A method and system for diagnosing the operating status of a constant humidity system

Abstract

This invention belongs to the field of data analysis technology, specifically relating to a method and system for diagnosing the operational status of a constant humidity system. The method includes: synchronously acquiring the humidity status of the cab and the engine speed time series; performing differential and standardization processing on the humidity data to construct three core diagnostic indicators: dynamic trajectory entropy, three-dimensional trajectory cyclicity, and resonance decoupling index. The dynamic trajectory entropy is used to initially screen for abnormal states with high uncertainty; the three-dimensional trajectory cyclicity is used to distinguish between periodic oscillations and random noise; the resonance decoupling index uses engine speed as an external reference source and, through spectrum analysis, accurately determines whether the source of periodic fluctuations is a true fault caused by internal control oscillations or a false fault caused by resonance interference from external sources such as the engine. Finally, through layer-by-layer logical judgment of these three indicators, accurate diagnosis of the constant humidity system status is achieved, effectively avoiding misjudgments.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology. More specifically, this invention relates to a method and system for diagnosing the operating status of a constant humidity system. Background Technology

[0002] As an important means of production, the cab of a heavy-duty truck is the core space where the driver works for long periods of time. In order to ensure driving comfort and driving safety, modern heavy-duty truck air conditioning systems generally integrate constant humidity function. By precisely controlling the air humidity, driver fatigue and window fogging are prevented. Therefore, the reliable operation of the constant humidity system is the key to realizing these functions. Thus, real-time and accurate status diagnosis of the constant humidity system is of great significance.

[0003] However, the working environment of heavy trucks is extremely harsh. Long-term strong vibration, bumps and complex electromagnetic interference pose a severe test to the sensors and actuators of the constant humidity system.

[0004] To diagnose the operating status of a constant humidity system, existing technologies often employ time series analysis to process humidity data. Using the Symbolic Aggregate Approximation (SAX) algorithm and its preceding Piecewise Aggregate Approximation (PAA) algorithm, high-dimensional raw time series data can be reduced in dimensionality and converted into symbolic strings. Its information entropy is then calculated to determine the degree of signal disorder. False faults in sensors caused by external interference such as vibration will make the signal random and chaotic, leading to an abnormally high information entropy. Conversely, true faults caused by the performance degradation of system components will make the signal monotonous and sluggish, leading to an abnormally low information entropy.

[0005] However, the aforementioned technical methods pose a risk of misjudging genuine faults in oscillating systems: when the control unit of the constant humidity system experiences a logic error or abnormal actuator response, the system may enter an unstable high-frequency oscillation state, with humidity fluctuating drastically around the target value. The signal generated by this genuine fault, due to its rapid and irregular changes, will also exhibit extremely high information entropy after SAX conversion. At the same time, under specific operating conditions of heavy trucks, such as at specific engine speeds, the false faults corresponding to the periodic resonance transmitted by the chassis and engine will also lead to high entropy in the humidity signal.

[0006] Existing technologies cannot distinguish between different high-entropy states, making it difficult for traditional diagnostic methods to effectively differentiate between true and false faults. This results in a high false alarm rate and a high false alarm rate, increasing unnecessary maintenance costs and downtime. Summary of the Invention

[0007] To address the technical problem that existing technologies cannot distinguish between different high-entropy states, which makes it difficult for traditional diagnostic methods to effectively differentiate between true and false faults, resulting in high false alarm and false alarm rates, this invention provides solutions in the following aspects.

[0008] In a first aspect, the present invention provides a method for diagnosing the operational status of a constant humidity system, comprising: acquiring a humidity state time series and an engine speed time series in the cab; performing differential processing and standardization processing on the humidity state time series to obtain a standardized state sequence, a standardized velocity sequence, and a standardized acceleration sequence; discretizing the standardized state sequence, standardized velocity sequence, and standardized acceleration sequence into a state symbol string, a velocity symbol string, and an acceleration symbol string; calculating the information entropy of the state symbol string as a dynamic trajectory entropy; combining the state symbols, velocity symbols, and acceleration symbols at the same moment to form a three-dimensional symbol trajectory sequence, calculating the three-dimensional trajectory cyclicity, wherein the three-dimensional trajectory cyclicity is inversely proportional to the number of non-repeating subsequence patterns in the three-dimensional symbol trajectory sequence and directly proportional to the total number of subsequences; performing spectral analysis on the engine speed time series to obtain the dominant frequency with the highest energy in its power spectral density; calculating the spectral coherence coefficient value of the humidity state and engine speed corresponding to the dominant frequency and defining it as a resonance decoupling index; and determining the fault status of the constant humidity system based on the dynamic trajectory entropy, the three-dimensional trajectory cyclicity, and the resonance decoupling index.

