Signal analysis method for electronic control single-cylinder diesel engine

Abstract

The invention relates to the technical field of data processing, in particular to a signal analysis method for an electronic control single-cylinder diesel engine. The method comprises the following steps: acquiring vibration amplitude regularity and vibration variation amplitude at a moment; the abnormal degree of the vibration amplitude at each moment is obtained according to the difference value of every two adjacent elements in the vibration change amplitude sequence at each moment and the vibration amplitude regularity at each moment; performing optimization according to the vibration amplitude and the vibration amplitude regularity corresponding to each moment in the judgment sequence of the vibration amplitude of each moment to obtain the optimization abnormal degree of the vibration amplitude of each moment; obtaining a real vibration amplitude sequence of each vibration sensor; further obtaining a real sound pressure sequence of each sound level meter; and analyzing according to the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter to obtain the fault type of the single-cylinder diesel engine. According to the method, noise can be accurately eliminated, and the fault type of the single-cylinder diesel engine can be accurately obtained.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a signal analysis method for an electronically controlled single-cylinder diesel engine. Background Art

[0002] As a power machine widely used in oil fields, railway traction, fixed power generation, engineering machinery and ships, diesel engines are increasingly developing towards large-scale, complex and precise. As the most commonly used power equipment, once a part or a link of the diesel engine fails, it will cause damage to key equipment and even cause huge economic losses. Therefore, it is necessary to analyze various signals collected by various vibration sensors when the diesel engine is working, find and diagnose faults in time and take effective measures to increase the safety and reliability of the diesel engine, reduce the maintenance cost of the diesel engine, reduce the losses caused by it, and prevent sudden accidents.

[0003] However, when analyzing various signals collected by various vibration sensors, since the diesel engine is disturbed by the complex operating environment, noise signals will exist in the various collected signals. The noise signals will interfere with the subsequent analysis of various signals, making the analysis results inaccurate. The existing method for removing noise signals is to filter the signals based on various filtering algorithms. However, since the filtering is performed on the entire signal, it is easy to filter out the normal signals, resulting in signal distortion, which will lead to inaccurate subsequent analysis results of the signal. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a signal analysis method for an electronically controlled single-cylinder diesel engine. The technical solution adopted is as follows: An embodiment of the present invention provides a signal analysis method for an electronically controlled single-cylinder diesel engine, the method comprising: The vibration amplitudes collected by vibration sensors at different positions at the same time and the distances between the corresponding vibration sensors and the vibration source constitute a vibration attenuation ratio sequence at that moment; the regularity of the vibration amplitude at that moment is obtained according to every three adjacent elements in the vibration attenuation ratio sequence; the vibration change amplitude at that moment is obtained according to the vibration amplitude at a moment collected by a vibration sensor, the vibration amplitudes at the moment before and the moment after that moment; The vibration change amplitudes at at least two moments before a moment and at least two moments after the moment constitute a vibration change amplitude sequence at the moment; the abnormality of the vibration amplitude at each moment is obtained according to the difference between every two adjacent elements in the vibration change amplitude sequence at each moment and the regularity of the vibration amplitude at each moment; Based on the position of the vibration amplitude at a moment in its cycle, the moments at the same position in a preset number of cycles adjacent to and before the cycle are obtained to form a judgment sequence of the vibration amplitude at the moment; according to the vibration amplitude corresponding to each moment in the judgment sequence of the vibration amplitude at each moment and the regularity of the vibration amplitude, the abnormality degree of the vibration amplitude at each moment is obtained to obtain the optimized abnormality degree of the vibration amplitude at each moment; The vibration amplitudes at different moments collected by a vibration sensor are screened based on the optimized abnormality degree of the vibration amplitudes at different moments collected by a vibration sensor to obtain the real vibration amplitude sequence of the vibration sensor; then the sound pressures at different moments collected by sound level meters at different positions are obtained, and the real sound pressure sequence of each sound level meter is screened to obtain; the fault type of a single-cylinder diesel engine is obtained by analyzing the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter.

