A method and system for evaluating noise quality of a flow electric drive system

Abstract

The application relates to the technical field of noise evaluation, and discloses a process-based noise quality evaluation method and system for an electric driving system, which comprises the following steps: collecting an in-vehicle sound pressure time domain signal of the electric driving system in a variable rotating speed process and a corresponding rotating speed signal, resampling the sound pressure time domain signal based on the rotating speed signal, determining the center frequency of different rotating speeds based on a target order spectrum, determining an order energy curve according to an order locking frequency band, expanding outward based on the center frequency as a symmetric reference, eliminating the frequency range of the order locking frequency band in the reference frequency band, searching for noise results along the rotating speed direction, identifying intervals with noise results higher than zero, and obtaining corresponding peak amplitudes and interval widths in each interval. The peak amplitudes and interval widths of all intervals are synthesized, and the noise quality of the electric driving system is evaluated based on a noise evaluation value. The application ensures the reliability of the noise quality evaluation of the electric driving system.

Description

Technical Field

[0001] This invention relates to the field of noise assessment technology, and more specifically, to a process-oriented method and system for evaluating the noise quality of electric drive systems. Background Technology

[0002] With the rapid development of new energy vehicles, electric drive systems are gradually replacing traditional engines as the power source for vehicles. During operation, electric drive systems contain multiple rotating components such as motors, electronic controls, and reduction gears. These components generate order noise related to speed changes under different rotational speeds. Because electric vehicles lack the masking effect of engine noise, order noise is more easily perceived in the in-vehicle environment, becoming a significant factor affecting the overall vehicle sound quality. Current technologies typically evaluate electric drive system noise using spectrum analysis or order tracking techniques to extract sound pressure levels of specific orders and compare them with empirical thresholds. However, in actual operation, electric drive system noise is superimposed on other order noises, structural noise, and environmental background noise. Under variable speed operating conditions, order noise exhibits a dynamic distribution that shifts with rotational speed. Judging solely based on the amplitude at a single moment or rotational speed is insufficient to reflect the continuous characteristics and trends of noise during speed changes, thus failing to accurately evaluate the noise of the electric drive system.

[0003] Therefore, it is necessary to design a process-oriented method and system for evaluating the noise quality of electric drive systems to solve the problems existing in the current technology. Summary of the Invention

[0004] In view of this, the present invention proposes a process-oriented method and system for evaluating the noise quality of electric drive systems. It aims to solve the problem that the noise of electric drive systems is superimposed with other order noises, structural noises and environmental background noises. Judging the noise based on the amplitude at a certain moment or a certain speed is difficult to reflect the continuous characteristics and changing trend of the noise during the speed change process, thus making it impossible to accurately evaluate the noise of electric drive systems.

[0005] In one aspect, the present invention proposes a method for evaluating the noise quality of a process-oriented electric drive system, comprising: The in-vehicle sound pressure time-domain signal and the corresponding speed signal of the electric drive system during the variable speed process are collected, and the sound pressure time-domain signal is resampled based on the speed signal to determine the target order spectrum; The center frequencies of different rotational speeds are determined based on the target order spectrum, and an order-locked frequency band is established based on the center frequency. The order energy curve is determined according to the order-locked frequency band, and a reference frequency band is determined by expanding outward with the center frequency as a symmetrical reference. The frequency range of the order-locked frequency band is removed from the reference frequency band, and the background energy curve is determined according to the remaining frequency. The noise result is determined based on the order energy curve and the background energy curve, and the noise result is searched along the rotation direction to identify the intervals where the noise result is higher than zero, and the corresponding peak amplitude and interval width are obtained in each interval. The peak amplitude and interval width of all intervals are combined to determine the noise evaluation value, and the noise quality of the electric drive system is evaluated based on the noise evaluation value.

[0006] Furthermore, when resampling the sound pressure time-domain signal to determine the target order spectrum, the process includes: Extract the instantaneous rotational speed increment of adjacent sampling points and determine the angular velocity. Based on the angular velocity, perform angular domain mapping on the sampling points of the sound pressure time domain signal, and determine the equiangular sampling position according to the angular step size. Interpolate the sound pressure time domain signal at the equiangular sampling position to determine the equiangular sound pressure sequence. The equal-angle sound pressure sequence is divided into sliding windows, and the data of each sliding window is transformed in the frequency domain to determine the order spectrum with the order as the horizontal axis and the rotation speed as the vertical axis. The order spectra of each sliding window are then spliced ​​together to determine the target order spectrum.

[0007] Furthermore, when determining the center frequency of different rotational speeds based on the target order spectrum, the process includes: In the target order spectrum, the spectral line position corresponding to the target order is selected as the initial search position, and the maximum point near the initial search position is searched segment by segment along the rotation direction to determine the peak order position corresponding to each rotation window. When the offset of adjacent peak positions exceeds the offset threshold, a secondary search range is set on both sides of the peak position of the previous rotation window and the peak position is repositioned. Based on the repositioned peak order position and the rotation speed of the rotation window, the center frequency sequence corresponding to the target order in the rotation window is determined, and the center frequency sequence is interpolated to complete the center frequency trajectory.

[0008] Furthermore, in establishing the order-locked frequency band and determining the order energy curve, the following steps are included: Within each rotation speed window, an upper and lower limit offset of the center frequency are set to determine the upper and lower limits of the order-locked frequency band. The spectral amplitude points between the upper and lower limits are extracted from the target order spectrum. Based on the amplitude points and the frequency step size, the order energy value of the rotation speed window is determined. The order energies of each rotation speed window are arranged to determine the order energy curve.

