Air conditioning unit fault type diagnosis method and system

Abstract

The invention relates to the technical field of air conditioning unit diagnosis, in particular to an air conditioning unit fault type diagnosis method and system, and the method comprises the steps: carrying out the state monitoring of a target air conditioning unit according to a preset sampling period based on a sensor group, so as to obtain a state data sequence of the target air conditioning unit, and carrying out the data fitting of the state data sequence, a state fitting curve is obtained; and matching the state fitting curve with a preset standard state fitting curve to obtain a matching degree of the state fitting curve and the preset standard state fitting curve, and comparing the matching degree with a preset matching degree threshold. According to the method, the state of the air conditioning unit is monitored through the sensor group, abnormal conditions can be found in time, early diagnosis and processing are facilitated, the equipment fault risk is reduced, the running state of the equipment can be recognized more accurately by combining fitting of the state data sequence with matching of the standard curve, and therefore the accuracy of fault detection is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning unit diagnosis, and in particular to an air conditioning unit fault type diagnosis method and system. Background Art

[0002] Air conditioning units are devices used to regulate air temperature, humidity and quality. They are widely used in commercial buildings, industrial facilities and homes. Their main functions include cooling, heating, ventilation and air purification.

[0003] Traditional methods usually rely on manual inspections and empirical judgments, and cannot achieve real-time monitoring, resulting in a long response time for fault discovery and handling. Traditional methods rely on visual inspections or simple instrument measurements, which are easily affected by human factors and subjective judgments, and may lead to misdiagnosis or missed diagnosis. Traditional methods often do not systematically collect and analyze equipment operation data and lack scientific basis, making fault diagnosis and prediction inaccurate. The process of manually analyzing and recording fault information is cumbersome and inefficient, especially when a large number of devices need to be processed, which can easily lead to information omissions and delays. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the disadvantages of the above-mentioned prior art and provide a method and system for diagnosing the fault type of an air-conditioning unit.

[0005] The technical solution adopted to solve the above technical problem is: a method for diagnosing the fault type of an air conditioning unit, comprising:

[0006] Based on the sensor group, the state of the target air-conditioning unit is monitored according to a preset sampling period to obtain a state data sequence of the target air-conditioning unit, and data fitting is performed on the state data sequence to obtain a state fitting curve;

[0007] Matching the state fitting curve with a preset standard state fitting curve to obtain a matching degree between the two, and comparing the matching degree with a preset matching degree threshold;

[0008] If the matching degree is lower than a preset threshold, the target air-conditioning unit is sound collected based on an audio collection device to obtain a sound signal of the target air-conditioning unit, and the sound signal is preprocessed to obtain a standard sound signal;

[0009] Performing feature extraction on the standard sound signal to obtain Mel-frequency cepstral coefficients of the standard sound signal, and performing fault classification according to the Mel-frequency cepstral coefficients based on a trained fault classification model to obtain a fault label of the target air-conditioning unit;

[0010] The fault label is matched with a preset fault state map set to obtain a plurality of neighborhood fault labels corresponding to the fault label, and a repair work order set is constructed based on the fault label, the plurality of neighborhood fault labels and the state data sequence.

[0011] Preferably, preprocessing the sound signal to obtain a standard sound signal includes:

[0012] The sound signal is subjected to emphasis processing, wherein the expression of the emphasis processing is as follows:

[0013]

[0014] Wherein, H(z) represents the filter for emphasis processing, a represents the emphasis parameter, and z represents the parameter of the filter for emphasis processing;

[0015] The sound signal after the emphasis processing is divided into a plurality of sub-sound signals, and each of the sub-sound signals has an overlapping part, wherein the expression of the overlapping part is as follows:

[0016]

[0017] Wherein, F represents the total number of sub-sound signals in the sound signal, w represents the length of the sub-sound signal, o represents the overlap rate, and n represents the total length of the sound signal;

[0018] The plurality of sub-sound signals are subjected to windowing processing to obtain a standard sound signal, wherein the expression of the windowing processing is as follows:

[0019]

[0020] Wherein, ω(n) represents the windowing process of the sub-sound signal, and L represents the window length of the windowing window.

