points are stored in a two-dimensional mapping diagram to form a complete filtering parameter table, so as to obtain the filtering coefficient set corresponding to the coordinate point set;

[0067] In the filter coefficient set, a first coordinate point is determined according to the speed and load of the engine when the knock sensor signal is generated at a preset time, and the filter coefficient corresponding to the first coordinate point is defined as a target filter coefficient.

[0068] In this embodiment S1.1, the target range of the speed and load to be calibrated is determined according to the performance parameters of the engine under different working conditions, so that representative speed and load points can be selected more accurately for calibration. The establishment of a two-dimensional mapping diagram makes the selection of coordinate points more intuitive and accurate, which helps to ensure that the calibration process covers the main working area of ​​the engine. The knock signal set is filtered using several filter coefficients, and the optimal filter coefficient can be selected according to actual conditions, so as to more effectively remove noise and interference and retain useful signals related to knock; and the selection of the optimal filter coefficient is based on the engine speed and load, which further improves the pertinence and accuracy of signal processing;

[0069] In addition, by calculating the best filter coefficient for each coordinate point during filtering, it can be ensured that the knock signal can be optimally processed under each specific speed and load combination. This method helps to improve the accuracy of knock identification and reduce false alarms and missed alarms. Moreover, for different speed and load combinations, selecting the best filter coefficient can more effectively remove noise and interference while retaining the important characteristics of the knock signal, which can improve the performance of the filter and make the filtered signal clearer and more accurate.

[0070] S1.2. Perform high-pass filtering on the knock sensor signal according to the target filter coefficient to obtain a filtered signal.

[0071] S2. Correct the filtered signal according to the target learning value to obtain the knock signal value; wherein the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions.

[0072] Step S2 of the embodiment of the present application includes S2.1 to S2.2, specifically:

[0073] S2.1. Establishing an initial learning value storage interval including a speed-related interval and a load-related interval based on the coordinate system; wherein, the "initial learning value storage interval" refers to a series of storage areas or ranges divided for storing and managing learning values;

[0074] Acquire the knock sensor signal of the engine under dynamic working condition and steady-state working condition to obtain the knock sensor signal set; wherein the dynamic working condition refers to the working condition when the engine is in acceleration, deceleration and the load exceeds the preset number of changes within a preset time, and the steady-state working condition refers to the working condition when the engine is in uniform speed and constant load operation;

[0075] According to the engine state information corresponding to the knock sensor signal set, a number of learning values ​​are calculated by a preset algorithm (such as machine learning, statistical method, etc.); wherein the number of learning values ​​reflects the correlation or mapping relationship between the knock sensor signal and the actual knock state of the engine under different working conditions;

[0076] A number of learning values ​​are stored in an initial learning value storage interval according to the rotation speed and the load to obtain a learning value storage interval.

[0077] A specific example is as follows: the initial learning value storage interval is set to 20 intervals, of which the load-correlation Y-axis is 4 intervals and the speed-correlation X-axis is 5 intervals. The acquired learning values ​​are stored in 20 different intervals to obtain the learning value storage interval; among them, the 4 intervals of the load-correlation Y-axis reflect the working state of the engine under different loads, and each interval corresponds to a certain load range; the 5 intervals of the speed-correlation X-axis reflect the working state of the engine at different speeds, and each interval corresponds to a certain speed range.

[0078] In this embodiment S2.1, based on the coordinate system, an initial learning value storage interval including a speed-related interval and a load-related interval is established, thereby realizing structured storage of learning values, which is larger than the capacity of a traditional learning value storage interval. This structured storage method helps to quickly find and update the corresponding learning value when the engine working state changes. The clear division of dynamic working conditions and steady-state working conditions helps to more accurately obtain the knock sensor signal set of the engine under different working conditions; and by considering the dynamic working conditions of the engine when accelerating, decelerating, and the load exceeds the preset number of changes within a preset time, as well as the steady-state working conditions of uniform speed driving and constant load operation, it is ensured that the learning value can cover the main working area of ​​the engine.

[0079] S2.2, calculating a first learning value according to the engine state information corresponding to the filtered signal, and storing the first learning value in a learning value storage interval;

[0080] The target learning value is obtained by iteratively calculating a number of learning values ​​in the learning value storage interval according to the superposition average algorithm.

[0081] The filtered signal is corrected according to the target learning value to obtain a more accurate knock signal value.