[0009] Furthermore, when determining the reference frequency band and background energy curve by extending outward with the center frequency as a symmetrical reference, the process includes: Within each speed window, the frequency is extended symmetrically to the high-frequency and low-frequency sides using the center frequency as a reference to determine the upper and lower boundaries. A reference frequency band is determined based on the upper and lower boundaries, and the reference frequency band is divided into a low-frequency reference sub-band and a high-frequency reference sub-band. Frequency intervals that overlap with the order-locked frequency band and frequency intervals that intersect with the center frequency trajectories of other orders are removed from the low-frequency and high-frequency reference sub-bands. The background energy value of the speed window is determined based on the removal results, and the background energy curve is determined based on the background energy value of each speed window.

[0010] Furthermore, when determining the noise result based on the order energy curve and the background energy curve, the following steps are included: The order energy curve and the background energy curve are aligned by rotational speed coordinates, and the order energy values ​​at the same rotational speed position are normalized with the background energy values. Sample points with values ​​greater than zero are marked as positive sample points, and sample points with values ​​less than or equal to zero are marked as non-positive sample points. The noise result curve is determined based on the positive sample points.

[0011] Furthermore, when identifying the interval where the noise result is above zero, the process includes: Traverse the noise result curve along the rotational speed direction, take the first occurrence of the positive sample point as the candidate starting point and the end of the positive sample point as the candidate ending point, determine several candidate intervals, and remove candidate intervals whose rotational speed span is less than the minimum interval width threshold, and determine the effective intervals based on the removal results.

[0012] Furthermore, when obtaining the corresponding peak amplitude and interval width within each interval, the following steps are included: Each valid interval is truncated to determine the peak amplitude, and the interval width is determined based on the difference between the starting and ending rotational speeds of the valid interval. The peak amplitude and interval width are then combined into interval parameter pairs according to the interval number.

[0013] Furthermore, when synthesizing the peak amplitude and interval width of all intervals to determine the noise evaluation value, and evaluating the noise quality of the electric drive system based on the noise evaluation value, the process includes: The peak amplitude and interval width corresponding to each effective interval are multiplied to determine the interval contribution value. The contribution values ​​of all intervals are weighted and calculated to determine the noise evaluation value. The noise evaluation value is compared with the noise evaluation threshold, and the evaluation report is determined based on the comparison results.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: By collecting the in-vehicle sound pressure time-domain signal and the corresponding rotational speed signal, and resampling the sound pressure time-domain signal based on the rotational speed, the order noise can be continuously tracked in a unified coordinate system under variable rotational speed conditions. An order-locked frequency band is established around the center frequency to extract order energy, and background energy in the adjacent frequency domain environment of the target order is obtained through a reference frequency band. This establishes a correspondence between the target order and surrounding non-target components at the same rotational speed position. Furthermore, by using interval recognition, discrete instantaneous changes are transformed into continuous segments with rotational speed ranges, reflecting the continuous characteristics of noise. The peak amplitude and interval width of each segment are uniformly synthesized, transforming the complex variable rotational speed order change process into a unified evaluation index. This transforms noise evaluation from a local judgment relying on a single moment or a single rotational speed point to a holistic analysis covering the entire variable rotational speed process. The evaluation results have consistent calculation logic across different test conditions and different vehicles. Simultaneously, since the noise evaluation value originates from the comparison between order energy and background energy, the evaluation results can reflect the prominence of the target order in the actual operating environment, thereby ensuring the reliability of the noise quality evaluation of the electric drive system.

[0015] On the other hand, this application also provides a noise quality evaluation system for a process-oriented electric drive system, used to apply the above-mentioned noise quality evaluation method for a process-oriented electric drive system, including: The acquisition unit is configured to acquire the in-vehicle sound pressure time-domain signal and the corresponding speed signal of the electric drive system during the variable speed process, and to resample the sound pressure time-domain signal based on the speed signal to determine the target order spectrum; The analysis unit is configured to determine the center frequency of different rotational speeds based on the target order spectrum, establish an order-locked frequency band based on the center frequency, determine the order energy curve according to the order-locked frequency band, expand outward with the center frequency as a symmetrical reference to determine a reference frequency band, remove the frequency range of the order-locked frequency band within the reference frequency band, and determine the background energy curve based on the remaining frequency. The processing unit is configured to determine the noise result based on the order energy curve and the background energy curve, and to search the noise result along the rotation direction to identify the intervals where the noise result is higher than zero, and to obtain the corresponding peak amplitude and interval width in each interval. The evaluation unit is configured to synthesize the peak amplitude and interval width of all intervals, determine the noise evaluation value, and evaluate the noise quality of the electric drive system based on the noise evaluation value.

[0016] It is understandable that the above-mentioned method and system for evaluating the noise quality of a process-oriented electric drive system have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a process-oriented method for evaluating the noise quality of an electric drive system provided in an embodiment of the present invention; Figure 2 This is a functional block diagram of a process-oriented electric drive system noise quality evaluation system provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] See Figure 1 As shown in some embodiments of this application, a method for evaluating the noise quality of a process-oriented electric drive system includes: S100: Acquires the in-vehicle sound pressure time-domain signal and the corresponding speed signal of the electric drive system during the variable speed process, and resamples the sound pressure time-domain signal based on the speed signal to determine the target order spectrum.

[0022] S200: Determine the center frequency of different rotational speeds based on the target order spectrum, establish an order-locked frequency band based on the center frequency, determine the order energy curve based on the order-locked frequency band, and expand outward with the center frequency as the symmetry reference to determine the reference frequency band. Eliminate the frequency range of the order-locked frequency band within the reference frequency band, and determine the background energy curve based on the remaining frequency.

[0023] S300: Determine the noise result based on the order energy curve and the background energy curve, and search for the noise result along the rotation direction to identify the interval where the noise result is higher than zero, and obtain the corresponding peak amplitude and interval width in each interval.