[0021] Preferably, feature extraction is performed on the standard sound signal to obtain Mel-frequency cepstral coefficients of the standard sound signal, including:

[0022] Performing a discrete Fourier transform on the standard sound signal to obtain frequency domain data of the standard sound signal, wherein the expression of the discrete Fourier transform is as follows:

[0023]

[0024] Wherein, X(k) represents the complex spectral component at the frequency index k of the standard sound signal, N represents the total number of samples of the standard sound signal, x(n) represents the sample value of the time domain signal at the time index n of the standard sound signal, and j represents an imaginary unit;

[0025] Calculate the spectral line energy of the frequency domain data, wherein the calculation formula of the spectral line energy is as follows:

[0026] P m =|X(k)| 2 ;

[0027] Among them, P m Represents the spectral line energy of the complex spectral component at the frequency index k of the standard sound signal;

[0028] The logarithmic spectrum of the frequency domain data is calculated based on the spectral line energy, wherein the calculation formula of the logarithmic spectrum is as follows:

[0029]

[0030] Among them, S m represents the logarithmic spectrum of the complex spectral component at frequency index k of the standard sound signal, H m (k) represents a triangular bandpass filter, and f(m) represents the center frequency;

[0031] Performing discrete cosine transformation on the logarithmic spectrum of the frequency domain data to obtain Mel-cepstral coefficients of the standard sound signal, wherein the expression of the Mel-cepstral coefficients is as follows:

[0032]

[0033] Wherein, C(n) represents the Mel-frequency cepstrum coefficient of the standard sound signal.

[0034] Preferably, the fault classification model adopts a feedforward neural network, and the fault classification model includes an input layer, a hidden layer and an output layer, the input layer and the hidden layer are fully connected, and the hidden layer and the output layer are sparsely connected, wherein the training method of the fault classification model includes:

[0035] Randomly initialize the weight matrix W, initialize the learning rate α, and initialize the stability factor m;

[0036] The feature vector corresponding to the Mel cepstral coefficient is input into the input layer, and the i-th training sample x in the feature vector is extracted. i , calculate the jth neuron and the training sample x i The Euclidean distance between them is calculated as follows:

[0037] d j =||x i -ω ij ||2;

[0038] Among them, d i Represents the jth neuron and the training sample x i The Euclidean distance between ij Represents the jth neuron and the training sample x i The weight between

[0039] Find the training sample x i The two hidden layer neurons j and k with the smallest Euclidean distance are selected, and their output types are marked as c j and c k ;

[0040] Determine the type of the output c j and c k Are they the same? If the output type is c j and c k If they are not the same, then the output type c is determined. j and the training sample x i Whether the corresponding label types are the same, if they are the same, the weight matrix W is updated by the first weight update formula, otherwise, the output type c is determined k and the training sample x i Whether the corresponding label types are the same, if they are the same, the weight matrix W is updated by the second weight update formula;

[0041] If the hidden layer neuron j is related to the training sample x i The Euclidean distance is less than the distance between the hidden layer neuron j and the training sample x i The above operation is repeated until the hidden layer neuron j is close to the training sample x. i The Euclidean distance is greater than or equal to the distance between the hidden layer neuron j and the training sample x i The Euclidean distance of

[0042] Information propagation is performed based on the updated weight matrix W to obtain a trained fault classification model.

[0043] Preferably, the first weight updating formula is as follows:

[0044]

[0045] Among them, ω ik Represents the kth neuron and the training sample x i The weight between , β represents the bias coefficient;

[0046] The second weight update formula is as follows:

[0047]

[0048] Preferably, the status data includes the name of the air-conditioning unit, monitoring data, weather data and energy consumption standardization value, wherein the monitoring data includes hot and cold water valve opening feedback, fresh air temperature, supply air temperature, return air temperature, supply air humidity and return air humidity.

[0049] Preferably, the calculation formula of the matching degree is as follows:

[0050]

[0051] Among them, M represents the matching degree between the state fitting curve and the preset standard state fitting curve, y i represents the data point of the standard state curve, represents the data points of the waiting state fitting curve, and B represents the total number of data points.

