Dairy beverage mechanical equipment fault diagnosis method based on data analysis

Abstract

The invention discloses a dairy beverage mechanical equipment fault diagnosis method based on data analysis, and relates to the technical field of dairy beverage mechanical equipment fault diagnosis, and the method comprises the steps: setting an initial ultrasonic frequency and an ultrasonic frequency range; acquiring a plurality of test ultrasonic frequencies in the ultrasonic frequency range; respectively transmitting a plurality of ultrasonic waves for testing ultrasonic frequencies to parts to be diagnosed of the pasteurization machine and the sterilization temperature control equipment, and performing fault diagnosis; respectively acquiring diagnosis accurate data and diagnosis efficiency data of fault diagnosis under different test ultrasonic frequencies; constructing an ultrasonic frequency evaluation model based on a bp neural network; and an ultrasonic frequency-diagnosis effect display model is constructed, the screened ultrasonic frequency is obtained, and the ultrasonic frequency-diagnosis effect display model is applied to fault diagnosis of dairy beverage mechanical equipment, so that the problem that the diagnosis effect of pasteurization machine and sterilization temperature control equipment faults is affected due to the fact that the setting of the existing ultrasonic frequency is not clear is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis of dairy beverage machinery and equipment, and more specifically, to a method for fault diagnosis of dairy beverage machinery and equipment based on data analysis. Background Art

[0002] Dairy and beverage machinery and equipment technology mainly covers the production, processing, packaging and cleaning of dairy products and beverages. In the processing process, common technologies include pasteurization and ultra-high temperature sterilization (UHT) to ensure product safety and extend shelf life; homogenization technology and fermentation technology are used to improve the taste and flavor of dairy products.

[0003] Modern dairy beverage production generally adopts automation and intelligent technologies, such as real-time monitoring through SCADA systems, and machine learning to optimize production processes, improve efficiency and reduce energy consumption. In terms of packaging, aseptic packaging technology is widely used to achieve long-term storage at room temperature, and environmentally friendly materials have also become a new direction for the development of the industry.

[0004] In terms of cleaning and disinfection, CIP (cleaning in place) technology simplifies the equipment cleaning process, while ultraviolet and ozone disinfection are used to maintain a sterile environment. In addition, innovative technologies such as high-pressure processing and ultrasonic treatment continue to improve product quality and meet consumers' demands for nutrition, taste and functionality.

[0005] With the development of the industry, energy conservation and consumption reduction, food safety and personalized customized production have become important trends, driving the continuous upgrading of dairy beverage machinery and equipment technology.

[0006] In pasteurization and ultra-high temperature sterilization, failure of the temperature control system may cause the sterilization temperature to deviate from the set value, affecting the quality and safety of the product. Scaling or blockage of the heat exchanger is also a common problem, which reduces thermal efficiency and increases energy consumption. Through real-time monitoring of temperature sensors and ultrasonic detection technology, potential problems in temperature control systems and heat exchangers can be effectively identified.

[0007] The above disclosed technical solutions have at least the following technical problems: In the diagnosis of pasteurization machine and sterilization temperature control equipment faults, the frequency of ultrasound will affect its penetration depth and resolution. Although high-frequency ultrasound can provide higher resolution, its penetration ability is weaker and it is suitable for detecting surface or shallow scaling, cracks and other problems. Low-frequency ultrasound can better penetrate materials and is suitable for detecting deeper problems, but its resolution is lower.

[0008] If too high an ultrasonic frequency is used, deeper heat exchanger problems, such as deeper blockages or scaling, may be missed; if too low an ultrasonic frequency is used, deeper anomalies may be detected, but the resolution may be low and small surface faults may not be accurately identified.

[0009] The existing ultrasonic frequency settings are unclear, which makes it impossible to find a balance between penetration depth and resolution, thus affecting the diagnostic effect of pasteurizer and sterilization temperature control equipment faults.

[0010] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0011] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a dairy beverage machinery and equipment fault diagnosis method based on data analysis, which performs real-time analysis and adjustment of the ultrasonic frequency to solve the problem that the existing ultrasonic frequency setting is unclear, resulting in an inability to find a balance between penetration depth and resolution, thereby affecting the diagnostic effect of pasteurizer and sterilization temperature control equipment faults.