[0024] S400: Combine the peak amplitude and interval width of all intervals to determine the noise evaluation value, and evaluate the noise quality of the electric drive system based on the noise evaluation value.

[0025] Specifically, the system synchronously acquires the in-vehicle sound pressure time-domain signal and the corresponding speed signal during the variable speed process of the electric drive system. The in-vehicle sound pressure time-domain signal represents the original waveform of the sound pressure changing with time, continuously recorded by the in-vehicle microphone in the time dimension. It can completely include the superposition result of electric drive system noise, vehicle body structure noise and environmental background noise. The speed signal represents the instantaneous speed change trajectory of the drive motor or equivalent rotating shaft during the variable speed process, which describes the law of order noise migration with speed. The sound pressure time-domain signal is resampled based on the rotational speed signal. Instead of sampling at the original equal time intervals, a new sampling benchmark is established according to the rotational speed change, so that the sound pressure time-domain signal forms a consistent sampling frequency in the rotational speed dimension. This transforms the order noise that moves with the rotational speed into spectral features aligned under a unified coordinate system. After resampling, the target order spectrum is determined. The target order spectrum is the energy distribution result with the order as the analysis axis and the rotational speed as the expansion axis. The order represents the harmonic relationship between the noise frequency and the rotational speed. When the rotational speed of the electric drive system changes, the noise frequency corresponding to the same order will change synchronously with the rotational speed. Therefore, the target order spectrum can fix the order noise that originally moved in the frequency to the order, so as to continuously track the energy change of the same order throughout the entire speed change process. The noise component is transformed into the target order spectrum that is continuously observed along the rotational speed, so that subsequent calculations no longer depend on a single moment or a single rotational speed point, reducing the error caused by order migration from the source.

[0026] Understandably, after obtaining the target order spectrum, the center frequency at different rotational speeds is determined based on the target order spectrum. The center frequency represents the frequency center position corresponding to the order at each rotational speed position. It changes with the rotational speed and is used to indicate the frequency point where the target order noise should be extracted at that rotational speed. The order-locked frequency band is a symmetrical frequency range set around the center frequency. It is used to completely delineate the energy of the order noise near that rotational speed position, avoiding mistaking adjacent orders or discrete background components as order noise. The order energy curve represents the curve result formed by continuously collecting the energy in the order-locked frequency band at each rotational speed position along the rotational speed sequence throughout the entire rotational speed process. The energy is the cumulative amount of sound pressure components in that frequency band. Therefore, the order energy curve directly represents the strength change of the order noise during the rotational speed process. The reference frequency band is determined by expanding outwards from the center frequency as a symmetrical reference. This wider frequency range, centered on the center frequency, provides a background energy reference for the target order within its adjacent frequency domain. The frequency range of the order-locked band is then removed from the reference band; that is, the frequency segment containing the target order itself is removed. The remaining frequencies contain noise components near but not belonging to the target order. The background energy curve represents the energy change trajectory of these remaining frequencies during the variable speed process. The noise result is determined based on the order energy curve and the background energy curve. This difference, calculated at the same speed position, reflects the prominence of the target order relative to the surrounding background. For example, if the order noise near a certain speed is significantly higher than the surrounding noise in the reference frequency band, the noise result at that speed position is positive. Conversely, if the target order noise is close to or covered by the surrounding background, the noise result at that speed position is not positive. The noise results are searched along the rotational speed direction to identify the intervals where the noise results are higher than zero. An interval higher than zero means that the target order is continuously higher than the surrounding background within this rotational speed range. The corresponding peak amplitude and interval width are obtained for each interval. The peak amplitude represents the highest level of the noise result within the interval, and the interval width represents the coverage range of the interval on the rotational speed axis. For example, if the noise result is continuously positive within a certain rotational speed range and there is a peak point in the middle, then the peak point corresponds to the peak amplitude, and the entire rotational speed range that is continuously positive corresponds to the interval width.The peak amplitude and interval width of all intervals are synthesized so that the noise evaluation value can simultaneously include two types of information: prominence and duration. The larger the noise evaluation value, the more significant the target order appears during the speed change process and the longer the positive value interval is covered. The smaller the noise evaluation value, the more briefly the target order appears or the difference from the background is not obvious. By eliminating the order-locked frequency band and the reference frequency band, the target order and the adjacent background are separated under the same speed coordinate, so that the noise result directly represents the difference between the target order and the background. Then, by extracting the peak amplitude and interval width of the interval with noise result above zero value and synthesizing them into a noise evaluation value, the noise quality evaluation of the electric drive system can cover the entire speed change process, ensuring the reliability of the evaluation results.

[0027] In some embodiments of this application, when resampling the sound pressure time-domain signal to determine the target order spectrum, the process includes: extracting the instantaneous rotational speed increment of adjacent sampling points and determining the angular velocity; performing angular domain mapping on the sampling points of the sound pressure time-domain signal based on the angular velocity; determining the iso-angle sampling positions according to the angular step size; interpolating the sound pressure time-domain signal at the iso-angle sampling positions to determine the iso-angle sound pressure sequence; dividing the iso-angle sound pressure sequence into sliding windows; performing frequency domain transformation on the data of each sliding window; determining the order spectrum with the order as the abscissa and the rotational speed as the ordinate; and splicing together all the order spectra of each sliding window to determine the target order spectrum.