[0052] Preferably, matching the fault label with a preset fault state map set to obtain a plurality of neighborhood fault labels corresponding to the fault label includes:

[0053] The support degree of the fault label and a plurality of fault state graphs in the preset fault state graph set is calculated, wherein the calculation formula of the support degree is as follows:

[0054]

[0055] Among them, sup i represents the support between the fault label and the i-th fault state graph in the preset fault state graph set, represents the number of occurrences of the fault label in the i-th fault state graph in the preset fault state graph set, and n represents the total number of nodes in the fault state graph;

[0056] The preset fault state graph set is screened based on the support to obtain a plurality of matching preset fault state graphs;

[0057] The plurality of matching preset fault state maps are compared for differences to obtain a set of difference coefficients, wherein the calculation formula of the difference coefficients is as follows:

[0058]

[0059] Among them, A i,j represents the difference coefficient between the i-th matching preset fault state map and the j-th matching preset fault state map, D i,j (v) represents the number of node differences between the i-th matching preset fault state graph and the j-th matching preset fault state graph, D i,j(u) represents the difference in the number of associated edges between the i-th matching preset fault state graph and the j-th matching preset fault state graph, N n Represents the sum of the number of vertices and the number of labels of the associated edges between the i-th matching preset fault state graph and the j-th matching preset fault state graph;

[0060] Sorting the difference coefficient set to obtain a sequential difference coefficient set, and obtaining matching preset fault state graphs corresponding to the first, middle and last difference coefficients in the sequential difference coefficient set;

[0061] Obtain the fault state nodes corresponding to the preset fault state map that match the fault label with the first, middle and last difference coefficients, and use the neighbor nodes of the fault state node as neighborhood fault labels to obtain multiple neighborhood fault labels.

[0062] Preferably, the expression of the repair work order set is as follows:

[0063]

[0064] Among them, RE represents the repair work order set. represents the name of the air conditioning unit, τ represents the monitoring data, v represents the weather data, represents the normalized value of energy consumption, φ represents the fault label, and χ represents multiple neighborhood fault labels.

[0065] The technical solution adopted to solve the above technical problem is: an air conditioning unit fault type diagnosis system, which is applicable to the air conditioning unit fault type diagnosis method, including:

[0066] A state monitoring unit, the state monitoring unit is used to monitor the state of the target air-conditioning unit based on the sensor group according to a preset sampling period to obtain a state data sequence of the target air-conditioning unit, and perform data fitting on the state data sequence to obtain a state fitting curve;

[0067] A curve matching unit, the curve matching unit is used to match the state fitting curve with a preset standard state fitting curve to obtain a matching degree between the two, and compare the matching degree with a preset matching degree threshold;

[0068] A sound collection unit, wherein if the sound is lower than a preset matching degree threshold, the sound collection unit collects sound from the target air-conditioning unit based on an audio collection device to obtain a sound signal of the target air-conditioning unit, and pre-processes the sound signal to obtain a standard sound signal;

[0069] A fault diagnosis unit, the fault diagnosis unit is used to extract features from the standard sound signal to obtain Mel cepstral coefficients of the standard sound signal, and perform fault classification according to the Mel cepstral coefficients based on a trained fault classification model to obtain a fault label of the target air-conditioning unit;

[0070] A work order construction unit, the work order construction unit is used to match the fault label with a preset fault state map set to obtain multiple neighborhood fault labels corresponding to the fault label, and to construct a repair work order set based on the fault label, the multiple neighborhood fault labels and the state data sequence.

[0071] The beneficial effects of the present invention are as follows: (1) The present invention monitors the status of the air-conditioning unit through a sensor group, can detect abnormal conditions in time, facilitate early diagnosis and processing, reduce the risk of equipment failure, and by combining the fitting of the status data sequence with the matching of the standard curve, can more accurately identify the operating status of the equipment, thereby improving the accuracy of fault detection; (2) The present invention can effectively capture the acoustic characteristics of equipment failures through audio acquisition and sound signal processing, using feature extraction techniques such as Mel-frequency cepstral coefficients, which provides strong data support for fault classification, and by matching with a preset fault state spectrum set, can identify neighboring faults related to the main fault, help maintenance personnel fully understand potential problems, and optimize maintenance strategies; (3) The present invention can realize the automatic processing of fault information by constructing a repair work order set based on fault labels and related data, improve work efficiency, reduce manual intervention, and through accurate fault diagnosis and early warning, can reduce the maintenance cost and downtime loss caused by sudden failures, thereby extending the service life of equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 A schematic flow chart of the steps of an overall method in an embodiment of the present invention;

[0073] Figure 2 A schematic diagram of the system architecture of an overall system in an embodiment of the present invention.