[0012] To achieve the above object, the present invention provides the following technical solutions:

[0013] A method for diagnosing faults of dairy and beverage machinery and equipment based on data analysis comprises the following steps: obtaining equipment data of dairy and beverage machinery and equipment, setting an initial ultrasonic frequency and an ultrasonic frequency range according to the equipment data; performing random number adjustments on the initial ultrasonic frequency several times to obtain several test ultrasonic frequencies within the ultrasonic frequency range; transmitting ultrasonic waves of several test ultrasonic frequencies to the components to be diagnosed of a pasteurizer and a sterilization temperature control device, respectively, and analyzing the intensity and spectrum of the corresponding signal echoes to perform fault diagnosis; respectively obtaining diagnostic accuracy data and diagnostic efficiency data of fault diagnosis under different test ultrasonic frequencies; constructing an ultrasonic frequency evaluation model based on a BP neural network according to the diagnostic effect data and the diagnostic efficiency data; constructing an ultrasonic frequency-diagnostic effect display model according to different test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model, and performing data analysis to obtain the screened ultrasonic frequency for application in the diagnosis of dairy and beverage machinery and equipment.

[0014] In a preferred embodiment, the equipment data of the dairy beverage machinery and equipment is obtained, and the initial ultrasonic frequency and ultrasonic frequency range are set according to the equipment data, specifically: the equipment data includes equipment material, wall thickness, temperature, working pressure and surface condition; data analysis is performed on the relationship between each equipment data and the ultrasonic frequency to obtain the constraints of each equipment data on the ultrasonic frequency, and the ultrasonic frequency range is set according to the constraints of each equipment data on the ultrasonic frequency; the initial ultrasonic frequency is set according to historical data.

[0015] In a preferred embodiment, the initial ultrasonic frequency is subjected to several random number adjustments to obtain several test ultrasonic frequencies within the ultrasonic frequency range, specifically: several random numbers are generated by a computer; the initial ultrasonic frequency is subjected to uniform random distribution processing according to each random number to generate several test ultrasonic frequencies; and the test ultrasonic frequencies outside the ultrasonic frequency range are eliminated.

[0016] In a preferred embodiment, the intensity and spectrum of the corresponding signal echo are analyzed to perform fault diagnosis, specifically: the echo intensity curve is fitted based on the least squares method, the deviation from the preset standard echo intensity curve is calculated, and the possibility of the fault is classified according to the deviation, and the possibility classification includes weak faults and serious faults; the echo signal is Fourier transformed to obtain a signal spectrum in the frequency domain, and the spectral features of the signal spectrum are extracted to perform spectral anomaly classification, and the spectral anomaly classification includes spectrum narrowing and spectrum offset.

[0017] In a preferred embodiment, the diagnostic accuracy data includes an echo resolution influence coefficient and a penetration fluctuation abnormality coefficient; the specific method for obtaining the echo resolution influence coefficient is as follows: extract time amplitude data from the ultrasonic signal echo; preprocess the ultrasonic echo, and calculate the amplitude distribution characteristics of the ultrasonic echo, the amplitude distribution characteristics include the echo mean and the echo variance; obtain the spectral density of the ultrasonic echo based on the amplitude distribution characteristics, obtain the high-frequency and low-frequency self-entropy of the ultrasonic echo according to the spectral density, and calculate the detail resolution entropy between the ultrasonic echoes based on the self-entropy of the ultrasonic echo; obtain the minimum comparison value of the echo mean and the echo standard deviation, and add the minimum comparison value combined with the detail resolution entropy to calculate the echo resolution influence coefficient.

[0018] In a preferred embodiment, the specific method for obtaining the abnormal coefficient of penetration fluctuation is as follows: real-time acquisition of the ultrasonic signal penetration depth and signal attenuation rate corresponding to the same time interval within a preset time range when ultrasonic fault detection is performed, and numbering the ultrasonic signal penetration depth and signal attenuation rate; calculating the penetration standard deviation and the penetration average value through the ultrasonic signal penetration depth and signal attenuation rate corresponding to the same time interval within a preset time range when ultrasonic fault detection is performed; calculating the penetration coefficient of variation based on the penetration standard deviation and the penetration average value; calculating the abnormal coefficient of penetration fluctuation based on the penetration coefficient of variation.