[0005] Preferably, obtaining the regularity of the vibration amplitude at that moment according to every three adjacent elements in the vibration attenuation ratio sequence includes: The absolute value of the difference between the first element and the second element and the absolute value of the difference between the second element and the third element in every three adjacent elements in the vibration attenuation ratio sequence are obtained; the absolute value of the difference between the first element and the second element is obtained, and the absolute value of the result of subtracting the absolute value of the difference between the second element and the third element is subtracted, and the opposite number of the absolute value of the result is used as the power exponent of the exponential function with the natural constant as the base to obtain the function value of the exponential function with the natural constant as the base corresponding to every three adjacent elements; the function values ​​of the exponential function with the natural constant as the base corresponding to every three adjacent elements in the vibration attenuation ratio sequence are averaged to obtain the regularity of the vibration amplitude.

[0006] Preferably, the vibration change amplitude at a moment is obtained based on the vibration amplitude at a moment collected by a vibration sensor, and the vibration amplitudes at the moment before and after the moment, including: the absolute value of the difference between the vibration amplitude at a moment and the vibration amplitude at the moment before the moment, plus the absolute value of the difference between the vibration amplitude at a moment and the vibration amplitude at the moment after the moment, to obtain an addition result; the ratio of the addition result to twice the absolute value of the vibration amplitude at the moment is the vibration change amplitude at the moment.

[0007] Preferably, the abnormality of the vibration amplitude at each moment is obtained according to the difference between every two adjacent elements in the vibration change amplitude sequence at each moment and the regularity of the vibration amplitude at each moment, including: taking the opposite of the absolute value of the difference between every two adjacent elements in the vibration change amplitude sequence as the power exponent of an exponential function with a natural constant as the base to obtain the function value of the exponential function; averaging the difference between the function values ​​corresponding to every two adjacent elements in the vibration change amplitude sequence at a moment, and multiplying it by the regularity of the vibration amplitude at that moment to obtain the abnormality of the vibration amplitude at that moment.

[0008] Preferably, the optimized abnormal degree of the vibration amplitude at each moment is: in, and They respectively represent the optimized abnormality degree and abnormality degree of the vibration amplitude at time T collected by a sensor; The number of moments in the judgment sequence representing the vibration amplitude at time T; represents the vibration amplitude at time T; The vibration amplitude corresponding to the i-th element in the judgment sequence representing the vibration amplitude at time T; Indicates the regularity of the vibration amplitude at time T; The regularity of the vibration amplitude corresponding to the i-th element in the judgment sequence representing the vibration amplitude at time T; and the elements in the judgment sequence are time; and e represents a natural constant.

[0009] Preferably, the vibration amplitudes collected at different moments by a vibration sensor are screened based on the optimized abnormality degree of the vibration amplitudes collected at different moments by the vibration sensor to obtain a real vibration amplitude sequence of the vibration sensor, including: setting a judgment threshold, removing the vibration amplitudes with optimized abnormality degrees greater than the judgment threshold from the vibration amplitudes collected at different moments by the vibration sensor, and forming the real vibration amplitude sequence of the vibration sensor.

[0010] Preferably, the fault type of the single-cylinder diesel engine is obtained by analyzing the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and the sound level meter, including: obtaining historical vibration amplitude data and sound pressure data; obtaining the probability of the vibration amplitude at a moment in the real vibration amplitude sequence of any vibration sensor appearing in the historical vibration amplitude data and the probability of the sound pressure at the same moment in the real sound pressure sequence of any sound level meter appearing in the historical sound pressure data, which are vibration history probability and sound pressure history probability respectively; The frequency of the vibration amplitude at the moment within a preset number of cycles adjacent to the cycle before and after the vibration amplitude at the moment is obtained, and the ratio of the frequency to the number of elements in the judgment sequence at the moment is recorded as the first probability; the frequency of the sound pressure at the moment within a preset number of cycles adjacent to the cycle before and after the sound pressure at the moment is obtained, and the ratio of the frequency to the number of elements in the judgment sequence at the moment is recorded as the second probability; The first probability sum at the moment is multiplied by the difference between the vibration history probability at the moment to obtain the vibration abnormality probability at the moment; the second probability sum at the moment is multiplied by the difference between the sound pressure history probability at the moment to obtain the sound pressure abnormality probability at the moment; the product of the vibration abnormality probability and the sound pressure abnormality probability at the moment is the fault probability at the moment; a first threshold is set, and if the fault probability at the moment is greater than the first threshold, the moment is the fault moment; the vibration amplitude and sound pressure at all fault moments in the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter are obtained to form a fault feature vector; the fault feature vector is input into the diesel engine fault classification neural network, and the fault type of the diesel engine is output.