[0009] This invention establishes a multi-dimensional, progressively deeper diagnostic framework by constructing three dimensions of diagnostic indicators: dynamic trajectory entropy, three-dimensional trajectory cyclicity, and resonance decoupling index. It can not only distinguish between random noise and periodic fluctuations in the system from the time domain and phase space, but also introduce engine speed as an external reference source. Through spectral coherence analysis, it accurately decouples the fault source and can clearly distinguish between true faults caused by internal control oscillations and false faults caused by resonance interference from external vibrations such as the engine. This makes the diagnostic results both highly robust and highly accurate, reducing maintenance costs.

[0010] Preferably, the differential and normalization processing of the humidity state time series includes: constructing a velocity series through first-order differencing, wherein time... speed Acceleration sequences are constructed using second-order differences, where time intervals are given. acceleration ; , , At time respectively , , The humidity state data were obtained; the humidity state time series, velocity series and acceleration series were Z-Score standardized to obtain standardized state series, standardized velocity series and standardized acceleration series.

[0011] Preferably, after performing differential processing on the humidity state time series, a common time window is extracted from the humidity state time series, velocity series, acceleration series, and engine speed time series. , To preset the length of the sliding time window, four dynamic sequences of equal length and with aligned time indices are generated: state processing sequence. Speed ​​processing sequence Acceleration processing sequence and rotation speed processing sequence .

[0012] Preferably, the standardized state sequence, standardized velocity sequence, and standardized acceleration sequence are discretized into state symbol strings, velocity symbol strings, and acceleration symbol strings, including: using a piecewise aggregation approximation method to reduce the dimensionality of the standardized state sequence, standardized velocity sequence, and standardized acceleration sequence to obtain three PAA sequences; using a symbol aggregation approximation method, based on a preset inclusion... Using the alphabet of symbols and the split points based on the standard normal distribution, the three PAA sequences are converted into state symbol strings, velocity symbol strings, and acceleration symbol strings, respectively.

[0013] This invention converts continuous time series data into discrete symbol strings, which can effectively reduce data dimensionality and sensitivity to noise while preserving the main morphological features of the sequence, providing a data foundation for subsequent symbol sequence-based calculations such as information entropy and trajectory cyclicity.

[0014] Preferably, the formula for calculating the information entropy of the state symbol string is: In the formula, For dynamic trajectory entropy, The first in the alphabet One symbol; For symbols In the status string The probability of it appearing in The size of the alphabet.

[0015] This invention quantifies the disorder of humidity state sequences by using information entropy, providing a first-level screening for fault diagnosis. This indicator can quickly determine from a statistical perspective whether the system is in an abnormal fluctuation, i.e., a high-entropy state, thereby efficiently filtering out normal or slowly changing states, allowing diagnostic resources to be concentrated on signal segments with potential problems.

[0016] Preferably, the formula for calculating the three-dimensional trajectory cyclicity is: In the formula, For the three-dimensional trajectory cyclicity, In the three-dimensional symbol trajectory sequence The total number of subsequences extracted; It represents the number of patterns of non-repeating subsequences among all extracted subsequences.

[0017] This invention effectively distinguishes between random noise (low cyclicity) and deterministic periodic oscillations (high cyclicity) by analyzing the repeatability of patterns in three-dimensional symbol trajectories. Pattern analysis in three-dimensional phase space has a stronger ability to suppress noise and a stronger robustness in distinguishing periodic behavior, laying the foundation for subsequent fault type determination.

[0018] Preferably, the formula for calculating the spectral coherence coefficient between the humidity state and engine speed at the dominant frequency is: ;in, In order to dominate frequency The corresponding humidity-speed spectrum coherence coefficient; In order to dominate frequency The corresponding humidity-speed cross-power spectral density at that location, In order to dominate frequency The corresponding humidity autopower spectral density at that location, In order to dominate frequency The corresponding engine speed and power spectral density.

[0019] This invention attributes the source of periodic fluctuations by calculating the spectral coherence coefficient value, quantifies the correlation strength between humidity fluctuations and engine speed fluctuations at the engine's dominant frequency, and thus provides a basis for judging the final distinction between internal control oscillations and external resonance interference.