[0082] To apply the embodiments of this application, please refer to Figure 2 , Figure 2 It is a structural design diagram provided in an embodiment of the present application, which shows the general process of processing the knock sensor signal of the engine in an embodiment of the present application to obtain a more accurate knock signal (knock signal value).

[0083] In this embodiment S2.2, the learning value is calculated based on the knock signal and state information of the engine under real working conditions, so it can more accurately reflect the knock characteristics of the engine under different speeds and loads. Through iterative calculation, the error and uncertainty in the learning value can be further reduced, and its accuracy and reliability can be improved.

[0084] Overall, this embodiment has the following beneficial effects:

[0085] This application determines the target filter coefficient according to the engine speed and load, which can optimize the frequency response characteristics of the filter to better match the spectral characteristics of the knock signal. This method can not only improve the filtering effect of the filter, but also reduce signal distortion and delay during the filtering process, further improving the real-time and accuracy of knock signal identification. The target learning value is calculated based on the accumulated learning value of the engine under dynamic conditions and steady-state conditions, which means that the knock signal characteristics of the engine under different conditions have been taken into account, and correction experience has been gradually accumulated through learning; therefore, when correcting the filtered signal, using the target learning value can make the correction more effective and closer to the knock signal characteristics of the engine under the current working conditions, ensuring that an accurate knock signal value can be obtained in the end;

[0086] In summary, the learning value is subdivided into learning values ​​under dynamic conditions and steady-state conditions. This measure accurately distinguishes different driving conditions, thereby significantly improving the pertinence of detection; through such subdivision, the learning value can more accurately fit the current actual conditions, avoiding the problem of learning value averaging caused by taking into account other conditions, thereby ensuring higher accuracy of learning values ​​and detection results. In addition, calibrable filter coefficients and two-dimensional mapping diagrams for noise compensation are introduced. These technical means can more flexibly adapt to the processing requirements of knock information under different conditions, effectively eliminate interference factors, and thus significantly improve the accuracy of knock detection.

[0087] Embodiment 2:

[0088] See also Figure 3 , an embodiment of the present application provides a knock signal noise correction device, comprising a filtering module 10 and a correction module 20;

[0089] The filter module 10 is used to collect the knock sensor signal of the engine at a preset time, and perform high-pass filtering on the knock sensor signal according to a target filter coefficient to obtain a filter signal; wherein the target filter coefficient is determined according to the speed and load of the engine;

[0090] The correction module 20 is used to correct the filtered signal according to the target learning value to obtain the knock signal value; wherein the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions.

[0091] In one embodiment, the filtering module 10 includes a parameter unit, a mapping unit, a signal unit, a filtering unit, a coefficient subunit, a coordinate subunit, and a high-pass filtering unit;

[0092] Wherein, the parameter unit is used to collect the knock sensor signal of the engine at a preset time;

[0093] The parameter unit is also used to determine the target range of the speed and load to be calibrated according to the performance parameters of the engine under different working conditions;

[0094] A mapping unit is used to select a series of representative target speed and target load points within the target range as coordinate points to establish a two-dimensional mapping diagram (two-dimensional map); and in the two-dimensional mapping diagram, the horizontal axis is the speed and the vertical axis is the load; wherein the "representative" target speed and target load points refer to points that can fully and accurately reflect the performance characteristics of the engine within the speed and load range;

[0095] A signal unit, used to start the engine according to a set of coordinate points (i.e., a combination set of speed and load) in the two-dimensional mapping diagram, and to make the engine run at a speed and load combination corresponding to the set of coordinate points, thereby obtaining a knock signal set;

[0096] The filtering unit is used to filter the knock signal set according to a number of filter coefficients to obtain a filtered signal set.

[0097] The coefficient subunit is used to calculate the best filter coefficient of each coordinate point in the coordinate point set when filtering according to the filtering effect and knock recognition rate of the filter signal set as the calibration value of the point, and store the best filter coefficients of all coordinate points in a two-dimensional mapping diagram to form a complete filter parameter table, so as to obtain the filter coefficient set corresponding to the coordinate point set;

[0098] The coordinate subunit is used to determine a first coordinate point in the filter coefficient set according to the speed and load of the engine when the knock sensor signal is generated at a preset time, and define the filter coefficient corresponding to the first coordinate point as a target filter coefficient.