[0011] The embodiments of the present invention have at least the following beneficial effects: the present invention obtains the regularity of the vibration amplitude at the same moment by analyzing the vibration amplitudes collected by vibration sensors located at different positions and the ratio of the distances between the vibration sensors at different positions and the vibration source, and on this basis, it can determine whether the vibration amplitude collected at each moment has regularity, which can help to more accurately screen out the vibration amplitudes belonging to noise points in the subsequent process; further, the difference between adjacent elements in the vibration change amplitude sequence corresponding to each moment is analyzed and combined with the regularity of the vibration amplitude at each moment to obtain the abnormality of the vibration amplitude at each moment, and combined with the characteristics of the vibration amplitude belonging to the noise point in one cycle, it can More accurately filter out the noise points in all vibration amplitudes collected by a vibration sensor; at the same time, optimize the obtained abnormality degree by combining the vibration amplitude and vibration amplitude regularity at the same position in different periods, and obtain a more accurate optimized abnormality degree. The vibration amplitude collected by each vibration sensor at different times is screened and the noise points are removed, so that the subsequent analysis of the vibration amplitude can obtain more accurate analysis results. At the same time, the sound pressure collected by the sound level meter at different positions at different times is also screened to remove the noise points, and the real sound pressure sequence of each sound pressure meter is obtained. Combined with the real vibration amplitude sequence for analysis, the fault type of the single-cylinder diesel engine can be more accurately determined. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0013] Figure 1 A method flow chart of a signal analysis method for an electronically controlled single-cylinder diesel engine provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0014] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a signal analysis method for an electronically controlled single-cylinder diesel engine proposed by the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.

[0015] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0016] A specific scheme of a signal analysis method for an electronically controlled single-cylinder diesel engine provided by the present invention is described in detail below with reference to the accompanying drawings.

[0017] Example: The main application scenario of the present invention is: in order to determine the fault type of the diesel engine, it is necessary to analyze the vibration signal and sound pressure signal of the diesel engine when it is working. However, when collecting the vibration signal and sound pressure signal, due to the complex working environment of the diesel engine and external interference, noise points will be generated. Therefore, it is necessary to filter out the interference signals in the collected vibration signal and sound pressure signal and eliminate them to obtain the real vibration signal and sound pressure signal, so that the subsequent analysis is more accurate.

[0018] See also Figure 1 , which shows a method flow chart of a signal analysis method for an electronically controlled single-cylinder diesel engine provided by an embodiment of the present invention, the method comprising the following steps: Step S1, the vibration amplitudes collected by vibration sensors at different positions at the same moment and the distances between the corresponding vibration sensors and the vibration source constitute a vibration attenuation ratio sequence at that moment; the regularity of the vibration amplitude at that moment is obtained according to every three adjacent elements in the vibration attenuation ratio sequence; the vibration change amplitude at that moment is obtained according to the vibration amplitude at a moment collected by a vibration sensor, the vibration amplitudes at the moment before and the moment after that moment.

[0019] In the embodiment of the present invention, a plurality of sensors are used to respectively collect signals of the diesel engine, including vibration signals and sound pressure signals. Two sensors are used in total, namely, a vibration sensor and a sound level meter. The installation positions of the two sensors should not affect the operating state of the diesel engine as much as possible, and at the same time, they can fully reflect the working information of the measured object. To ensure the acquisition accuracy, each type of sensor has multiple sensors, wherein the vibration sensor collects the vibration signal of the diesel engine. Preferably, in this embodiment, the vibration amplitude ZD is obtained from the vibration signal for analysis, and the sound signal of the diesel engine is collected by the sound level meter. Preferably, the sound signal in this embodiment is the sound pressure SY. At the same time, the power signal Y of the current diesel engine is obtained to ensure that the diesel engine is running at the same power.