[0020] Preferably, the humidity-speed cross-power spectral density, humidity auto-power spectral density, and engine speed auto-power spectral density are obtained by processing the state sequence using the Welch method. and rotation speed processing sequence Obtained through frequency domain correlation analysis.

[0021] Preferably, determining the fault state of the constant humidity system based on the dynamic trajectory entropy, three-dimensional trajectory cyclicity, and resonance decoupling index includes: if the dynamic trajectory entropy The system is then assessed to determine if the humidity control system is in normal or experiencing slow, low-frequency changes, and the diagnostic process is terminated. For high entropy threshold; if And the three-dimensional trajectory cyclicity The humidity control system was determined to be in a false fault state. The preset cycle count threshold; if and To decouple the resonance index With resonance threshold Comparison: If If the humidity control system is determined to be in a false fault state; The humidity control system was determined to be in a true fault state.

[0022] Secondly, the present invention provides an operational status diagnostic system for a constant humidity system, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned operational status diagnostic method for a constant humidity system is implemented.

[0023] By adopting the above technical solution, a computer program is generated from the above-mentioned method for diagnosing the operating status of a constant humidity system and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and the processor for convenient use.

[0024] The beneficial effects of this invention are as follows:

[0025] This invention establishes a multi-dimensional, progressively deeper diagnostic framework by constructing three dimensions of diagnostic indicators: dynamic trajectory entropy, three-dimensional trajectory cyclicity, and resonance decoupling index. It can not only distinguish between random noise and periodic fluctuations in the system from the time domain and phase space, but also introduce engine speed as an external reference source. Through spectral coherence analysis, it accurately decouples the fault source and can clearly distinguish between true faults caused by internal control oscillations and false faults caused by resonance interference from external vibrations such as the engine. This makes the diagnostic results both highly robust and highly accurate, reducing maintenance costs. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a method for diagnosing the operating status of a constant humidity system according to the present invention;

[0027] Figure 2 It is a schematic diagram illustrating the time series of humidity and engine speed under normal operating conditions;

[0028] Figure 3 This is a schematic diagram illustrating the time series of humidity and engine speed under a false fault caused by random vibration;

[0029] Figure 4 This is a schematic diagram illustrating the time series of humidity and engine speed under a false fault caused by external resonance;

[0030] Figure 5 This is a schematic diagram illustrating the time series of humidity and engine speed under a true fault caused by internal control oscillation;

[0031] Figure 6It is an illustrative representation of the... Figure 4 A schematic diagram of the power spectral density obtained after frequency domain decoupling;

[0032] Figure 7 It is shown schematically. Figure 6 A schematic diagram of the humidity-speed spectrum coherence coefficients at different frequencies.

[0033] Figure 8 It is an illustrative representation of the... Figure 5 A schematic diagram of the power spectral density obtained after frequency domain decoupling;

[0034] Figure 9 It is shown schematically. Figure 8 A schematic diagram of the humidity-speed spectrum coherence coefficients at different frequencies. Detailed Implementation

[0035] 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, not all, of the embodiments of the present invention. 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.

[0036] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] This invention discloses a method for diagnosing the operating status of a constant humidity system, referring to... Figure 1 This includes steps S1-S6:

[0038] S1: Obtain the time series of humidity status in the cab and the engine speed time series; perform differential processing and standardization on the humidity status time series to obtain the standardized state series, standardized velocity series and standardized acceleration series.

[0039] Specifically, humidity data of the air conditioning system in the cab is acquired at various times via a humidity sensor, and engine speed is acquired at various times via the vehicle's CAN bus, with a sampling frequency of 10Hz. This results in a humidity state time series and an engine speed time series within a preset sliding time window, the length of which is defined as... The time index is .

[0040] Among them, the length of the time window To meet the frequency resolution requirements of the spectrum estimation algorithm in subsequent steps, the length It cannot be too small; in order to meet the real-time requirements of the constant humidity system for fault diagnosis, the length... It also can't be too large, so choosing a 1-minute diagnostic window results in a short duration. It equals the product of the sampling frequency and the time length, i.e., 60 seconds, which is the length. Data points.