[0074] Figure numerals: 1. state monitoring unit; 2. curve matching unit; 3. sound collection unit; 4. fault diagnosis unit; 5. work order building unit. DETAILED DESCRIPTION

[0075] Embodiment 1, as Figure 1 As shown, the air conditioning unit fault type diagnosis method proposed by the present invention includes:

[0076] S1. Based on the sensor group, the target air-conditioning unit is monitored according to a preset sampling period to obtain a state data sequence of the target air-conditioning unit, and the state data sequence is fitted to obtain a state fitting curve;

[0077] S2, matching the state fitting curve with a preset standard state fitting curve to obtain a matching degree between the two, and comparing the matching degree with a preset matching degree threshold;

[0078] S3. If the matching degree is lower than a preset matching degree threshold, the target air-conditioning unit is sound collected based on an audio collection device to obtain a sound signal of the target air-conditioning unit, and the sound signal is preprocessed to obtain a standard sound signal;

[0079] S4, extracting features of the standard sound signal to obtain Mel-frequency cepstral coefficients of the standard sound signal, and performing fault classification based on the trained fault classification model according to the Mel-frequency cepstral coefficients to obtain a fault label of the target air-conditioning unit;

[0080] S5. Match the fault label with a preset fault state map set to obtain multiple neighborhood fault labels corresponding to the fault label, and construct a repair work order set based on the fault label, the multiple neighborhood fault labels and the state data sequence.

[0081] In the present invention, the standard state fitting curve refers to a pre-set ideal state curve for comparison with actual data; the audio acquisition device refers to a device for recording sound signals, such as a microphone, for collecting the operating sound of the air-conditioning unit; the Mel-frequency cepstral coefficient (MFCC) refers to a common sound feature representation method, which is particularly suitable for speech and audio signal analysis and can capture the frequency and energy characteristics of sound; the fault state spectrum set refers to a set of pre-defined various faults and their manifestation characteristics for comparison and analysis.

[0082] Embodiment 2, the method for diagnosing the fault type of an air conditioner unit proposed by the present invention, compared with embodiment 1, this embodiment further includes: preprocessing the sound signal to obtain a standard sound signal, including:

[0083] A1. Perform emphasis processing on the sound signal, wherein the expression of emphasis processing is as follows:

[0084]

[0085] Wherein, H(z) represents the filter for emphasis processing, a represents the emphasis parameter, and z represents the parameter of the filter for emphasis processing;

[0086] A2. Divide the emphasized sound signal into a plurality of sub-sound signals, and each sub-sound signal has an overlapping portion, wherein the expression of the overlapping portion is as follows:

[0087]

[0088] Wherein, F represents the total number of sub-sound signals in the sound signal, w represents the length of the sub-sound signal, o represents the overlap rate, and n represents the total length of the sound signal;

[0089] A3. Windowing is performed on the multiple sub-sound signals to obtain a standard sound signal, wherein the expression of the windowing process is as follows:

[0090]

[0091] Wherein, ω(n) represents the windowing process of the sub-sound signal, and L represents the window length of the windowing window.

[0092] In an optional embodiment, feature extraction is performed on the standard sound signal to obtain Mel-frequency cepstrum coefficients of the standard sound signal, including:

[0093] B1. Perform discrete Fourier transform on the standard sound signal to obtain frequency domain data of the standard sound signal, wherein the expression of discrete Fourier transform is as follows:

[0094]

[0095] Wherein, X(k) represents the complex spectral component at the frequency index k of the standard sound signal, N represents the total number of samples of the standard sound signal, x(n) represents the sample value of the time domain signal at the time index n of the standard sound signal, and j represents an imaginary unit;