[0020] When the diesel engine is subject to external interference, since the amount of collected data is large, the collected data often contains noise data, that is, noise points. Therefore, the present invention completes data analysis and identification based on the regularity and periodicity between the multiple collected data to eliminate the interference of noise data.

[0021] In the present invention, the continuous data collected by the same type of sensor is first analyzed. Here, the vibration amplitude ZD is taken as an example. Assuming that there are N vibration amplitudes collected by vibration sensors at different positions of the cylinder head, the vibration signal can be recorded as , It represents the vibration amplitude collected by the i-th vibration sensor at time T.

[0022] Then multiple vibration sensors can obtain the vibration amplitude at the same time, forming a sequence represented as , … In addition, it should be noted that the vibration amplitude and sound pressure collected by each vibration sensor and sound level meter are equal, and the vibration signal and sound signal are periodic. Even if a fault occurs, there will be a certain periodicity.

[0023] Ideally, the vibration amplitudes collected by vibration sensors at the same distance from the vibration source at the same time should be the same. However, during actual collection, due to different positions, that is, different distances from the vibration source, the data collected by the sensor may be interfered with, resulting in inconsistent detected data. Or when the vibration sensor is interfered with by external factors, the data detected by the sensor may also be inconsistent. In this case, noise occurs.

[0024] Multiple vibration amplitudes at the same time correspond to multiple vibration sensors, and the coordinate information of the vibration source is obtained ( , , ) and the coordinate information of each vibration sensor ( , ),in,( , ) represents the coordinate information of the i-th vibration sensor.

[0025] The distance between the vibration sensor and the vibration source can be obtained : in,( , ) is the coordinate information of the i-th vibration sensor, ( , , ) is the coordinate information of the vibration source, It represents the distance between the i-th vibration sensor and the vibration source, and the vibration sensors are sorted according to the distance between the vibration sensor and the vibration source.

[0026] Thus, the ratio of the vibration amplitude corresponding to one vibration sensor among all the vibration amplitudes of vibration sensors at different positions at the same time to the distance between the vibration sensor and the vibration source can be obtained. The ratio is the vibration attenuation ratio of the vibration sensor at that moment, which is expressed by the formula: in, It represents the vibration amplitude collected by the i-th vibration sensor at time T; represents the distance between the i-th vibration sensor and the vibration source; represents the vibration attenuation ratio of the i-th vibration sensor at time T; all vibration attenuation ratios corresponding to time T constitute a vibration attenuation ratio sequence.

[0027] Furthermore, based on the difference in the change amplitude between every three adjacent elements in the vibration attenuation ratio sequence, the regularity of the vibration amplitude at the corresponding moment of the vibration attenuation ratio sequence is obtained, specifically: the absolute value of the difference between the first element and the second element, and the absolute value of the difference between the second element and the third element in every three adjacent elements in the vibration attenuation ratio sequence are obtained; the opposite of the absolute value of the result of subtracting the absolute value of the difference between the second element and the third element from the absolute value of the difference between the first element and the second element is used as the power exponent of an exponential function with a natural constant as the base, and the function value of the exponential function with the natural constant as the base corresponding to every three adjacent elements is obtained; the function values ​​of the exponential function with the natural constant as the base corresponding to every three adjacent elements in the vibration attenuation ratio sequence are averaged to obtain the regularity of the vibration amplitude.

[0028] The formula is: in, Indicates the regularity of the vibration amplitude corresponding to time T; Indicates the number of vibration sensors; , and They represent the i-th, i+1-th and i+2-th elements in the vibration attenuation ratio sequence corresponding to time T respectively; e represents a natural constant, It indicates the difference between the change amplitudes of the first and second elements, and the second and third elements in three adjacent elements. The smaller the value, the greater the regularity of the vibration amplitude. The larger it is, the less likely it is that the vibration amplitude collected at that moment is a noise point.