[0028] Specifically, the instantaneous speed increments of adjacent sampling points are extracted from the speed signal, and the angular velocity is determined. The instantaneous speed increment represents the increment of speed change between two adjacent sampling moments, reflecting the uniformity of speed change during the variable speed process. The angular velocity represents the rate of change of angular displacement per unit time, corresponding to the instantaneous speed increment, enabling subsequent processing to convert the time-domain sampling points to angular-domain positions. Based on the angular velocity, angular-domain mapping is performed on the sampling points of the sound pressure time-domain signal. Angular-domain mapping assigns a corresponding angular position to each sampling point in the sound pressure time-domain signal, so that the sampling point no longer represents only the sound pressure at a certain moment, but rather the sound pressure at a certain angular position. For example, when the electric drive system experiences speed fluctuations during acceleration, the angular increments corresponding to adjacent sampling points in time may vary greatly. Angular-domain mapping can explicitly reflect this non-uniformity in the angular coordinates, thus preparing for subsequent equal-angle sampling. The isoangular sampling positions are determined based on the angular step size, which represents the fixed angular interval between two adjacent target angular sampling positions. The isoangular sampling positions represent a set of angular coordinate points arranged at this fixed angular interval along the rotation angle. Switching the sampling reference from equal time intervals to equal angular intervals ensures that sound pressure data from different speed ranges are extracted with the same angular resolution during the variable speed process. Interpolation is performed on the sound pressure time-domain signal at the isoangular sampling positions to determine the isoangular sound pressure sequence. Interpolation is performed when the angular position of the original sound pressure time-domain signal's sampling point does not completely coincide with the isoangular sampling position; the sound pressure value at the isoangular sampling position is reconstructed in the angular domain using the sound pressure values ​​of adjacent sampling points. The isoangular sound pressure sequence represents the sound pressure sequence obtained by arranging the isoangular sampling positions in order, with each sample data corresponding to a fixed angular step size. A sliding window is used to divide the isoangular sound pressure sequence; the sliding window is a continuous segment extracted from the isoangular sound pressure sequence, and adjacent segments move at a fixed step size. The sliding window allows the variable speed process to be divided into multiple local intervals, ensuring that the isoangular sound pressure sequence within each interval has processing length and local stability. Frequency domain transformation is performed on the data of each sliding window. This transformation converts the isoangular sound pressure sequence within the sliding window from the angle domain / isoangular sampling domain to the frequency-dependent spectral domain, causing periodic components to appear as peaks in the spectral domain. In order analysis, since the isoangular sound pressure sequence is synchronized with the rotation angle, the horizontal axis of the spectral domain output by the frequency domain transformation corresponds to the order. This determines the order spectrum with the order as the horizontal axis and the rotation speed as the vertical axis. The order spectrum represents the energy distribution of different order components at different rotation speed positions. The horizontal axis of the order indicates that the spectral domain position corresponds to the order component that is multiple of the rotation speed, and the vertical axis of the rotation speed represents the trajectory of the order component as it unfolds along the variable rotation speed process.Finally, the order spectra of all sliding windows are spliced ​​together to determine the target order spectrum. The splicing is performed according to the order of the sliding windows during the variable speed process, continuously connecting the order spectra obtained from each sliding window along the speed dimension to form a continuous order spectrum covering the entire variable speed process. The target order spectrum represents the spectral information related to the target order retained from the spliced ​​order spectrum or the order spectrum result with the target order as the object of interest. By using angular domain mapping and equal-angle sampling positions, the sound pressure time-domain signal is switched from the time reference to the rotation angle reference, ensuring continuous tracking of components corresponding to the same order in the spectral domain during the variable speed process. Furthermore, equal-angle sound pressure sequences are obtained through interpolation, ensuring that the data within the sliding windows meet the equal-angle sampling condition. This facilitates stable order spectra obtained through frequency domain transformation, ensuring the reliability of the target order spectrum covering the entire variable speed process.

[0029] In some embodiments of this application, when determining the center frequency of different rotational speeds based on the target order spectrum, the method includes: selecting the spectral line position corresponding to the target order in the target order spectrum as the initial search position, and searching for the maximum point near the initial search position segment by segment along the rotational speed direction to determine the peak order position corresponding to each rotational speed window. When the offset between adjacent peak positions exceeds the offset threshold, a secondary search range is set on both sides of the peak position of the previous rotational speed window and the peak position is repositioned. The center frequency sequence corresponding to the target order in the rotational speed window is determined based on the repositioned peak order position and the rotational speed of the rotational speed window, and the center frequency sequence is interpolated to complete the center frequency trajectory.