[0019] In a preferred embodiment, the diagnostic efficiency data includes a detection efficiency influence coefficient; a specific method for obtaining the detection efficiency influence coefficient is as follows: based on the historical diagnostic equipment fault types of fault diagnosis at different ultrasonic frequencies, a diagnostic target set for fault diagnosis at different ultrasonic frequencies is obtained, wherein the diagnostic target set for fault diagnosis at different ultrasonic frequencies includes each equipment fault type for fault diagnosis and the diagnostic efficiency of each equipment fault type at different ultrasonic frequencies; a preset priority weight is assigned to each diagnostic equipment fault type within the diagnostic target set for fault diagnosis at different ultrasonic frequencies; a regression model is constructed based on the diagnostic target set for fault diagnosis at different ultrasonic frequencies and the corresponding priority weights to generate the detection efficiency influence coefficient.

[0020] In a preferred embodiment, the ultrasonic frequency-diagnostic effect display model is constructed according to different test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model, specifically: the test ultrasonic frequency is used as the x-axis on the rectangular coordinate system, and the output of the ultrasonic frequency evaluation model corresponding to the fault diagnosis under the test ultrasonic frequency is used as the y-axis of the rectangular coordinate system; several test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model are mapped to the rectangular coordinate system to obtain several data points; several data points are connected using a smooth curve to construct the ultrasonic frequency-diagnostic effect display model.

[0021] In a preferred embodiment, several test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model are mapped to a rectangular coordinate system to obtain several data points; the several data points are connected using a smooth curve to construct an ultrasonic frequency-diagnostic effect display model;

[0022] The data analysis is performed to obtain the screened ultrasonic frequency, which is applied to the fault diagnosis of dairy and beverage machinery and equipment. Specifically, a corresponding mathematical expression is established according to the ultrasonic frequency-diagnostic effect display model; each peak value in the ultrasonic frequency-diagnostic effect display model is obtained, and each peak value is quadratically differentiated based on the mathematical expression, and the test ultrasonic frequency corresponding to the peak value with the smallest quadratic derivative result is used as the screened ultrasonic frequency, which is applied to the fault diagnosis of dairy and beverage machinery and equipment.

[0023] Technical effects and advantages of the dairy beverage machinery equipment fault diagnosis method based on data analysis of the present invention:

[0024] 1. The present invention improves diagnostic accuracy and efficiency by optimizing frequency selection. The steps include data analysis, random adjustment of ultrasonic frequency, signal echo analysis, and data collection of diagnostic accuracy and efficiency. In particular, the echo resolution influence coefficient and the penetration fluctuation abnormal coefficient are introduced to optimize the resolution and penetration ability of ultrasonic frequency in equipment fault detection respectively. Through this method, the frequency can be flexibly adjusted according to the different materials, structures and working environments of the equipment to improve the accuracy and efficiency of fault detection. This method can not only accurately identify different types of faults, but also effectively reduce the misdiagnosis rate and missed diagnosis rate.

[0025] 2. The present invention adopts BP neural network to construct an ultrasonic frequency evaluation model, uses diagnostic effect data and diagnostic efficiency data to evaluate ultrasonic frequency, and optimizes the fault diagnosis process of dairy beverage machinery and equipment. Specifically, the ultrasonic frequency evaluation coefficient is generated through the echo resolution influence coefficient, the penetration fluctuation abnormal coefficient and the detection efficiency influence coefficient, and then the most suitable ultrasonic frequency for fault diagnosis is obtained. On this basis, an ultrasonic frequency-diagnostic effect display model is constructed, and its data is smoothed to further screen out effective ultrasonic frequencies. By taking the secondary derivative of the peak value, the optimal frequency is accurately identified, which improves the accuracy and efficiency of equipment fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The present invention is a flow chart of a method for diagnosing faults in dairy beverage machinery and equipment based on data analysis. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Embodiment 1, Figure 1 The present invention provides a method for diagnosing dairy beverage machinery faults based on data analysis, which includes the following steps:

[0029] S1, obtaining equipment data of dairy beverage machinery and equipment, and setting an initial ultrasonic frequency and an ultrasonic frequency range according to the equipment data.