[0041] Furthermore, dynamic change information of the constant humidity system is extracted from single humidity state data, and then velocity characteristics reflecting the motion trend of the constant humidity system and acceleration characteristics reflecting the motion change trend are constructed. Specifically, a velocity sequence is constructed by approximating the velocity using first-order difference, where at time... speed The sequence is in The above is defined; by approximating the acceleration using the second-order difference, an acceleration sequence is constructed, where, at time... acceleration The sequence is in There is a definition above; , , At time respectively , , Humidity status data.

[0042] Simultaneously, to construct a three-dimensional phase space with consistent time indices, it is necessary to extract a common time window that is defined for all three sequences, i.e. The effective length of this window is Specifically, generate three [units] of length [missing information]. Furthermore, dynamic sequences aligned to time indices: state processing sequences Speed ​​processing sequence and acceleration processing sequence .

[0043] Furthermore, the physical dimensions and numerical ranges of these three sequences are different, requiring standardization to eliminate the influence of dimensions and make them comparable. This is a prerequisite for subsequent unified symbolic mapping; therefore, for , and The three sequences were Z-score standardized to obtain standardized sequences with a mean of 0 and a standard deviation of 1. , and ,in , For a standardized state sequence, For the standardized velocity sequence, This is a standardized acceleration sequence.

[0044] It should be noted that three sequences of equal length, time alignment, and uniform dimensions were obtained. , and Together they describe the constant humidity system in Complete dynamic behavior within a time period.

[0045] Finally, the engine speed time series is also truncated to obtain the speed processing sequence. This makes it consistent with the state processing sequence. They are of equal length and time-aligned.

[0046] S2: Discretize the standardized state sequence, standardized velocity sequence, and standardized acceleration sequence into state symbol strings, velocity symbol strings, and acceleration symbol strings, and calculate the information entropy of the state symbol strings as the dynamic trajectory entropy.

[0047] It should be noted that in order to quantify the uncertainty and disorder of the humidity signal in the time domain, it is first necessary to reduce the dimension and discretize the one-dimensional state sequence and convert it into a symbol string.

[0048] Specifically, the segmented aggregation approximation method is first applied to... Length The sequence is compressed into a length of PAA sequence Then, the PAA sequence was processed. Perform symbolic aggregation approximation processing, specifically: based on a preset inclusion... The PAA sequence is constructed using an alphabet of symbols and split points based on a standard normal distribution. Convert to a symbol string and record it as a status string. .

[0049] The segmented aggregation approximation method is a sequence dimensionality reduction technique, which involves dividing a sequence into segments of length 10 ... The sequence is divided into 1. Frames of equal length are used, and the mean of the data within each frame is calculated to obtain a length of . The mean sequence is obtained; this method significantly reduces the data dimensionality while preserving the main shape of the sequence.

[0050] The symbolic aggregation approximation method is a time series discretization technique based on the assumption that the numerical distribution of the standardized time series approximates a standard normal distribution. This is achieved by setting a value on the standard normal distribution. Divide the area under the normal curve into equal parts at each dividing point. There are 3 regions, each corresponding to a symbol; when a mean in the PAA sequence falls into a certain region, it is assigned the corresponding symbol.

[0051] Optional, parameters The length of the PAA sequence, its value range is: In this embodiment, it is set to ;parameter The size of the alphabet, its value range is: In this embodiment, it is set to Then the symbols in the alphabet can be a, b, c, and d.

[0052] Furthermore, based on the status string Calculate its information entropy, which is defined as dynamic trajectory entropy. Used to quantize state strings The degree of disorder is used as a preliminary screening indicator for fault diagnosis. The specific calculation formula is as follows:

[0053]

[0054] In the formula, For dynamic trajectory entropy, The first in the alphabet One symbol; For symbols In the status string The probability of it appearing in, i.e. Chinese symbol The number of occurrences divided by the status string Total length , The size of the alphabet.

[0055] Among them, when When the value increases, it indicates that the randomness and uncertainty of the humidity signal increases, and the constant humidity system may be in an abnormal state caused by random noise interference or control oscillation.

[0056] Furthermore, only high-entropy states pose a risk of noise and oscillation being mixed together; low-entropy states corresponding to stable or slowly drifting signals do not need to be considered. Therefore, a high-entropy threshold is set. When the calculated dynamic trajectory entropy If the condition is met, the deep diagnostic step S3 is triggered; otherwise, it is determined that the constant humidity system has not experienced a high entropy anomaly. Through the preliminary screening of the constant humidity system's state, computational resources are concentrated on processing suspicious signals with high uncertainty.