[0029] Finally, the amplitude of the vibration amplitude change at a moment is calculated and analyzed to obtain the vibration amplitude change at a moment, specifically: the absolute value of the difference between the vibration amplitude at moment T and the vibration amplitude at the moment before this moment, plus the absolute value of the difference between the vibration amplitude at moment T and the vibration amplitude at the moment after this moment, to obtain the addition result; the ratio of the addition result to twice the absolute value of the vibration amplitude at moment T is the vibration amplitude change at this moment; it can be expressed as: in, It represents the vibration change amplitude of the vibration amplitude collected by any vibration sensor at time T, that is, the vibration change amplitude at time T; , and They respectively represent the vibration amplitudes at time T-1, time T, and time T+1 collected by a vibration sensor.

[0030] Step S2, the vibration change amplitudes of at least two moments before a moment and at least two moments after the moment constitute a vibration change amplitude sequence at the moment; the abnormality of the vibration amplitude at each moment is obtained according to the difference between every two adjacent elements in the vibration change amplitude sequence at each moment and the regularity of the vibration amplitude at each moment.

[0031] When determining whether the vibration amplitude at each moment collected by a vibration sensor is noise data, it is necessary to consider the vibration change amplitudes at least two moments in a period before the moment and the vibration change amplitudes at least two moments in a period after the moment. Specifically, the vibration change amplitudes at least two moments in a preset period before moment T of a vibration sensor and the vibration change amplitudes at least two moments in a preset period after moment T of a vibration sensor are obtained to form a vibration change amplitude sequence at moment T, which is expressed as { … … }, Represents a vibration sensor The amplitude of vibration change at each moment.

[0032] According to the vibration amplitude sequence at time T and the regularity of the vibration amplitude at time T Calculate the abnormality of the vibration amplitude of a vibration sensor at time T, specifically: take the opposite of the absolute value of the difference between every two adjacent elements in the vibration change amplitude sequence at time T as the power exponent of an exponential function with a natural constant as the base, and obtain the function value of the exponential function; average the difference between the function values ​​corresponding to every two adjacent elements in the vibration change amplitude sequence at time T, and multiply it by the regularity of the vibration amplitude at time T to obtain the abnormality of the vibration amplitude at time T; expressed by the formula: in, Indicates the abnormal degree of the vibration amplitude of a vibration sensor at time T; represents the Tth moment; and Represents the vibration change amplitude of the vibration amplitude collected by any vibration sensor at time T and time T+1; It indicates the regularity of the vibration amplitude corresponding to time T; the larger the YC is, the more likely the vibration amplitude at that moment is a noise signal; e represents a natural constant.

[0033] Step S3, based on the position of the vibration amplitude at a moment in its cycle, obtain the moments at the same position in a preset number of cycles adjacent to and before the cycle, to form a judgment sequence of the vibration amplitude at the moment; according to the vibration amplitude corresponding to each moment in the judgment sequence of the vibration amplitude at each moment and the regularity of the vibration amplitude, the abnormality degree of the vibration amplitude at each moment obtains the optimized abnormality degree of the vibration amplitude at each moment.

[0034] Since the present embodiment analyzes the vibration amplitude at a moment, and the vibration amplitude at a moment is within a cycle, it is necessary to further analyze the vibration amplitude within a preset number of cycles adjacent to and before the cycle to optimize the abnormality of the vibration amplitude at a moment; based on the position of the vibration amplitude at a moment in its cycle, the moments at the same position within a preset number of cycles adjacent to and before the cycle are obtained to form a judgment sequence of the vibration amplitude at the moment, and the elements in the judgment sequence are all moments, wherein the number of elements in the judgment sequence is M, for example, if moment T is the third moment in its corresponding cycle, then the third moment in each of the first three cycles of the cycle corresponding to moment T is taken, and the third moment in each of the last three cycles of the cycle corresponding to moment T is taken to form a judgment sequence, and the number of elements M in the judgment sequence is 6; in addition, it should be noted that the value of M can be adjusted by the implementer according to the specific situation.

[0035] The abnormal degree of the vibration amplitude at each moment is optimized by combining the judgment sequence to obtain the optimized abnormal degree of the vibration amplitude at each moment: in, and They respectively represent the optimized abnormality degree and abnormality degree of the vibration amplitude at time T collected by a sensor; The number of moments in the judgment sequence representing the vibration amplitude at time T; represents the vibration amplitude at time T; The vibration amplitude corresponding to the i-th element in the judgment sequence representing the vibration amplitude at time T; Indicates the regularity of the vibration amplitude at time T; The regularity of the vibration amplitude corresponding to the i-th element in the judgment sequence representing the vibration amplitude at time T; e represents a natural constant.