[0030] Specifically, the spectral line position corresponding to the target order is selected as the initial search position in the target order spectrum. The spectral line position corresponding to the target order represents the continuous high-energy trajectory related to the target order in the energy distribution result. The initial search position represents the order coordinate region used to limit the search starting point when the tracking begins, in order to avoid mistakenly selecting non-target order components when order noise is superimposed or background energy fluctuates. Subsequently, the maximum point near the initial search position is searched segment by segment along the rotation direction. The segment by segment along the rotation direction means dividing the target order spectrum into multiple adjacent rotation window according to the order of rotation from low to high or from high to low. The rotation window represents a continuous segment divided on the rotation axis. The maximum point represents the local maximum energy point that appears in the order range close to the initial search position within the rotation window. This maximum point corresponds to the most significant energy concentration position of the target order within the rotation window. By performing a search on each rotational speed window, the peak order position corresponding to each window is determined. The peak order position represents the position of the maximum point in order within that rotational speed window, reflecting the actual shift of the target order within that window. It should be noted that multiple adjacent spectral lines or local energy spikes may appear in the actual order spectrum. Relying solely on a single search can easily lead to misjudging local peaks of adjacent orders as the target order. Therefore, when the shift of adjacent peak positions exceeds the shift threshold, that is, when the difference in order between the peak order positions of two adjacent rotational speed windows exceeds the maximum allowable shift range of the target order spectral line between adjacent rotational speed windows, it means that the peak order position of the current window is... Inconsistencies in continuity with the previous window may be caused by instantaneous noise superposition, structural resonance, or interference from adjacent orders. In such cases, a secondary search range is set on both sides of the peak position of the previous rotational speed window, and the peak position is repositioned. A narrower order search interval is obtained by expanding outwards from the peak order position of the previous rotational speed window. By locking the search range near the previous window, the repositioned peak order position maintains continuity with the historical positioning results. For example, when strong background energy appears in a certain rotational speed window, causing two local peaks to appear near the initial search position, the secondary search range limits the search to the neighborhood of the peak order position of the previous rotational speed window, ensuring that the repositioned peak order position falls on a continuous spectral line. Based on the repositioned peak order position and the rotational speed of the rotational speed window, the center frequency sequence corresponding to the target order in that rotational speed window is determined. The center frequency sequence is obtained by converting the peak order position of each rotational speed window with the rotational speed of that window, so that the frequency center position of the target order at different rotational speeds is determined window by window. Finally, the center frequency sequence is interpolated and completed to form a continuous curve of center frequency change without any breaks. The center frequency trajectory represents the continuous path of center frequency change with speed covering the entire variable speed range.The peak order position of each speed window is obtained by searching the initial search position and the segmented maximum point search, so that the center frequency sequence can be extracted as the speed changes. Furthermore, the secondary search range repositioning triggered by the offset threshold suppresses the jump caused by adjacent order interference or local protrusions. This ensures that the corresponding center frequency can be obtained at each speed position when establishing the order-locked frequency band. This ensures that the extraction process of the order energy curve and the background energy curve has a consistent frequency alignment basis, further ensuring the reliability of the noise quality evaluation of the electric drive system.

[0031] In some embodiments of this application, the process of establishing an order-locked frequency band and determining the order energy curve includes: setting an upper limit offset and a lower limit offset of the center frequency in each rotation speed window, determining the upper limit frequency and lower limit frequency of the order-locked frequency band, extracting the spectral amplitude points between the upper limit frequency and the lower limit frequency in the target order spectrum, determining the order energy value of the rotation speed window based on the amplitude points and the frequency step size, and arranging the order energies of each rotation speed window to determine the order energy curve.

[0032] Specifically, for each speed window, an upper and lower center frequency offset are set to allow for segmented processing of the variable speed process. These offsets represent frequency deviations set around the center frequency towards higher and lower frequencies, respectively, to limit the frequency range to be extracted. This prevents the target order energy from being truncated due to an excessively narrow extraction range, or from being mixed with adjacent orders and background components due to an excessively wide extraction range. Based on these offsets, the upper and lower limits of the order-locked frequency band are determined. The upper frequency represents the frequency boundary after adding the upper offset to the center frequency, and the lower frequency represents the frequency boundary after subtracting the lower offset from the center frequency. The order-locked frequency band is the frequency range jointly defined by the upper and lower limits, symmetrically set around the center frequency within each speed window to ensure that the target order's frequency point within that window is always covered. The spectral amplitude points between the upper and lower frequency limits are extracted from the target order spectrum. These amplitude points represent discrete frequency points within the speed window that fall within the order-locked frequency band. The extraction of amplitude points is not limited to a single frequency point, but rather involves multiple frequency points within a frequency band to cover the frequency spread and sideband components generated by the target order under actual noise superposition conditions. For example, when the electric drive system experiences slight modulation or structural coupling within a certain speed window, the target order energy may not be concentrated at a single point but distributed across multiple frequency points near the center frequency. The order-locked frequency band can incorporate these distributions. The order energy value of the speed window is determined based on the amplitude points and the frequency step size. The frequency step size represents the frequency interval between adjacent spectral amplitude points, determining the integration scale when accumulating amplitude points. The order energy value represents the result of integrating the spectral amplitude points between the upper and lower frequency limits according to the frequency step size, essentially restoring the discrete spectral amplitude points to the total energy quantization value of the frequency band, thus obtaining the energy of the target order within the speed window. The energy values ​​of each order are connected sequentially according to the speed window. Therefore, the order energy curve represents the continuous curve result of the target order energy changing with the speed, which further ensures the reliability of the noise quality evaluation of the electric drive system.

[0033] In some embodiments of this application, when determining the reference frequency band and background energy curve by expanding outward with the center frequency as a symmetric reference, the process includes: expanding outward with the center frequency as a symmetric reference towards the high-frequency side and the low-frequency side within each speed window to determine the upper and lower boundaries; determining the reference frequency band based on the upper and lower boundaries; dividing the reference frequency band into a low-frequency reference sub-band and a high-frequency reference sub-band; eliminating frequency intervals that overlap with the order-locked frequency bands in the low-frequency and high-frequency reference sub-bands, as well as frequency intervals that intersect with the center frequency trajectories of other orders; determining the background energy value of the speed window based on the elimination results; and determining the background energy curve based on the background energy value of each speed window.

[0034] Specifically, within each speed window, the frequency is extended symmetrically towards the high-frequency and low-frequency sides from the center frequency to determine upper and lower boundaries. The upper boundary represents the frequency boundary formed by extending from the center frequency towards higher frequencies, and the lower boundary represents the frequency boundary formed by extending from the center frequency towards lower frequencies. A reference frequency band is determined based on these upper and lower boundaries. The reference frequency band is a frequency range distributed around the center frequency, used to cover the adjacent frequency domain environment around the target order. The reference frequency band is divided into a low-frequency reference sub-band and a high-frequency reference sub-band. The low-frequency reference sub-band represents the reference frequency range located on the low-frequency side of the center frequency, and the high-frequency reference sub-band represents the reference frequency range located on the high-frequency side of the center frequency. The reference frequency band is thus divided into two... Subbands prevent background energy from being biased due to originating from only one frequency domain. Frequency intervals overlapping with the order-locked frequency band are removed from the low-frequency and high-frequency reference subbands. The order-locked frequency band represents the frequency range where the target order itself resides; removing overlapping frequency intervals prevents the target order energy from being repeatedly included in the background energy. Simultaneously, frequency intervals intersecting with the center frequency trajectories of other orders are removed from the low-frequency and high-frequency reference subbands to avoid including the energy of other orders in the background energy. The background energy value represents the sum of energy within the remaining frequency range after removal. For example, when adjacent orders have strong energy within a certain speed window, removing intersecting intervals ensures that the background energy only reflects environmental noise not dominated by a specific order. Finally, the background energy curve is determined based on the background energy values ​​for each speed window. The background energy curve is a continuous curve formed by the change of background energy with speed. By constructing a reference frequency band with the center frequency as a symmetrical reference, the background energy and the target order are placed in the same frequency domain environment, thus ensuring the reliability of the noise results.