[0057] Among them, due to The range of values ​​is Therefore, the high entropy threshold The range of values ​​is In this embodiment, the high entropy threshold is... Set as .

[0058] S3: Combine the state symbol, velocity symbol, and acceleration symbol at the same moment to form a three-dimensional symbol trajectory sequence.

[0059] It should be noted that after the constant humidity system is determined to be in a high-entropy state by step S2, in order to distinguish whether its internal driving force is random noise or deterministic oscillation, the analysis dimension needs to be increased from one-dimensional time domain to three-dimensional space.

[0060] Specifically, the standardized velocity sequence obtained in step S1 is used. and standardized acceleration sequences Using standardized state sequences The PAA algorithm and SAX processing parameters are exactly the same, i.e., the length of the PAA sequence is... The size of the alphabet is Standardized velocity sequence and standardized acceleration sequences Convert them into symbol strings respectively, and denote them as speed symbol strings. and acceleration symbol string .

[0061] Furthermore, index the same time period. The state symbols, velocity symbols, and acceleration symbols are paired to form a three-dimensional symbol trajectory sequence. In the formula, It is the index after PAA segmentation. ; , and Representing the first The state, velocity, and acceleration sign for each time period; Each element in Each is a three-dimensional symbolic tuple, representing the three-dimensional phase space coordinates of the constant humidity system at that moment. This allows us to obtain the state information of the constant humidity system at any given time. Speed ​​information and acceleration information By combining these elements, we can fully depict its trajectory in three-dimensional phase space, which is the key premise for distinguishing between randomness and determinism.

[0062] S4: Calculate the cyclicity of the three-dimensional trajectory.

[0063] It should be noted that the trajectory of random noise, such as road bumps, in three-dimensional phase space is chaotic and non-repeating; while whether it is internal controlled oscillation or external resonant interference, its trajectory will appear as a low-dimensional manifold, such as a closed loop, resulting in a high degree of repetition of its trajectory pattern. Therefore, random noise and periodic fluctuations can be distinguished by quantifying the repeatability of the three-dimensional symbol trajectory sequence.

[0064] Specifically, from the three-dimensional symbol trajectory sequence Extract a subsequence of a specific length, where the length of the subsequence is... It is a hyperparameter, and its value is usually set to 1. Then each subsequence consists of three consecutive three-dimensional symbol tuples, for example... .

[0065] Furthermore, statistical three-dimensional symbol trajectory sequences Total number of neutron sequences The number of patterns of non-repeating subsequences This leads to the construction of a three-dimensional trajectory cyclicity. This is used to measure the periodicity and determinism of the three-dimensional dynamic trajectory of a constant humidity system. The specific calculation formula is as follows:

[0066]

[0067] In the formula, For the three-dimensional trajectory cyclicity, In the three-dimensional symbol trajectory sequence The total number of subsequences extracted; It represents the number of patterns of non-repeating subsequences among all extracted subsequences.

[0068] When the phase space trajectory is a random walk, the resulting subsequence patterns, corresponding to random noise, will be diverse. It will approach This leads to the cyclicity of the three-dimensional trajectory. The value approaches 0; conversely, when the trajectory falls into a periodic cycle, corresponding to control oscillation or resonance disturbance, the vast majority of subsequences are repetitions of a few patterns. It will be much smaller This leads to the cyclicity of the three-dimensional trajectory. The value increases significantly, approaching 1.

[0069] It should be noted that random trajectories in three-dimensional analysis are less likely to repeat; therefore, the obtained three-dimensional trajectory cyclicity... It has stronger noise suppression while maintaining the repeatability of deterministic trajectories, thus making the three-dimensional trajectory cyclicity feature more discriminative.

[0070] S5: Perform spectral analysis on the engine speed time series to obtain the dominant frequency with the highest energy in its power spectral density; calculate the spectral coherence coefficient value between the humidity state and the engine speed at the dominant frequency, and define it as the resonance decoupling index.

[0071] It should be noted that the features constructed in step S4 have separated out random noise, but left a vague state of periodic fluctuations. This state may be control oscillations caused by real faults or resonance interference caused by false faults. In order to distinguish between the two, external vibration source data specific to the heavy truck scenario is introduced.