[0030] The device data of the dairy beverage machine is obtained, and the initial ultrasonic frequency and ultrasonic frequency range are set according to the device data, specifically:

[0031] The equipment data includes equipment material, wall thickness, temperature, working pressure and surface condition;

[0032] Perform data analysis on the relationship between each device data and the ultrasonic frequency to obtain the constraint conditions of each device data on the ultrasonic frequency, and set the ultrasonic frequency range according to the constraint conditions of each device data on the ultrasonic frequency;

[0033] Set the initial ultrasonic frequency based on historical data.

[0034] It should be noted that the material of the equipment: such as stainless steel, aluminum alloy, etc., the density and sound speed of the material affect the propagation of ultrasonic waves; wall thickness: the wall thickness of the equipment components affects the penetration ability of ultrasonic waves. The thicker the wall, the lower the required ultrasonic frequency is usually; temperature: temperature changes will affect the elastic modulus and density of the material, thereby affecting the propagation speed of ultrasonic waves, and changes in the working environment temperature must be considered; working pressure: high-pressure environment will affect the structure and stress distribution of the equipment, thereby affecting ultrasonic detection; surface condition: whether the surface of the equipment is smooth or has scaling, corrosion, etc. These factors affect the reflection and transmission characteristics of ultrasonic waves.

[0035] It should be noted that in existing mature technologies, the relationship analysis between equipment data and ultrasonic frequency usually uses multivariate regression analysis, machine learning algorithms or statistical analysis methods to establish models. By analyzing the impact of factors such as equipment material, wall thickness, temperature, working pressure and surface condition on ultrasonic frequency, the constraints of various equipment parameters on frequency selection can be obtained. For example, the sound velocity characteristics of the material, the influence of wall thickness, the impact of operating temperature and pressure changes on sound wave propagation, etc. are all factors that need to be considered in the diagnostic model. The data analysis process usually includes the following steps:

[0036] Data collection and organization: Use sensors to obtain real-time data on the working status of the equipment, including material, wall thickness, temperature, pressure, and surface condition; Feature engineering and variable selection: Analyze how these equipment data affect the ultrasonic frequency based on historical failure data and equipment properties. For example, thicker equipment wall thickness may require a lower frequency to ensure sufficient penetration depth, while high temperature or high pressure may require corresponding adjustments to the frequency to ensure measurement accuracy; Modeling and regression analysis: Use regression models, support vector machines, neural networks and other machine learning algorithms to explore the relationship between various factors and ultrasonic frequency. These models can output a suitable frequency range based on different states of the equipment. Optimization and constraints: Adjust the results through optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.) to obtain specific constraints on the ultrasonic frequency of the equipment data, thereby providing a theoretical basis for the subsequent frequency range setting.

[0037] This analysis method has been widely used in fault diagnosis and health monitoring systems of various types of equipment, ensuring that the ultrasonic frequency can be adjusted in a targeted manner under various working environments, thereby improving diagnostic accuracy and efficiency. Therefore, these analysis steps are already mature and effective technologies in ultrasonic testing of industrial equipment and will not be repeated.

[0038] S2, performing random number adjustments on the initial ultrasonic frequency several times to obtain several test ultrasonic frequencies within the ultrasonic frequency range.

[0039] The initial ultrasonic frequency is adjusted by random numbers several times to obtain several test ultrasonic frequencies within the ultrasonic frequency range, specifically:

[0040] Generate a number of random numbers by computer;

[0041] The initial ultrasonic frequency is uniformly and randomly distributed according to each random number to generate a number of test ultrasonic frequencies;

[0042] The test ultrasonic frequencies outside the ultrasonic frequency range are eliminated.

[0043] S3, several ultrasonic waves of test ultrasonic frequencies are respectively transmitted to the components to be diagnosed of the pasteurizer and the sterilization temperature control equipment, and the intensity and spectrum of the corresponding signal echoes are analyzed to perform fault diagnosis.

[0044] The intensity and spectrum of the corresponding signal echo are analyzed to perform fault diagnosis, specifically:

[0045] Fit the echo intensity curve based on the least square method, calculate the deviation from the preset standard echo intensity curve, and classify the possibility of faults according to the deviation. The possibility classification includes weak faults and serious faults.