[0099] The parameter unit, mapping unit, signal unit, filtering unit, coefficient sub-unit and coordinate sub-unit of this embodiment determine the target range of speed and load to be calibrated according to the performance parameters of the engine under different working conditions, and can more accurately select representative speed and load points for calibration. The establishment of a two-dimensional mapping diagram makes the selection of coordinate points more intuitive and accurate, which helps to ensure that the calibration process covers the main working area of ​​the engine. The knock signal set is filtered using several filter coefficients, and the optimal filter coefficient can be selected according to actual conditions, so as to more effectively remove noise and interference and retain useful signals related to knock; and the selection of the optimal filter coefficient is based on the engine speed and load, which further improves the pertinence and accuracy of signal processing;

[0100] In addition, by calculating the best filter coefficient for each coordinate point during filtering, it can be ensured that the knock signal can be optimally processed under each specific speed and load combination. This method helps to improve the accuracy of knock identification and reduce false alarms and missed alarms. Moreover, for different speed and load combinations, selecting the best filter coefficient can more effectively remove noise and interference while retaining the important characteristics of the knock signal, which can improve the performance of the filter and make the filtered signal clearer and more accurate.

[0101] The high-pass filter unit is used to perform high-pass filtering on the knock sensor signal according to a target filter coefficient to obtain a filtered signal.

[0102] In one embodiment, the correction module 20 includes an interval subunit, a data subunit, a calculation subunit, a storage subunit, a storage unit, and a calculation unit;

[0103] The interval subunit is used to establish an initial learning value storage interval including a speed-related interval and a load-related interval based on the coordinate system; wherein the "initial learning value storage interval" refers to a series of storage areas or ranges divided for storing and managing learning values;

[0104] A data subunit is used to obtain the knock sensor signal of the engine under dynamic working conditions and steady-state working conditions to obtain a knock sensor signal set; wherein the dynamic working condition refers to the working state when the engine is accelerating, decelerating, and the load exceeds a preset number of changes within a preset time, and the steady-state working condition refers to the working state when the engine is running at a constant speed and a constant load;

[0105] A calculation subunit, for calculating a number of learning values ​​by a preset algorithm (such as machine learning, statistical method, etc.) according to the engine state information corresponding to the knock sensor signal set; wherein the number of learning values ​​reflects the association or mapping relationship between the knock sensor signal and the actual knock state of the engine under different working conditions;

[0106] The storage subunit is used to store a number of learning values ​​in an initial learning value storage interval according to the rotation speed and the load to obtain a learning value storage interval.

[0107] A specific example is as follows: the initial learning value storage interval is set to 20 intervals, of which the load-correlation Y-axis is 4 intervals and the speed-correlation X-axis is 5 intervals. The acquired learning values ​​are stored in 20 different intervals to obtain the learning value storage interval; among them, the 4 intervals of the load-correlation Y-axis reflect the working state of the engine under different loads, and each interval corresponds to a certain load range; the 5 intervals of the speed-correlation X-axis reflect the working state of the engine at different speeds, and each interval corresponds to a certain speed range.

[0108] In this embodiment, the interval subunit, data subunit, calculation subunit and storage subunit establish an initial learning value storage interval including a speed-related interval and a load-related interval based on the coordinate system, realizing structured storage of learning values, which is larger than the capacity of a traditional learning value storage interval. This structured storage method helps to quickly find the corresponding learning value and update it when the engine working state changes. The clear division of dynamic working conditions and steady-state working conditions helps to more accurately obtain the knock sensor signal set of the engine under different working conditions; and by considering the dynamic working conditions of the engine when accelerating, decelerating, and the load exceeds the preset number of changes within a preset time, as well as the steady-state working conditions of uniform speed driving and constant load operation, it is ensured that the learning value can cover the main working area of ​​the engine.

[0109] A storage unit, used for calculating a first learning value according to the engine state information corresponding to the filtered signal, and storing the first learning value in a learning value storage interval;

[0110] The calculation unit is used to iteratively calculate a number of learning values ​​in the learning value storage interval according to a superposition average algorithm to obtain a target learning value.

[0111] The calculation unit is also used to correct the filtered signal according to the target learning value to obtain a more accurate knock signal value.

[0112] To apply the embodiments of this application, please refer to Figure 2 , Figure 2 It is a structural design diagram provided in an embodiment of the present application, which shows the general process of processing the knock sensor signal of the engine in the second embodiment of the present application to obtain a more accurate knock signal (knock signal value).