[0072] To analyze the correlation between two sequences in the frequency domain, we first need to use the Welch method to process the aligned state sequences. and rotation speed processing sequence Frequency domain correlation analysis was performed by segmenting the long sequence, windowing it, calculating the Fourier transform separately, and then averaging the results to obtain a smoother power spectral density estimate with smaller variance. This estimate includes three key power spectral density functions:

[0073] (1) Humidity autopower spectral density This indicates that the energy of the humidity signal itself is in the frequency range. The corresponding distribution at that location.

[0074] (2) Engine speed power spectral density This indicates the rotational speed signal. Its own energy at frequency The corresponding distribution at that location.

[0075] (3) Humidity-rotation speed cross power spectral density This indicates the frequency difference between the humidity and rotation speed signals. The correlation on the surface includes both amplitude and phase information.

[0076] Furthermore, based on the estimated three power spectral density functions, the humidity-speed spectral coherence coefficients at different frequencies are calculated. To identify the degree of linear correlation between the humidity and engine speed signals at different frequencies, the specific calculation formula is as follows:

[0077]

[0078] in, In frequency The corresponding humidity-speed spectrum coherence coefficient at that location has a range of values. , An equal value of 1 indicates that at a frequency The two signals are completely linearly correlated, meaning that the fluctuation in the humidity signal is entirely caused by the engine. A value of 0 indicates that the two are completely unrelated; In frequency The corresponding humidity-speed cross-power spectral density at that location, In frequency The corresponding humidity autopower spectral density at that location, In frequency The corresponding engine speed and power spectral density.

[0079] Finally, since the process of decoupling the source of periodic fluctuations focuses on the overall contribution of engine speed to the humidity signal, and this contribution is concentrated in the dominant frequency of the engine, the dominant frequency with the highest energy is found in the obtained engine speed self-power spectral density. The spectral coherence coefficient value at the dominant frequency is calculated as a resonant decoupling index. ,Right now .

[0080] Among them, when the resonant decoupling index When the value approaches 1, it indicates that the periodic fluctuation energy in the humidity signal is highly correlated with the dominant frequency of engine speed, and the fluctuation is most likely caused by engine resonance; when the resonance decoupling index When the value approaches 0, it indicates that the periodicity of the humidity signal is unrelated to the engine and is generated by the self-excitation of the constant humidity system.

[0081] It should be noted that utilizing the scenario information of the heavy-duty truck CAN bus to achieve precise decoupling of the source of periodic fluctuations is the key to solving the problem of misjudgment.

[0082] S6: Determine the fault status of the constant humidity system based on dynamic trajectory entropy, three-dimensional trajectory cyclicity and resonance decoupling index.

[0083] Specifically, the comprehensive dynamic trajectory entropy 3D trajectory cyclicity Harmonic decoupling index Execute the following three-layer diagnostic logic:

[0084] First layer: If the dynamic trajectory entropy If the humidity system is determined to be in normal or slowly changing at low frequency, the diagnosis is complete. This is the high entropy threshold.

[0085] Second layer: If the dynamic trajectory entropy And the three-dimensional trajectory cyclicity If the constant humidity system is found to be in a false fault state, it is because it is affected by random vibrations such as road bumps.

[0086] Third layer: If the dynamic trajectory entropy And the three-dimensional trajectory cyclicity At this point, the constant humidity system is in a state of high entropy and high structure, which in turn affects the resonance decoupling index. With resonance threshold Comparison:

[0087] (1) If the resonance decoupling index If the constant humidity system is found to be in a false fault state, it is because it is affected by external resonance interference such as engine resonance.

[0088] (2) If the resonance decoupling index If the constant humidity system is found to be in a state of true fault, the cause is internal control oscillation due to ECU or actuator failure or other reasons.

[0089] in, The preset cycle degree threshold has a value range of [value range missing]. In this embodiment, it is set to ; This is the resonance correlation threshold, and its value range is... In this embodiment, it is set to This means that the humidity signal is only attributed to resonance when the correlation between the humidity signal and the engine speed exceeds 80%.

[0090] For example, Figures 2 to 5 The original signals under four operating conditions are shown, and fault condition diagnosis is performed for these four operating conditions, specifically as follows:

[0091] (1) Targeting Figure 2 The original signal under the corresponding operating condition is calculated to have a dynamic trajectory entropy of 1.1999, which is used for the first level of diagnosis. Since its dynamic trajectory entropy is less than the high entropy threshold, this is not considered a high-entropy signal. Therefore, the humidity control system is judged to be in a normal state or undergoing slow, low-frequency changes. Figure 2 This is a schematic diagram of the time series of humidity and engine speed under normal operating conditions.