[0096] B2. Calculate the spectral line energy of the frequency domain data, where the calculation formula of the spectral line energy is as follows:

[0097] P m =|X(k)| 2 ;

[0098] Among them, P m Represents the spectral line energy of the complex spectral component at the frequency index k of the standard sound signal;

[0099] B3. Calculate the logarithmic spectrum of the frequency domain data based on the spectral line energy, where the calculation formula of the logarithmic spectrum is as follows:

[0100]

[0101] Among them, S m represents the logarithmic spectrum of the complex spectral component at frequency index k of the standard sound signal, H m (k) represents a triangular bandpass filter, and f(m) represents the center frequency;

[0102] B4. Perform discrete cosine transformation on the logarithmic spectrum of the frequency domain data to obtain the Mel-cepstral coefficient of the standard sound signal, where the expression of the Mel-cepstral coefficient is as follows:

[0103]

[0104] Wherein, C(n) represents the Mel-frequency cepstrum coefficient of the standard sound signal.

[0105] In an optional embodiment, the fault classification model adopts a feedforward neural network, and the fault classification model includes an input layer, a hidden layer, and an output layer. The input layer and the hidden layer are fully connected, and the hidden layer and the output layer are sparsely connected. The training method of the fault classification model includes:

[0106] C1. Randomly initialize the weight matrix W, initialize the learning rate α, and initialize the stability factor m;

[0107] C2. Input the feature vector corresponding to the Mel cepstral coefficient into the input layer and extract the i-th training sample x in the feature vector. i , calculate the jth neuron and the training sample x i The Euclidean distance between them is calculated as follows:

[0108] d j =||x i -ω ij ||2;

[0109] Among them, d i Represents the jth neuron and the training sample x i The Euclidean distance between ij Represents the jth neuron and the training sample x i The weight between

[0110] C3, find the training sample x i The two hidden layer neurons j and k with the smallest Euclidean distance are selected, and their output types are marked as c j and c k ;

[0111] C4. Determine the output type c j and c k Are they the same? If the output type is c j and c k If they are not the same, then determine the output type c j And training samples x i Whether the corresponding label types are the same, if they are the same, the weight matrix W is updated by the first weight update formula, otherwise, the output type c is determined kAnd training samples x i Whether the corresponding label types are the same, if they are the same, the weight matrix W is updated by the second weight update formula;

[0112] C5, if the hidden layer neuron j and the training sample x i The Euclidean distance is less than the distance between the hidden layer neuron j and the training sample x i The Euclidean distance between the hidden layer neuron j and the training sample x is repeated until i The Euclidean distance is greater than or equal to the distance between the hidden layer neuron j and the training sample x i The Euclidean distance of

[0113] C6. Perform information propagation based on the updated weight matrix W to obtain a trained fault classification model.

[0114] In an optional embodiment, the first weight updating formula is as follows:

[0115]

[0116] Among them, ω ik Represents the kth neuron and the training sample x i The weight between , β represents the bias coefficient;

[0117] The second weight update formula is as follows:

[0118]

[0119] In an optional embodiment, the status data includes the air conditioning unit name, monitoring data, weather data and energy consumption standardization value, wherein the monitoring data includes hot and cold water valve opening feedback, fresh air temperature, supply air temperature, return air temperature, supply air humidity and return air humidity.

[0120] In an optional embodiment, the calculation formula of the matching degree is as follows:

[0121]

[0122] Among them, M represents the matching degree between the state fitting curve and the preset standard state fitting curve, y i represents the data point of the standard state curve, represents the data points of the waiting state fitting curve, and B represents the total number of data points.

[0123] In an optional embodiment, the fault label is matched with a preset fault state map set to obtain a plurality of neighborhood fault labels corresponding to the fault label, including:

[0124] D1. Calculate the support between the fault label and multiple fault state graphs in the preset fault state graph set, where the calculation formula of the support is as follows:

[0125]

[0126] Among them, sup i represents the support between the fault label and the i-th fault state graph in the preset fault state graph set, represents the number of occurrences of the fault label in the i-th fault state graph in the preset fault state graph set, and n represents the total number of nodes in the fault state graph;

[0127] D2. Screening the preset fault state graph set based on the support to obtain a plurality of matching preset fault state graphs;