[0036] and The larger the value is, the higher the optimized abnormality of the vibration amplitude at that moment is, which means that the vibration amplitude at that moment is more likely to be noise and needs to be filtered out. The abnormality degree is optimized based on the data corresponding to each element in the judgment sequence, and a more accurate optimized abnormality degree is obtained, which makes the subsequent elimination of the vibration amplitude more accurate.

[0037] Step S4, screening the vibration amplitudes at different moments collected by a vibration sensor based on the optimized abnormality degree of the vibration amplitudes at different moments collected by a vibration sensor to obtain a real vibration amplitude sequence of the vibration sensor; then obtaining the sound pressures at different moments collected by sound level meters at different positions, screening to obtain a real sound pressure sequence of each sound level meter; and obtaining the fault type of the single-cylinder diesel engine by analyzing the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter.

[0038] First, the vibration amplitudes collected at different times by a vibration sensor are screened based on the optimized abnormality of the vibration amplitudes collected at different times by a vibration sensor to obtain the real vibration amplitude sequence of the vibration sensor, specifically: a judgment threshold is set, and the vibration amplitudes with optimized abnormality greater than the judgment threshold are removed from the vibration amplitudes collected at different times by a vibration sensor to form the real vibration amplitude sequence of the vibration sensor, so each vibration sensor will obtain a real vibration amplitude sequence, and in step S1, the sound pressure at different times is collected by different sound level meters, and the sound pressure at different times collected is processed in the same way as the vibration amplitude processing method, and the sound pressure data belonging to noise in the sound pressure data is screened out, so that the real sound pressure sequence of each sound level meter can be obtained. At this time, the real vibration amplitude sequence of each vibration sensor and the real sound pressure sequence of each sound level meter have eliminated the data belonging to noise, and more accurate analysis results can be obtained in the subsequent analysis process.

[0039] Furthermore, the fault type of the single-cylinder diesel engine is obtained by analyzing the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter, specifically: obtaining historical vibration amplitude data and sound pressure data. It should be noted that obtaining historical vibration amplitude data and sound pressure data is a well-known technology and will not be elaborated on here.

[0040] The probability that the vibration amplitude at a certain moment in the real vibration amplitude sequence of any vibration sensor appears in the historical vibration amplitude data and the probability that the sound pressure at the same moment in the real sound pressure sequence of any sound level meter appears in the historical sound pressure data are obtained, which are the vibration history probabilities and sound pressure history probability ; Obtain the frequency of the vibration amplitude at this moment within a preset number of cycles adjacent to and before the cycle where the vibration amplitude at this moment is located, and the ratio of the frequency to the number of elements in the judgment sequence at this moment is recorded as the first probability ; Obtain the frequency of the sound pressure at this moment in the preset number of cycles adjacent to the cycle before and after the sound pressure at this moment, and the ratio of the frequency to the number of elements in the judgment sequence at this moment is recorded as the second probability The first probability sum at this moment is multiplied by the difference in the vibration history probability at this moment to obtain the vibration abnormality probability at this moment, which is expressed as follows: in, Indicates the probability of abnormal vibration at a certain moment; represents the first probability corresponding to the moment, Represents the vibration history probability corresponding to the moment. The larger the value is, the greater the frequency of the vibration amplitude at this moment in the adjacent cycles is. At the same time, the smaller the frequency of the vibration amplitude at this moment in the historical data is, the greater the possibility that the vibration amplitude at this moment is noise data.

[0041] The probability of abnormal sound pressure at that moment is obtained by multiplying the difference between the second probability sum and the sound pressure history probability at that moment; it can be expressed as: in, represents the probability of abnormal sound pressure at a moment; represents the second probability corresponding to this moment, Represents the sound pressure history probability corresponding to that moment.