[0092] (2) Regarding Figure 3 The original signal under the corresponding operating condition is calculated to have a dynamic trajectory entropy of 1.571, which is used for the first level of diagnosis. Since its dynamic trajectory entropy is greater than the high entropy threshold... This triggers the second-level diagnosis, further calculating its three-dimensional trajectory cyclicity to be 0.3926. Since its three-dimensional trajectory cyclicity is less than the preset cyclicity threshold, this is incorrect. The humidity control system was determined to be in a false fault state because it was affected by random vibrations such as road bumps. Figure 3 This is a schematic diagram of the time series of humidity and engine speed under a false fault caused by random vibration.

[0093] (3) Targeting Figure 4 The original signal under the corresponding operating condition is calculated to have a dynamic trajectory entropy of 1.3092, which is used for the first level of diagnosis. Since its dynamic trajectory entropy is greater than the high entropy threshold... This triggers the second-level diagnosis, further calculating its three-dimensional trajectory cyclicity to be 0.5293. Since its three-dimensional trajectory cyclicity is greater than the preset cyclicity threshold... This triggers the third layer of diagnosis, and through frequency domain decoupling, the power spectral density diagram is shown below. Figure 6 As shown, through Figure 6 The dominant frequency was determined to be 1.48 Hz, and then the humidity-speed spectrum coherence coefficients at different frequencies were calculated, as shown in the schematic diagram. Figure 7 As shown, Figure 7 The spectral coherence coefficient value corresponding to the dominant frequency is used as a resonance decoupling index. Therefore, its resonance decoupling index is equal to 0.97, which is greater than the resonance correlation threshold. The humidity control system was determined to be in a false fault state because it was affected by external resonance interference such as engine resonance. Figure 4 This is a schematic diagram of the time series of humidity and engine speed under a false fault caused by external resonance.

[0094] (4) Targeting Figure 5 The original signal under the corresponding operating condition was calculated to have a dynamic trajectory entropy of 1.8938, and the first level of diagnosis was performed. However, since its dynamic trajectory entropy was greater than the high entropy threshold... This triggers the second-level diagnosis, further calculating its three-dimensional trajectory cyclicity to be 0.5161. Since its three-dimensional trajectory cyclicity is greater than the preset cyclicity threshold... This triggers the third layer of diagnosis, and through frequency domain decoupling, the power spectral density diagram is shown below. Figure 8 As shown, through Figure 8 The dominant frequency was determined to be 1.48 Hz, and then the humidity-speed spectrum coherence coefficients at different frequencies were calculated, as shown in the schematic diagram. Figure 9 As shown, Figure 9 The spectral coherence coefficient value corresponding to the dominant frequency is used as a resonance decoupling index. Therefore, its resonance decoupling index is equal to 0.1, which is less than the resonance correlation threshold. The humidity control system was determined to be in a true fault state, caused by internal control oscillations due to ECU or actuator malfunctions. Therefore... Figure 5 This is a schematic diagram of the time series of humidity and engine speed under a true fault caused by internal control oscillation.

[0095] In summary, the method of this invention can correctly distinguish between low-entropy normal states and high-entropy abnormal states. In high-entropy states, the method can distinguish between random fluctuations and periodic fluctuations by using three-dimensional trajectory cyclicity. In periodic fluctuation states, the method can successfully distinguish between false faults caused by external resonance and true faults caused by internal control oscillations by introducing engine speed and performing frequency domain decoupling. Therefore, the diagnostic logic of this invention covers all high-entropy states and solves the key problem of periodic fluctuation attribution by using scene data engine speed, thus achieving accurate diagnosis.

[0096] This invention also discloses an operational status diagnostic system for a constant humidity system, comprising a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an operational status diagnostic method for a constant humidity system according to the present invention.