[0035] In some embodiments of this application, when determining the noise result based on the order energy curve and the background energy curve, the following steps are included: aligning the order energy curve and the background energy curve with rotational speed coordinates, normalizing the order energy value and the background energy value at the same rotational speed position, marking sample points greater than zero as positive sample points, marking sample points less than or equal to zero as non-positive sample points, and determining the noise result curve based on the positive sample points.

[0036] Specifically, the order energy curve and the background energy curve are first aligned using rotational speed coordinates. This means mapping the two curves to the same rotational speed system, ensuring a one-to-one correspondence between the order energy value and the background energy value at any given rotational speed, thus avoiding distortion in energy comparison caused by differences in rotational speed window divisions. At the same rotational speed, the order energy value and the background energy value are normalized. They are calculated proportionally according to a unified energy benchmark, placing them within the same directly comparable dimension. This eliminates the influence of overall energy level differences across different rotational speed ranges on the judgment results. The normalization process does not change the relative magnitude relationship between the order energy value and the background energy value; rather, it ensures that the comparison process is unaffected by changes in absolute energy magnitude. The normalized results are then subjected to sign determination. Sample points with values ​​greater than zero are marked as positive sample points, while those less than or equal to zero are marked as non-positive sample points. Each sample point represents the calculation result corresponding to a discrete rotational speed position on the rotational speed axis. A positive sample point indicates that the order energy value at that rotational speed position is dominant relative to the background energy value, while a non-positive sample point indicates that the order energy value at that rotational speed position does not exceed the background energy value. For example, when the target order noise near a certain rotational speed position is significantly higher than the surrounding environmental noise, the sample point corresponding to that rotational speed position is marked as a positive sample point. Conversely, when the target order is covered by background noise or the two are close, the corresponding sample point is marked as a non-positive sample point. Based on the positive sample points, the noise result curve is determined, which can directly identify the continuous region on the rotational speed axis where the target order is dominant relative to the background energy, thus laying a data foundation for determining the noise evaluation value.

[0037] In some embodiments of this application, when identifying the interval where the noise result is higher than zero, the process includes: traversing the noise result curve along the rotational speed direction, taking the position where the positive sample point first appears as the candidate starting point and the position where the positive sample point ends as the candidate ending point, determining several candidate intervals, and eliminating candidate intervals where the rotational speed span is less than the minimum interval width threshold, and determining the effective interval based on the elimination results.

[0038] In some embodiments of this application, when obtaining the corresponding peak amplitude and interval width in each interval, the process includes: truncating each effective interval, determining the peak amplitude, determining the interval width based on the difference between the end speed and the starting speed of the effective interval, and constructing an interval parameter pair by the interval number using the peak amplitude and the interval width.

[0039] Specifically, the noise result curve is traversed along the rotational speed direction. The noise result curve is a continuous curve formed by positive sample points in sequence along the rotational speed. Traversing along the rotational speed direction involves checking the changing state point by point in order of increasing rotational speed. During the traversal, the position where the first positive sample point appears is taken as the candidate starting point. A positive sample point indicates that the order energy is dominant relative to the background energy at that rotational speed position. The candidate starting point represents the starting rotational speed position that may form an effective noise segment. When the continuous positive sample points end and turn into non-positive sample points, the position where the positive sample points end is taken as the candidate ending point. The candidate ending point represents the end position of this continuous dominant state on the rotational speed axis. Several candidate intervals are determined by pairing candidate starting points and candidate ending points. The candidate interval is a continuous range on the rotational speed axis defined by the candidate starting point and candidate ending point. Candidate intervals are screened, and those with a rotational speed span smaller than the minimum interval width threshold are removed. The rotational speed span is the distance between the candidate's endpoint and starting point on the rotational speed axis. The minimum interval width threshold is used to exclude the smallest continuous range of instantaneous fluctuations or isolated spikes. This removal process avoids mistaking transient unstable noise for valid intervals. Each valid interval is truncated by extracting the corresponding continuous segment from the noise result curve, thereby obtaining the corresponding peak amplitude and interval width. The peak amplitude represents the highest level of noise within the interval, and the interval width represents the coverage area of ​​the interval on the rotational speed axis. The interval width is determined by the difference between the endpoint rotational speed and the starting rotational speed of the valid interval. When the first non-positive sample switches to the position of a positive sample, the rotational speed corresponding to this position is the starting rotational speed. When a positive sample switches to a non-positive sample... If the position of the last positive sample is determined, then the rotational speed corresponding to the position of the last positive sample is the endpoint rotational speed. When constructing interval parameter pairs by interval numbering the peak amplitude and interval width, each valid interval is first assigned a unique interval number according to the order of appearance of the valid intervals on the rotational speed axis. The interval number is used to distinguish different valid intervals, so that each valid interval has a clear correspondence in the subsequent calculation process. The peak amplitude and interval width belonging to the same interval number are bound together, so that the peak amplitude and interval width form a one-to-one correspondence in the data structure. Thus, interval parameter pairs containing peak amplitude and interval width are obtained with interval number as index. By interval identification, discrete positive sample points are organized into continuous segments, and the influence of occasional interference is weakened by the interval width constraint, ensuring that the noise evaluation can simultaneously reflect the prominence of noise and the characteristics of continuous intervals.