[0113] In the storage unit and the calculation unit of this embodiment, the learning value is calculated based on the knock signal and state information of the engine under real working conditions, so it can more accurately reflect the knock characteristics of the engine under different speeds and loads. Through iterative calculation, the error and uncertainty in the learning value can be further reduced, and its accuracy and reliability can be improved.

[0114] Overall, this embodiment has the following beneficial effects:

[0115] This application determines the target filter coefficient according to the engine speed and load, which can optimize the frequency response characteristics of the filter to better match the spectral characteristics of the knock signal. This method can not only improve the filtering effect of the filter, but also reduce signal distortion and delay during the filtering process, further improving the real-time and accuracy of knock signal identification. The target learning value is calculated based on the accumulated learning value of the engine under dynamic conditions and steady-state conditions, which means that the knock signal characteristics of the engine under different conditions have been taken into account, and correction experience has been gradually accumulated through learning; therefore, when correcting the filtered signal, using the target learning value can make the correction more effective and closer to the knock signal characteristics of the engine under the current working conditions, ensuring that an accurate knock signal value can be obtained in the end;

[0116] In summary, the learning value is subdivided into learning values ​​under dynamic conditions and steady-state conditions. This measure accurately distinguishes different driving conditions, thereby significantly improving the pertinence of detection; through such subdivision, the learning value can more accurately fit the current actual conditions, avoiding the problem of learning value averaging caused by taking into account other conditions, thereby ensuring higher accuracy of learning values ​​and detection results. In addition, calibrable filter coefficients and two-dimensional mapping diagrams for noise compensation are introduced. These technical means can more flexibly adapt to the processing requirements of knock information under different conditions, effectively eliminate interference factors, and thus significantly improve the accuracy of knock detection.

[0117] Embodiment three:

[0118] An embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the knock signal noise correction method;

[0119] Wherein, the knock signal noise correction method, if implemented in the form of a software functional unit and used as an independent product, can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0120] The above are preferred embodiments of the present invention. It should be noted that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A knock signal noise correction method, characterized in that: include: Collecting a knock sensor signal of the engine at a preset time, and performing high-pass filtering on the knock sensor signal according to a target filter coefficient to obtain a filtered signal; wherein the target filter coefficient is determined according to the speed and load of the engine; The filtered signal is corrected according to a target learning value to obtain a knock signal value; wherein the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions.

2. A knock signal noise correction method according to claim 1, characterized in that: The target filter coefficient is determined according to the speed and load of the engine, specifically: Determining the target range of the speed and load to be calibrated according to the performance parameters of the engine under different working conditions; Selecting a target speed and a target load point within the target range as coordinate points to establish a two-dimensional mapping diagram; Starting the engine according to the coordinate point set in the two-dimensional map, and making the engine run at a speed and load combination corresponding to the coordinate point set, to obtain a knock signal set; Performing filtering processing on the knock signal set according to a number of filtering coefficients to obtain a filtered signal set; An optimal filter coefficient for a corresponding signal in the filter signal set is selected based on the rotation speed and load of the engine as the target filter coefficient.

3. A knock signal noise correction method as claimed in claim 2, characterized in that: Based on the speed and load of the engine, the optimal filter coefficient of the corresponding signal in the filter signal set is selected as the target filter coefficient, which is specifically: According to the knock recognition rate of the filter signal set, the filter coefficient of each coordinate point in the coordinate point set that performs best during filtering is calculated to obtain a filter coefficient set corresponding to the coordinate point set; In the filter coefficient set, a first coordinate point is determined according to the rotation speed and load of the engine when the knock sensor signal is generated, and the filter coefficient corresponding to the first coordinate point is defined as the target filter coefficient.

4. A knock signal noise correction method as claimed in claim 1, characterized in that: The target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions, specifically: A first learning value is calculated based on the engine state information corresponding to the filtered signal, and the first learning value is stored in a learning value storage interval; wherein the learning value storage interval is divided according to the speed and the load, and is established by calculating the knock signal learning value of the engine under the dynamic working condition and the steady-state working condition; The target learning value is obtained by iteratively calculating a plurality of learning values ​​in the learning value storage interval according to a preset algorithm.