[0042] The product of the probability of abnormal vibration and the probability of abnormal sound pressure at a moment is the failure probability at that moment, which can be expressed as: in, Indicates the probability of abnormal vibration at a moment; Indicates the probability of abnormal sound pressure at a moment; Indicates the failure probability at that moment. The greater the failure probability, the greater the possibility of failure at that moment. Set a first threshold. If the failure probability at that moment is greater than the first threshold, then that moment is the failure moment. Preferably, in the embodiment of the present invention, the value of the first threshold is 0.87. The implementer can adjust the value of the first threshold according to the actual situation.

[0043] Finally, the vibration amplitude and sound pressure at all fault moments in the real vibration amplitude sequence and real sound pressure sequence of any vibration sensor and sound level meter are obtained to form a fault feature vector; the fault feature vector is input into the diesel engine fault classification neural network, and the fault type of the diesel engine is output. The diesel engine fault classification neural network training process is a well-known technology and will not be repeated here, because it is used for classification, and the loss function of this neural network is the cross entropy loss function.

[0044] According to the above steps, the diesel engine signal is collected by the sensor, and the signal is preprocessed to obtain the corresponding fault feature vector, the fault is identified according to the diesel engine fault classification neural network, and relevant repairs are performed according to the fault.

[0045] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0046] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0047] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the present invention should be included in the protection scope of the present invention.

Claims

1. A signal analysis method for an electronically controlled single-cylinder diesel engine, characterized in that: The method includes: The vibration amplitudes collected by vibration sensors at different positions at the same time and the distances between the corresponding vibration sensors and the vibration source constitute a vibration attenuation ratio sequence at that moment; the regularity of the vibration amplitude at that moment is obtained according to every three adjacent elements in the vibration attenuation ratio sequence; the vibration change amplitude at that moment is obtained according to the vibration amplitude at a moment collected by a vibration sensor, the vibration amplitudes at the moment before and the moment after that moment; The vibration change amplitudes at at least two moments before a moment and at least two moments after the moment constitute a vibration change amplitude sequence at the moment; the abnormality of the vibration amplitude at each moment is obtained according to the difference between every two adjacent elements in the vibration change amplitude sequence at each moment and the regularity of the vibration amplitude at each moment; Based on the position of the vibration amplitude at a moment in its cycle, the moments at the same position in a preset number of cycles adjacent to and before the cycle are obtained to form a judgment sequence of the vibration amplitude at the moment; according to the vibration amplitude corresponding to each moment in the judgment sequence of the vibration amplitude at each moment and the regularity of the vibration amplitude, the abnormality degree of the vibration amplitude at each moment is obtained to obtain the optimized abnormality degree of the vibration amplitude at each moment; The vibration amplitudes at different moments collected by a vibration sensor are screened based on the optimized abnormality degree of the vibration amplitudes at different moments collected by a vibration sensor to obtain the real vibration amplitude sequence of the vibration sensor; then the sound pressures at different moments collected by sound level meters at different positions are obtained, and the real sound pressure sequence of each sound level meter is screened to obtain; the fault type of a single-cylinder diesel engine is obtained by analyzing the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter.

2. A signal analysis method for an electronically controlled single-cylinder diesel engine according to claim 1, characterized in that: The step of obtaining the regularity of the vibration amplitude at the moment according to every three adjacent elements in the vibration attenuation ratio sequence includes: The absolute value of the difference between the first element and the second element and the absolute value of the difference between the second element and the third element in every three adjacent elements in the vibration attenuation ratio sequence are obtained; the absolute value of the difference between the first element and the second element is obtained, and the absolute value of the result of subtracting the absolute value of the difference between the second element and the third element is subtracted, and the opposite number of the absolute value of the result is used as the power exponent of the exponential function with the natural constant as the base to obtain the function value of the exponential function with the natural constant as the base corresponding to every three adjacent elements; the function values ​​of the exponential function with the natural constant as the base corresponding to every three adjacent elements in the vibration attenuation ratio sequence are averaged to obtain the regularity of the vibration amplitude.