[0040] In some embodiments of this application, when synthesizing the peak amplitude and interval width of all intervals to determine the noise evaluation value, and evaluating the noise quality of the electric drive system based on the noise evaluation value, the process includes: multiplying the peak amplitude and interval width corresponding to each effective interval to determine the interval contribution value; performing a weighted calculation on the contribution values ​​of all intervals to determine the noise evaluation value; comparing the noise evaluation value with the noise evaluation threshold; and determining an evaluation report based on the comparison result.

[0041] Specifically, for each valid interval, the peak amplitude and interval width corresponding to that interval are read. The valid interval is a continuous speed range retained after interval filtering. The peak amplitude and interval width corresponding to each valid interval are multiplied to determine the interval contribution value, which represents the individual contribution of the valid interval to the overall noise evaluation. This ensures that intervals with high peak values ​​but short durations and intervals with low peak values ​​but long durations have different impacts on the contribution results. Next, all interval contribution values ​​are weighted and processed to avoid any one interval from having an excessive impact on the overall evaluation result, thereby determining the noise evaluation value. The noise evaluation value is a single numerical result that summarizes the contribution information of multiple valid intervals. The noise evaluation value is compared with the noise evaluation threshold. When the noise evaluation value is greater than or equal to the noise evaluation threshold, it indicates that the noise quality of the electric drive system is poor. When the noise evaluation value is less than the noise evaluation threshold, it indicates that the noise quality of the electric drive system is good. An evaluation report is generated based on this, so that the evaluation results no longer depend on a single speed point, but cover the entire variable speed process, thereby achieving a unified quantitative evaluation of the noise quality of the electric drive system.

[0042] In summary, the beneficial effects of this invention are as follows: By collecting the in-vehicle sound pressure time-domain signal and the corresponding rotational speed signal, and resampling the sound pressure time-domain signal based on the rotational speed, the order noise can be continuously tracked in a unified coordinate system under variable rotational speed conditions. An order-locked frequency band is established around the center frequency to extract order energy, and background energy in the adjacent frequency domain environment of the target order is obtained through a reference frequency band. This establishes a correspondence between the target order and surrounding non-target components at the same rotational speed position. Furthermore, by using interval recognition, discrete instantaneous changes are transformed into continuous segments with rotational speed ranges, reflecting the continuous characteristics of noise. The peak amplitude and interval width of each segment are uniformly synthesized, transforming the complex variable rotational speed order change process into a unified evaluation index. This transforms noise evaluation from a local judgment relying on a single moment or a single rotational speed point to a holistic analysis covering the entire variable rotational speed process. The evaluation results have consistent calculation logic across different test conditions and different vehicles. Simultaneously, since the noise evaluation value originates from the comparison between order energy and background energy, the evaluation results can reflect the prominence of the target order in the actual operating environment, thereby ensuring the reliability of the noise quality evaluation of the electric drive system.

[0043] In another preferred embodiment based on the above embodiments, see [reference] Figure 2 As shown, this embodiment provides a noise quality evaluation system for a process-oriented electric drive system, used to apply a noise quality evaluation method for process-oriented electric drive systems, including: The acquisition unit is configured to acquire the in-vehicle sound pressure time-domain signal and the corresponding speed signal of the electric drive system during the variable speed process, and to resample the sound pressure time-domain signal based on the speed signal to determine the target order spectrum.

[0044] The analysis unit is configured to determine the center frequency of different rotational speeds based on the target order spectrum, establish an order-locked frequency band based on the center frequency, determine the order energy curve based on the order-locked frequency band, expand outward with the center frequency as a symmetrical reference to determine a reference frequency band, eliminate the frequency range of the order-locked frequency band within the reference frequency band, and determine the background energy curve based on the remaining frequency.

[0045] The processing unit is configured to determine the noise result based on the order energy curve and the background energy curve, and to search for the noise result along the rotation direction, identify the intervals where the noise result is higher than zero, and obtain the corresponding peak amplitude and interval width in each interval.

[0046] The evaluation unit is configured to synthesize the peak amplitude and interval width of all intervals, determine the noise evaluation value, and evaluate the noise quality of the electric drive system based on the noise evaluation value.

[0047] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for evaluating the noise quality of a process-oriented electric drive system, characterized in that, include: The in-vehicle sound pressure time-domain signal and the corresponding speed signal of the electric drive system during the variable speed process are collected, and the sound pressure time-domain signal is resampled based on the speed signal to determine the target order spectrum; The center frequencies of different rotational speeds are determined based on the target order spectrum, and an order-locked frequency band is established based on the center frequency. The order energy curve is determined according to the order-locked frequency band, and a reference frequency band is determined by expanding outward with the center frequency as a symmetrical reference. The frequency range of the order-locked frequency band is removed from the reference frequency band, and the background energy curve is determined according to the remaining frequency. The noise result is determined based on the order energy curve and the background energy curve, and the noise result is searched along the rotation direction to identify the intervals where the noise result is higher than zero, and the corresponding peak amplitude and interval width are obtained in each interval. The peak amplitude and interval width of all intervals are combined to determine the noise evaluation value, and the noise quality of the electric drive system is evaluated based on the noise evaluation value.