[0128] D3. Compare the differences of multiple matching preset fault state graphs to obtain a set of difference coefficients, wherein the calculation formula of the difference coefficient is as follows:

[0129]

[0130] Among them, A i,j represents the difference coefficient between the i-th matching preset fault state map and the j-th matching preset fault state map, D i,j (v) represents the number of node differences between the i-th matching preset fault state graph and the j-th matching preset fault state graph, D i,j (u) represents the difference in the number of associated edges between the i-th matching preset fault state graph and the j-th matching preset fault state graph, N n Represents the sum of the number of vertices and the number of labels of the associated edges between the i-th matching preset fault state graph and the j-th matching preset fault state graph;

[0131] D4. Sort the difference coefficient set to obtain a sequential difference coefficient set, and obtain matching preset fault state graphs corresponding to the first, middle, and last difference coefficients in the sequential difference coefficient set;

[0132] D5. Obtain the fault state nodes corresponding to the preset fault state map that match the fault label with the first, middle and last difference coefficients, and use the neighbor nodes of the fault state node as neighborhood fault labels to obtain multiple neighborhood fault labels.

[0133] In an optional embodiment, the expression of the repair work order set is as follows:

[0134]

[0135] Among them, RE represents the repair work order set. represents the name of the air conditioning unit, τ represents the monitoring data, v represents the weather data, represents the normalized value of energy consumption, φ represents the fault label, and χ represents multiple neighborhood fault labels.

[0136] Embodiment three, as Figure 2 As shown, the air conditioning unit fault type diagnosis system proposed by the present invention is applicable to the air conditioning unit fault type diagnosis method, including:

[0137] The state monitoring unit 1 is used to monitor the state of the target air-conditioning unit based on the sensor group according to a preset sampling period to obtain a state data sequence of the target air-conditioning unit, and perform data fitting on the state data sequence to obtain a state fitting curve;

[0138] The curve matching unit 2 is used to match the state fitting curve with a preset standard state fitting curve to obtain a matching degree between the two, and compare the matching degree with a preset matching degree threshold;

[0139] The sound collection unit 3 is used to collect sound from the target air-conditioning unit based on the audio collection device if the sound is lower than a preset matching degree threshold, so as to obtain a sound signal of the target air-conditioning unit, and pre-process the sound signal to obtain a standard sound signal;

[0140] A fault diagnosis unit 4 is used to extract features from the standard sound signal to obtain the Mel cepstral coefficients of the standard sound signal, and to classify the faults according to the Mel cepstral coefficients based on the trained fault classification model to obtain a fault label of the target air-conditioning unit;

[0141] The work order construction unit 5 is used to match the fault label with a preset fault state map set to obtain multiple neighborhood fault labels corresponding to the fault label, and to construct a repair work order set based on the fault label, multiple neighborhood fault labels and state data sequence.

[0142] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. A method for diagnosing fault types of air conditioner units, characterized in that: include: Based on the sensor group, the state of the target air-conditioning unit is monitored according to a preset sampling period to obtain a state data sequence of the target air-conditioning unit, and data fitting is performed on the state data sequence to obtain a state fitting curve; Matching the state fitting curve with a preset standard state fitting curve to obtain a matching degree between the two, and comparing the matching degree with a preset matching degree threshold; If the matching degree is lower than a preset threshold, the target air-conditioning unit is sound collected based on an audio collection device to obtain a sound signal of the target air-conditioning unit, and the sound signal is preprocessed to obtain a standard sound signal; Performing feature extraction on the standard sound signal to obtain Mel-frequency cepstral coefficients of the standard sound signal, and performing fault classification according to the Mel-frequency cepstral coefficients based on a trained fault classification model to obtain a fault label of the target air-conditioning unit; The fault label is matched with a preset fault state map set to obtain a plurality of neighborhood fault labels corresponding to the fault label, and a repair work order set is constructed based on the fault label, the plurality of neighborhood fault labels and the state data sequence.