[0046] The echo signal is subjected to Fourier transform to obtain a signal spectrum in the frequency domain, and spectrum features are extracted from the signal spectrum to perform spectrum anomaly classification, which includes spectrum narrowing and spectrum shift.

[0047] It should be noted that if the echo intensity slightly deviates from the normal value, it is judged as a weak fault, including slight scaling or reduced heat exchange efficiency; if the echo intensity drops significantly, it is judged as a serious fault, including serious blockage, cracks or other damage. It should be noted that spectrum narrowing: If the echo spectrum width is significantly reduced, it may be due to scaling or local temperature / pressure abnormalities that cause reduced heat exchange efficiency; spectrum shift: If the spectrum peak shifts significantly, it may be due to damage to equipment components or cracks, bubbles and other faults.

[0048] The above judgment of what constitutes a substantial drop, a minor deviation, a significant shift, and a significant reduction in width is based on experience with other existing technologies, and there are already a variety of mature methods in the existing technology, which will not be repeated here.

[0049] S4, respectively obtaining diagnostic accuracy data and diagnostic efficiency data of fault diagnosis under different test ultrasonic frequencies.

[0050] The diagnostic accuracy data includes the echo resolution influence coefficient and the penetration fluctuation abnormality coefficient; the diagnostic efficiency data includes the detection efficiency influence coefficient.

[0051] The echo resolution influence coefficient is used to evaluate the ability of different ultrasonic frequencies to resolve the details of echo signals in fault diagnosis. High-frequency ultrasound has a higher resolution and can clearly detect small defects such as tiny cracks, scaling, corrosion, etc. on the surface or shallow layer of the material, so it is suitable for detecting small-scale faults on the surface or near the surface. However, high-frequency ultrasound has poor penetration and it is difficult to deeply detect thick-walled structures or deep defects. In contrast, low-frequency ultrasound has stronger penetration and can detect deep scaling, blockage or other large-area faults, but its resolution is lower and may not be able to clearly identify small-scale cracks or minor corrosion problems. Therefore, the echo resolution influence coefficient plays a role in balancing the contradiction between resolution and penetration in fault diagnosis, and can help select the appropriate frequency to improve the accuracy and efficiency of diagnosis.

[0052] Analyzing the echo resolution influence coefficient for evaluating the effect of the diagnostic method can solve the problem that the existing ultrasonic frequency setting is unclear, resulting in the inability to find a balance between penetration depth and resolution, which in turn affects the diagnostic effect of pasteurizer and sterilization temperature control equipment failures. It has the following advantages:

[0053] Optimize frequency selection: By introducing the echo resolution influence coefficient, the effect of different ultrasonic frequencies in actual diagnosis can be accurately evaluated, so as to effectively select the appropriate frequency. This helps to find the best balance between resolution and penetration depth, avoid relying solely on experience or simple rules for frequency selection, and improve the accuracy of diagnosis.

[0054] Improve fault detection accuracy: The echo resolution influence factor can help identify and distinguish small defects such as micro cracks, slight corrosion or scaling on the equipment surface, especially for complex equipment such as pasteurizers and sterilization temperature control equipment. Through the optimized use of high-resolution ultrasonic frequencies, it can ensure that small-scale faults can also be accurately detected.

[0055] Enhanced deep defect recognition capability: Low-frequency ultrasonic frequencies can penetrate thicker walls or pipe structures, and are more effective in detecting deep scaling, blockages or other large-scale faults. By analyzing the echo resolution influence coefficient, it can ensure that the advantage of low-frequency ultrasonic penetration depth is fully utilized, while avoiding missing subtle surface faults due to low resolution.

[0056] Improve diagnostic efficiency: By properly setting the frequency range and frequency selection strategy, excessive testing and unnecessary repeated diagnosis can be avoided, thereby improving the efficiency of the entire fault diagnosis process. This helps to quickly locate problems and make accurate diagnoses, reducing equipment downtime and maintenance costs.

[0057] Dynamic adjustment and optimization: The echo resolution influence coefficient provides a scientific basis for the dynamic adjustment of the ultrasonic frequency, and can adjust the frequency setting in real time according to different equipment conditions, working environment and fault type. This allows for more flexible frequency selection based on the characteristics of different faults, avoiding the negative impact of improper frequency setting on the diagnostic results.