[0097] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A method of diagnosing an operating state of a constant humidity system, characterized by, The method comprises the following steps: acquiring a humidity state time sequence and an engine speed time sequence of a driver's cabin; performing differential processing and standardization processing on the humidity state time sequence to obtain a standardized state sequence, a standardized speed sequence, and a standardized acceleration sequence; discretizing the standardized state sequence, the standardized speed sequence, and the standardized acceleration sequence into a state symbol string, a speed symbol string, and an acceleration symbol string; calculating the information entropy of the state symbol string as a dynamic trajectory entropy; combining the state symbol, the speed symbol, and the acceleration symbol at the same time to form a three-dimensional symbol trajectory sequence, and calculating a three-dimensional trajectory cyclic degree, which is inversely proportional to the number of non-repeated subsequence modes in the three-dimensional symbol trajectory sequence and is proportional to the total number of subsequence; performing spectrum analysis on the engine speed time sequence to obtain a dominant frequency with the highest energy in a power spectrum density of the engine speed time sequence; calculating a spectral coherence coefficient value of the humidity state and the engine speed at the dominant frequency, and defining the spectral coherence coefficient value as a resonance decoupling index; determining the fault state of the constant humidity system based on the dynamic trajectory entropy, the three-dimensional trajectory cyclic degree and the resonance decoupling index, including: if the dynamic trajectory entropy , determining that the constant humidity system is in a normal or low-frequency slowly varying state, and ending the diagnosis, wherein, is a high entropy threshold; if and the three-dimensional trajectory cyclic degree , determining that the constant humidity system is in a false fault state, wherein, is a preset cyclic degree threshold; if and , comparing the resonance decoupling index with a resonance threshold : if , determining that the constant humidity system is in a false fault state; if , determining that the constant humidity system is in a true fault state.

2. The operating state diagnosis method for a constant humidity system according to claim 1, wherein the differential processing and the standardization processing on the humidity state time sequence comprise the following steps: The velocity sequence is constructed by first-order difference, wherein the velocity at time point The acceleration sequence is constructed by second-order difference, wherein the acceleration at time point , , are humidity state data at time points , , , respectively.​​​ performing Z-Score standardization processing on the humidity state time sequence, the speed sequence, and the acceleration sequence to obtain the standardized state sequence, the standardized speed sequence, and the standardized acceleration sequence.

3. The operating state diagnosis method for a constant humidity system according to claim 2, wherein The humidity state time sequence is subjected to differential processing, and a common time window of the humidity state time sequence, the speed sequence, the acceleration sequence and the engine speed time sequence is intercepted , The length of the preset sliding time window is generated. Four dynamic sequences with the same length and time index alignment are generated: state processing sequence , speed processing sequence , acceleration processing sequence and speed processing sequence .

4. The operating state diagnosis method for a constant humidity system according to claim 1, wherein discretizing the standardized state sequence, the standardized speed sequence, and the standardized acceleration sequence into the state symbol string, the speed symbol string, and the acceleration symbol string comprises the following steps: The standardized state sequence, the standardized speed sequence and the standardized acceleration sequence are reduced in dimension using a piecewise aggregate approximation method to obtain three PAA sequences; the three PAA sequences are respectively converted into a state symbol string, a speed symbol string and an acceleration symbol string using a symbolic aggregate approximation method according to a preset alphabet containing symbols and a split point based on a standard normal distribution.

5. The operating state diagnosis method for a constant humidity system according to claim 4, wherein the calculation formula of the information entropy of the state symbol string is: ； where is the dynamic trajectory entropy, is the i-th symbol in the alphabet; is the i-th symbol in the alphabet; is the i-th symbol in the alphabet; is the probability of the state string is the probability of the state string is the size of the alphabet.

6. The operating state diagnosis method for a constant humidity system according to claim 1, wherein the calculation formula of the three-dimensional trajectory cyclic degree is: ； wherein is the three-dimensional trajectory cycle degree, is the total number of sub-sequences extracted in the three-dimensional symbolic trajectory sequence is the total number of sub-sequences extracted in the three-dimensional symbolic trajectory sequence is the number of pattern species of non-repeated sub-sequences among all extracted sub-sequences.

7. The operating state diagnosis method for a constant humidity system according to claim 1, wherein the calculation formula of the spectral coherence coefficient value of the humidity state and the engine speed at the dominant frequency is: ； wherein, is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency is the corresponding humidity-rotation speed cross power spectral density at the dominant frequency 8. The operating state diagnosis method for a constant humidity system according to claim 3 or 7, characterized in that, The humidity-rotation speed cross power spectral density, humidity self power spectral density and engine rotation speed self power spectral density are obtained by Welch method for state processing sequence and rotation speed processing sequence frequency domain correlation analysis.

9. An operating state diagnosis system for a constant humidity system, characterized by comprising: The method comprises the following steps: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for diagnosing the running state of a constant humidity system according to any one of claims 1-8 is realized.

Images