5. A knock signal noise correction method as claimed in claim 4, characterized in that: The learning value storage interval is divided according to the speed and load, and is established by calculating the knock signal learning value of the engine under dynamic working conditions and steady-state working conditions, specifically: An initial learning value storage interval including a speed-related interval and a load-related interval is established based on the coordinate system; Acquire the knock sensor signal of the engine under dynamic working condition and steady-state working condition to obtain a knock sensor signal set; wherein the dynamic working condition refers to the working condition of the engine when it is accelerating, decelerating and the load exceeds a preset number of changes within a preset time, and the steady-state working condition refers to the working condition of the engine when it is running at a constant speed and a constant load; Calculating a number of learning values ​​according to the engine state information corresponding to the knock sensor signal set; The plurality of learning values ​​are stored in the initial learning value storage interval according to the rotation speed and the load to obtain the learning value storage interval.

6. A knock signal noise correction device, characterized in that: Including filtering module and correction module; The filtering module is used to collect the knock sensor signal of the engine at a preset time, and perform high-pass filtering on the knock sensor signal according to a target filtering coefficient to obtain a filtered signal; wherein the target filtering coefficient is determined according to the speed and load of the engine; The correction module is used to correct the filtered signal according to a target learning value to obtain a knock signal value; wherein the target learning value is calculated based on the accumulated learning values ​​of the engine under dynamic conditions and steady-state conditions.

7. A knock signal noise correction device as claimed in claim 6, characterized in that: The filtering module includes a parameter unit, a mapping unit, a signal unit, a filtering unit and a coefficient unit; Wherein, the parameter unit is used to determine the target range of the speed and load to be calibrated according to the performance parameters of the engine under different working conditions; The mapping unit is used to select a target speed and a target load point as a coordinate point within the target range to establish a two-dimensional mapping diagram; The signal unit is used to start the engine according to the coordinate point set in the two-dimensional map, and make the engine run at the speed and load combination corresponding to the coordinate point set to obtain a knock signal set; The filtering unit is used to filter the knock signal set according to a number of filtering coefficients to obtain a filtered signal set; The coefficient unit is used to select the best filter coefficient of the corresponding signal in the filter signal set based on the speed and load of the engine as the target filter coefficient.

8. A knock signal noise correction device as claimed in claim 7, characterized in that: The coefficient unit includes a coefficient sub-unit and a coordinate sub-unit; The coefficient subunit is used to calculate the best filter coefficient of each coordinate point in the coordinate point set when filtering according to the knock recognition rate of the filter signal set, so as to obtain the filter coefficient set corresponding to the coordinate point set; The coordinate subunit is used to determine a first coordinate point in the filter coefficient set according to the rotation speed and load of the engine when generating the knock sensor signal, and define the filter coefficient corresponding to the first coordinate point as the target filter coefficient.

9. A knock signal noise correction device as claimed in claim 6, characterized in that: The correction module includes a storage unit and a calculation unit; The storage unit is used to calculate a first learning value according to the engine state information corresponding to the filtered signal, and store the first learning value in a learning value storage interval; wherein the learning value storage interval is divided according to the speed and the load, and is established by calculating the knock signal learning value of the engine under the dynamic working condition and the steady-state working condition; The calculation unit is used to iteratively calculate a number of learning values ​​in the learning value storage interval according to a preset algorithm to obtain the target learning value.

10. A knock signal noise correction device as claimed in claim 9, characterized in that: The storage unit includes an interval subunit, a data subunit, a calculation subunit and a storage subunit; Wherein, the interval subunit is used to establish an initial learning value storage interval including a speed-related interval and a load-related interval based on a coordinate system; The data subunit is used to obtain the knock sensor signal of the engine under dynamic working conditions and steady-state working conditions to obtain a knock sensor signal set; wherein the dynamic working condition refers to the working state of the engine when it is accelerating, decelerating, and the load exceeds a preset number of changes within a preset time period, and the steady-state working condition refers to the working state of the engine when it is running at a constant speed and a constant load; The calculation subunit is used to calculate a number of learning values ​​according to the engine state information corresponding to the knock sensor signal set; The storage subunit is used to store the plurality of learning values ​​in the initial learning value storage interval according to the rotation speed and the load to obtain the learning value storage interval.

11. A storage medium, characterized in that: The storage medium stores a computer program, which is called and executed by a computer to implement a knock signal noise correction method as described in any one of claims 1 to 5 above.

Images