[0058] Reduce misdiagnosis and missed diagnosis rates: By accurately evaluating the echo resolution influence coefficient at different frequencies, it is possible to effectively reduce misdiagnosis and missed diagnosis caused by frequency mismatch, especially in complex equipment fault diagnosis, to ensure the reliability and accuracy of the diagnostic results.

[0059] In summary, the introduction of the echo resolution influence coefficient can effectively solve the problem of poor diagnostic effect caused by unclear ultrasonic frequency selection in the prior art, and improve the accuracy, efficiency and reliability of diagnosis by optimizing the frequency setting.

[0060] The specific method for obtaining the echo resolution influence coefficient is as follows:

[0061] Extracting time amplitude data from ultrasonic signal echoes;

[0062] Preprocessing the ultrasonic echo and calculating the amplitude distribution characteristics of the ultrasonic echo, wherein the amplitude distribution characteristics include the echo mean and the echo variance;

[0063] The spectrum density of the ultrasonic echo is obtained based on the amplitude distribution characteristics, the high-frequency and low-frequency self-entropy of the ultrasonic echo is obtained according to the spectrum density, and the detail resolution entropy between the ultrasonic echoes is calculated based on the self-entropy of the ultrasonic echo;

[0064] The minimum comparison value of the echo mean and the echo standard deviation is obtained, and the minimum comparison value is added and combined with the detail resolution entropy to calculate the echo resolution influence coefficient.

[0065] The specific calculation formula of the self-entropy of the ultrasonic echo is as follows:

[0066] H(Pi )=-∑ x P i ×logP i ;

[0067] The specific calculation formula of the detail resolution entropy between ultrasonic echoes is as follows:

[0068] H(P i , P j )=-∑ x P i ×logP j ;

[0069] The specific calculation formula of the echo resolution influence coefficient is as follows:

[0070]

[0071] In the formula, H(P i ) is the ultrasonic echo P i The entropy of x is the ultrasonic echo value, H(P i , P j ) is the ultrasonic echo P i and P j The details of the parsing entropy, is the total number of 2 echoes selected from n echoes, α is the echo resolution influence coefficient, n is the total number of ultrasonic echoes, min(V ij ) is the minimum value of the difference between the echo mean and the echo variance, V ij is the sum of the difference between the echo mean and the echo variance.

[0072] The penetration fluctuation anomaly coefficient is used to evaluate the penetration ability of different ultrasonic frequencies in fault diagnosis, and reflects the difference in the performance of ultrasonic signals when penetrating different materials. High-frequency ultrasonic waves have weak penetration ability and are suitable for detecting faults in thin-walled materials or shallow structures, such as tiny cracks or surface corrosion. However, when used to detect thicker equipment structures (such as thick-walled pipes in heat exchangers), the signal may attenuate too much, resulting in a detection blind area, making it difficult to identify deep scaling or blockage problems. In contrast, low-frequency ultrasonic waves have stronger penetration ability and are suitable for the detection of thick-walled structures and deep defects, such as deep scaling or severe blockage, but their strong penetration ability may cause excessive signal penetration in thin-walled materials, making it difficult to focus on the target area and affecting the accuracy of diagnosis. This coefficient helps analyze and quantify the impact of frequency selection on penetration depth and signal quality, thereby optimizing the effectiveness of fault diagnosis methods.

[0073] Analyzing the abnormal coefficient of penetration fluctuation is helpful for evaluating the effect of diagnostic methods, thus solving the problem that the existing ultrasonic frequency setting is unclear, resulting in the inability to find a balance between penetration depth and resolution, which in turn affects the diagnostic effect of pasteurizer and sterilization temperature control equipment failures. It has the following advantages:

[0074] Clear frequency selection criteria: The penetration fluctuation anomaly coefficient can clearly quantify the penetration ability of different ultrasonic frequencies in different equipment materials and structures, so that the frequency selection is no longer based on experience, but on a data-driven scientific basis. This provides a clear standard for the setting of ultrasonic frequency and avoids differences in diagnostic effects caused by improper frequency setting.