2. The method for diagnosing fault types of air conditioner units according to claim 1, characterized in that: Preprocessing the sound signal to obtain a standard sound signal includes: The sound signal is subjected to emphasis processing, wherein the expression of the emphasis processing is as follows: Wherein, H(z) represents the filter for emphasis processing, a represents the emphasis parameter, and z represents the parameter of the filter for emphasis processing; The sound signal after the emphasis processing is divided into a plurality of sub-sound signals, and each of the sub-sound signals has an overlapping part, wherein the expression of the overlapping part is as follows: Wherein, F represents the total number of sub-sound signals in the sound signal, w represents the length of the sub-sound signal, o represents the overlap rate, and n represents the total length of the sound signal; The plurality of sub-sound signals are subjected to windowing processing to obtain a standard sound signal, wherein the expression of the windowing processing is as follows: Wherein, ω(n) represents the windowing process of the sub-sound signal, and L represents the window length of the windowing window.

3. The method for diagnosing fault types of air conditioner units according to claim 2, characterized in that: Extracting features from the standard sound signal to obtain Mel-frequency cepstrum coefficients of the standard sound signal includes: Performing a discrete Fourier transform on the standard sound signal to obtain frequency domain data of the standard sound signal, wherein the expression of the discrete Fourier transform is as follows: Wherein, X(k) represents the complex spectral component at the frequency index k of the standard sound signal, N represents the total number of samples of the standard sound signal, x(n) represents the sample value of the time domain signal at the time index n of the standard sound signal, and j represents an imaginary unit; Calculate the spectral line energy of the frequency domain data, wherein the calculation formula of the spectral line energy is as follows: P m =|X(k)| 2 ; Among them, P m Represents the spectral line energy of the complex spectral component at the frequency index k of the standard sound signal; The logarithmic spectrum of the frequency domain data is calculated based on the spectral line energy, wherein the calculation formula of the logarithmic spectrum is as follows: Among them, S m represents the logarithmic spectrum of the complex spectral component at frequency index k of the standard sound signal, H m (k) represents a triangular bandpass filter, and f(m) represents the center frequency; Performing discrete cosine transformation on the logarithmic spectrum of the frequency domain data to obtain Mel-cepstral coefficients of the standard sound signal, wherein the expression of the Mel-cepstral coefficients is as follows: Wherein, C(n) represents the Mel-frequency cepstrum coefficient of the standard sound signal.

4. The method for diagnosing fault types of air conditioner units according to claim 1, characterized in that: The fault classification model adopts a feedforward neural network, and the fault classification model includes an input layer, a hidden layer and an output layer. The input layer and the hidden layer are fully connected, and the hidden layer and the output layer are sparsely connected. The training method of the fault classification model includes: Randomly initialize the weight matrix W, initialize the learning rate α, and initialize the stability factor m; The feature vector corresponding to the Mel cepstral coefficient is input into the input layer, and the i-th training sample x in the feature vector is extracted. i , calculate the jth neuron and the training sample x i The Euclidean distance between them is calculated as follows: d j =||x i -oh ij ||2; Among them, d i Represents the jth neuron and the training sample x i The Euclidean distance between ij Represents the jth neuron and the training sample x i The weight between Find the training sample x i The two hidden layer neurons j and k with the smallest Euclidean distance are selected, and their output types are marked as c j and c k ; Determine the type of the output c j and c k Are they the same? If the output type is c j and c k If they are not the same, then the output type c is determined. j and the training sample x i Whether the corresponding label types are the same, if they are the same, the weight matrix W is updated by the first weight update formula, otherwise, the output type c is determined k and the training sample x i Whether the corresponding label types are the same, if they are the same, the weight matrix W is updated by the second weight update formula; If the hidden layer neuron j is related to the training sample x i The Euclidean distance is less than the distance between the hidden layer neuron j and the training sample x i The above operation is repeated until the hidden layer neuron j is close to the training sample x. i The Euclidean distance is greater than or equal to the distance between the hidden layer neuron j and the training sample x i The Euclidean distance of Information propagation is performed based on the updated weight matrix W to obtain a trained fault classification model.