3. The signal analysis method for an electronically controlled single-cylinder diesel engine according to claim 1, characterized in that: The method of obtaining the vibration change amplitude at a moment based on the vibration amplitude at a moment collected by a vibration sensor and the vibration amplitudes at the moment before and after the moment includes: adding the absolute value of the difference between the vibration amplitude at a moment and the vibration amplitude at the moment before the moment, and the absolute value of the difference between the vibration amplitude at a moment and the vibration amplitude at the moment after the moment, to obtain an addition result; the ratio of the addition result to twice the absolute value of the vibration amplitude at the moment is the vibration change amplitude at the moment.

4. The signal analysis method for an electronically controlled single-cylinder diesel engine according to claim 1, characterized in that: The method of obtaining the abnormality of the vibration amplitude at each moment according to the difference between every two adjacent elements in the vibration change amplitude sequence at each moment and the regularity of the vibration amplitude at each moment includes: taking the opposite number of the absolute value of the difference between every two adjacent elements in the vibration change amplitude sequence as the power exponent of an exponential function with a natural constant as the base to obtain the function value of the exponential function; averaging the difference between the function values ​​corresponding to every two adjacent elements in the vibration change amplitude sequence at a moment, and multiplying the average value by the regularity of the vibration amplitude at the moment to obtain the abnormality of the vibration amplitude at the moment.

5. The signal analysis method for an electronically controlled single-cylinder diesel engine according to claim 1, characterized in that: The optimized abnormal degree of the vibration amplitude at each moment is: in, and They respectively represent the optimized abnormality degree and abnormality degree of the vibration amplitude at time T collected by a sensor; The number of moments in the judgment sequence representing the vibration amplitude at time T; represents the vibration amplitude at time T; The vibration amplitude corresponding to the i-th element in the judgment sequence representing the vibration amplitude at time T; Indicates the regularity of the vibration amplitude at time T; The regularity of the vibration amplitude corresponding to the i-th element in the judgment sequence representing the vibration amplitude at time T; and the elements in the judgment sequence are time; and e represents a natural constant.

6. The signal analysis method for an electronically controlled single-cylinder diesel engine according to claim 1, characterized in that: The method of screening the vibration amplitudes at different moments collected by a vibration sensor based on the optimized abnormality degree of the vibration amplitudes at different moments collected by the vibration sensor to obtain the real vibration amplitude sequence of the vibration sensor includes: setting a judgment threshold, removing the vibration amplitudes at different moments collected by the vibration sensor whose optimized abnormality degree is greater than the judgment threshold, and forming the real vibration amplitude sequence of the vibration sensor.

7. The signal analysis method for an electronically controlled single-cylinder diesel engine according to claim 1, characterized in that: The method of analyzing the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and the sound level meter to obtain the fault type of the single-cylinder diesel engine includes: obtaining historical vibration amplitude data and sound pressure data; obtaining the probability of the vibration amplitude at a moment in the real vibration amplitude sequence of any vibration sensor appearing in the historical vibration amplitude data and the probability of the sound pressure at the same moment in the real sound pressure sequence of any sound level meter appearing in the historical sound pressure data, which are vibration history probability and sound pressure history probability respectively; The frequency of the vibration amplitude at the moment within a preset number of cycles adjacent to the cycle before and after the vibration amplitude at the moment is obtained, and the ratio of the frequency to the number of elements in the judgment sequence at the moment is recorded as the first probability; the frequency of the sound pressure at the moment within a preset number of cycles adjacent to the cycle before and after the sound pressure at the moment is obtained, and the ratio of the frequency to the number of elements in the judgment sequence at the moment is recorded as the second probability; The first probability sum at the moment is multiplied by the difference between the vibration history probability at the moment to obtain the vibration abnormality probability at the moment; the second probability sum at the moment is multiplied by the difference between the sound pressure history probability at the moment to obtain the sound pressure abnormality probability at the moment; the product of the vibration abnormality probability and the sound pressure abnormality probability at the moment is the fault probability at the moment; a first threshold is set, and if the fault probability at the moment is greater than the first threshold, the moment is the fault moment; the vibration amplitude and sound pressure at all fault moments in the real vibration amplitude sequence and the real sound pressure sequence of any vibration sensor and sound level meter are obtained to form a fault feature vector; the fault feature vector is input into the diesel engine fault classification neural network, and the fault type of the diesel engine is output.

Images