2. The noise quality evaluation method for a process-oriented electric drive system according to claim 1, characterized in that, When resampling the sound pressure time-domain signal to determine the target order spectrum, the following steps are included: Extract the instantaneous rotational speed increment of adjacent sampling points and determine the angular velocity. Based on the angular velocity, perform angular domain mapping on the sampling points of the sound pressure time domain signal, and determine the equiangular sampling position according to the angular step size. Interpolate the sound pressure time domain signal at the equiangular sampling position to determine the equiangular sound pressure sequence. The equal-angle sound pressure sequence is divided into sliding windows, and the data of each sliding window is transformed in the frequency domain to determine the order spectrum with the order as the horizontal axis and the rotation speed as the vertical axis. The order spectra of each sliding window are then spliced ​​together to determine the target order spectrum.

3. The noise quality evaluation method for a process-oriented electric drive system according to claim 2, characterized in that, Determining the center frequency of different rotational speeds based on the target order spectrum includes: In the target order spectrum, the spectral line position corresponding to the target order is selected as the initial search position, and the maximum point near the initial search position is searched segment by segment along the rotation direction to determine the peak order position corresponding to each rotation window. When the offset of adjacent peak positions exceeds the offset threshold, a secondary search range is set on both sides of the peak position of the previous rotation window and the peak position is repositioned. Based on the repositioned peak order position and the rotation speed of the rotation window, the center frequency sequence corresponding to the target order in the rotation window is determined, and the center frequency sequence is interpolated to complete the center frequency trajectory.

4. The noise quality evaluation method for a process-oriented electric drive system according to claim 3, characterized in that, Establishing the order-locked frequency band and determining the order energy curve includes: Within each rotation speed window, an upper and lower limit offset of the center frequency are set to determine the upper and lower limits of the order-locked frequency band. The spectral amplitude points between the upper and lower limits are extracted from the target order spectrum. Based on the amplitude points and the frequency step size, the order energy value of the rotation speed window is determined. The order energies of each rotation speed window are arranged to determine the order energy curve.

5. The noise quality evaluation method for a process-oriented electric drive system according to claim 4, characterized in that, When determining the reference frequency band and background energy curve by extending outward with the center frequency as a symmetrical reference, the process includes: Within each speed window, the frequency is extended symmetrically to the high-frequency and low-frequency sides using the center frequency as a reference to determine the upper and lower boundaries. A reference frequency band is determined based on the upper and lower boundaries, and the reference frequency band is divided into a low-frequency reference sub-band and a high-frequency reference sub-band. Frequency intervals that overlap with the order-locked frequency band and frequency intervals that intersect with the center frequency trajectories of other orders are removed from the low-frequency and high-frequency reference sub-bands. The background energy value of the speed window is determined based on the removal results, and the background energy curve is determined based on the background energy value of each speed window.

6. The noise quality evaluation method for a process-oriented electric drive system according to claim 5, characterized in that, When determining the noise result based on the order energy curve and the background energy curve, the following steps are included: The order energy curve and the background energy curve are aligned by rotational speed coordinates, and the order energy values ​​at the same rotational speed position are normalized with the background energy values. Sample points with values ​​greater than zero are marked as positive sample points, and sample points with values ​​less than or equal to zero are marked as non-positive sample points. The noise result curve is determined based on the positive sample points.

7. The noise quality evaluation method for a process-oriented electric drive system according to claim 6, characterized in that, When identifying the interval where the noise result is higher than zero, the following is included: Traverse the noise result curve along the rotational speed direction, take the first occurrence of the positive sample point as the candidate starting point and the end of the positive sample point as the candidate ending point, determine several candidate intervals, and remove candidate intervals whose rotational speed span is less than the minimum interval width threshold, and determine the effective intervals based on the removal results.

8. The noise quality evaluation method for a process-oriented electric drive system according to claim 7, characterized in that, When obtaining the corresponding peak amplitude and interval width within each interval, the following is included: Each valid interval is truncated to determine the peak amplitude, and the interval width is determined based on the difference between the starting and ending rotational speeds of the valid interval. The peak amplitude and interval width are then combined into interval parameter pairs according to the interval number.

9. The method for evaluating the noise quality of a process-oriented electric drive system according to claim 8, characterized in that, When synthesizing the peak amplitude and interval width of all intervals to determine the noise evaluation value, and evaluating the noise quality of the electric drive system based on the noise evaluation value, the process includes: The peak amplitude and interval width corresponding to each effective interval are multiplied to determine the interval contribution value. The contribution values ​​of all intervals are weighted and calculated to determine the noise evaluation value. The noise evaluation value is compared with the noise evaluation threshold, and the evaluation report is determined based on the comparison results.

10. A noise quality evaluation system for a process-oriented electric drive system, used to apply the noise quality evaluation method for a process-oriented electric drive system as described in any one of claims 1-9, characterized in that, include: The acquisition unit is configured to acquire the in-vehicle sound pressure time-domain signal and the corresponding speed signal of the electric drive system during the variable speed process, and to resample the sound pressure time-domain signal based on the speed signal to determine the target order spectrum; The analysis unit is configured to determine the center frequency of different rotational speeds based on the target order spectrum, establish an order-locked frequency band based on the center frequency, determine the order energy curve according to the order-locked frequency band, expand outward with the center frequency as a symmetrical reference to determine a reference frequency band, remove the frequency range of the order-locked frequency band within the reference frequency band, and determine the background energy curve based on the remaining frequency. The processing unit is configured to determine the noise result based on the order energy curve and the background energy curve, and to search the noise result along the rotation direction to identify the intervals where the noise result is higher than zero, and to obtain the corresponding peak amplitude and interval width in each interval. The evaluation unit is configured to synthesize the peak amplitude and interval width of all intervals, determine the noise evaluation value, and evaluate the noise quality of the electric drive system based on the noise evaluation value.

Images