5. The method for diagnosing fault types of air conditioner units according to claim 4, characterized in that: The first weight update formula is as follows: Among them, ω ik Represents the kth neuron and the training sample x i The weight between , β represents the bias coefficient; The second weight update formula is as follows:

6. The method for diagnosing fault types of air conditioner units according to claim 1, characterized in that: The status data includes the name of the air-conditioning unit, monitoring data, weather data and energy consumption standardization value, wherein the monitoring data includes hot and cold water valve opening feedback, fresh air temperature, supply air temperature, return air temperature, supply air humidity and return air humidity.

7. The method for diagnosing fault types of air conditioner units according to claim 1, characterized in that: The calculation formula of the matching degree is as follows: Among them, M represents the matching degree between the state fitting curve and the preset standard state fitting curve, y i represents the data point of the standard state curve, represents the data points of the waiting state fitting curve, and B represents the total number of data points.

8. The method for diagnosing fault types of air conditioner units according to claim 1, characterized in that: Matching the fault label with a preset fault state map set to obtain a plurality of neighborhood fault labels corresponding to the fault label, including: The support degree of the fault label and a plurality of fault state graphs in the preset fault state graph set is calculated, wherein the calculation formula of the support degree is as follows: Among them, sup i represents the support between the fault label and the i-th fault state graph in the preset fault state graph set, represents the number of occurrences of the fault label in the i-th fault state graph in the preset fault state graph set, and n represents the total number of nodes in the fault state graph; The preset fault state graph set is screened based on the support to obtain a plurality of matching preset fault state graphs; The plurality of matching preset fault state maps are compared for differences to obtain a set of difference coefficients, wherein the calculation formula of the difference coefficients is as follows: Among them, A i,j represents the difference coefficient between the i-th matching preset fault state map and the j-th matching preset fault state map, D i,j (v) represents the number of node differences between the i-th matching preset fault state graph and the j-th matching preset fault state graph, D i,j (u) represents the difference in the number of associated edges between the i-th matching preset fault state graph and the j-th matching preset fault state graph, N n Represents the sum of the number of vertices and the number of labels of the associated edges between the i-th matching preset fault state graph and the j-th matching preset fault state graph; Sorting the difference coefficient set to obtain a sequential difference coefficient set, and obtaining matching preset fault state graphs corresponding to the first, middle and last difference coefficients in the sequential difference coefficient set; Obtain the fault state nodes corresponding to the preset fault state map that match the fault label with the first, middle and last difference coefficients, and use the neighbor nodes of the fault state node as neighborhood fault labels to obtain multiple neighborhood fault labels.

9. The method for diagnosing fault types of air conditioner units according to claim 1, characterized in that: The expression of the repair work order set is as follows: Among them, RE represents the repair work order set. represents the name of the air conditioning unit, τ represents the monitoring data, υ represents the weather data, represents the normalized value of energy consumption, φ represents the fault label, and χ represents multiple neighborhood fault labels.

10. An air conditioning unit fault type diagnosis system, which is applicable to the air conditioning unit fault type diagnosis method according to any one of claims 1 to 9, characterized in that: include: A state monitoring unit (1), the state monitoring unit (1) being used to monitor the state of a target air-conditioning unit based on a sensor group according to a preset sampling period to obtain a state data sequence of the target air-conditioning unit, and to perform data fitting on the state data sequence to obtain a state fitting curve; A curve matching unit (2), the curve matching unit (2) being used to match the state fitting curve with a preset standard state fitting curve to obtain a matching degree between the two, and to compare the matching degree with a preset matching degree threshold; A sound collection unit (3), the sound collection unit (3) being used for collecting sound from a target air-conditioning unit based on an audio collection device if the sound is lower than a preset matching degree threshold, so as to obtain a sound signal of the target air-conditioning unit, and pre-processing the sound signal to obtain a standard sound signal; A fault diagnosis unit (4), the fault diagnosis unit (4) being used to extract features from the standard sound signal to obtain Mel-frequency cepstral coefficients of the standard sound signal, and to perform fault classification based on the Mel-frequency cepstral coefficients based on a trained fault classification model to obtain a fault label of the target air-conditioning unit; A work order construction unit (5), the work order construction unit (5) is used to match the fault label with a preset fault state map set to obtain multiple neighborhood fault labels corresponding to the fault label, and to construct a repair work order set based on the fault label, the multiple neighborhood fault labels and the state data sequence.

Images