[0075] Optimize equipment fault detection: By evaluating the abnormal coefficient of penetration fluctuation, the penetration effect of different frequencies on complex structure equipment such as pasteurizers and sterilization temperature control equipment can be accurately analyzed. When encountering problems such as thick-walled pipes or deep scaling, the penetration advantage of low-frequency ultrasound can help detect deep faults and avoid blind spots caused by insufficient penetration of high-frequency ultrasound.

[0076] Balancing penetration depth and resolution: The penetration fluctuation anomaly coefficient helps find the best balance between penetration depth and echo resolution. Low-frequency ultrasound can penetrate deeper layers, but may result in reduced resolution and difficulty in identifying tiny cracks or surface faults; high-frequency ultrasound provides higher resolution but weaker penetration. By analyzing the penetration fluctuation anomaly coefficient, the penetration depth and resolution can be optimized at different frequencies, thereby improving the overall effect of fault diagnosis.

[0077] Improve the scope of fault detection: Through reasonable frequency selection, it can ensure that faults from the surface of the equipment to deep faults can be accurately identified. Low-frequency ultrasound can detect deep scaling or blockage, while high-frequency ultrasound is suitable for the precise location of problems such as fine surface cracks and minor corrosion. The flexible application of this frequency significantly expands the detection range of fault diagnosis.

[0078] Improve diagnostic accuracy: The abnormal coefficient of penetration fluctuation can reduce the negative impact of frequency selection on diagnostic results, thereby improving overall diagnostic accuracy. Especially in equipment such as pasteurizers and sterilization temperature control equipment, the complex structure and working environment make the performance of ultrasound waves of different frequencies vary greatly when penetrating different materials. By quantifying these differences, the most suitable frequency can be accurately determined, thereby improving the accuracy of fault location.

[0079] Improve diagnostic efficiency: Optimized frequency selection not only improves diagnostic accuracy, but also improves diagnostic efficiency. It avoids excessive frequency tests and excessive repeated tests, reduces time costs and equipment downtime, and ensures that equipment failures can be diagnosed quickly and accurately, avoiding diagnostic delays caused by improper frequency settings.

[0080] Enhanced early warning capability for faults: The introduction of the abnormal coefficient of penetration fluctuation makes it possible to diagnose deep faults in equipment at an early stage, especially in pasteurizers and sterilization temperature control equipment, where deep scaling, blockage and other problems often require strong penetration for effective detection. By quantifying the impact of penetration, earlier fault detection and warning can be achieved, thus avoiding equipment damage or production interruption.

[0081] In summary, the penetration force fluctuation anomaly coefficient provides a scientific evaluation standard for ultrasonic frequency selection, solves the diagnostic effect problem caused by unclear frequency selection in the prior art, thereby optimizing the fault diagnosis process of pasteurizers and sterilization temperature control equipment, and improving the accuracy, efficiency and reliability of diagnosis.

[0082] The specific method for obtaining the abnormal coefficient of penetration fluctuation is as follows:

[0083] Real-time acquisition of ultrasonic signal penetration depth and signal attenuation rate corresponding to the same time interval within a preset time range when ultrasonic fault detection is performed, and numbering the ultrasonic signal penetration depth and signal attenuation rate;

[0084] The standard deviation and the average value of the penetration force are calculated by the penetration depth and the signal attenuation rate of the ultrasonic signal corresponding to the same time interval within the preset time range when the ultrasonic fault detection is performed;

[0085] The coefficient of variation of penetration force was calculated based on the standard deviation of penetration force and the mean penetration force;

[0086] The penetration force fluctuation anomaly coefficient is calculated based on the penetration force variation coefficient.

[0087] The specific calculation formula of the penetration standard deviation is as follows:

[0088]

[0089] The specific calculation formula of the average penetration force is as follows:

[0090]

[0091] The specific calculation formula of the abnormal coefficient of penetration fluctuation is as follows:

[0092]

[0093] Where M1 is the standard deviation of penetration, M2 is the average penetration, H is the total number of ultrasonic signal penetration depth and signal attenuation rate, h is the number of ultrasonic signal penetration depth and signal attenuation rate, O h is the penetration depth of ultrasonic signal, Y h is the signal attenuation rate, and β is the abnormal coefficient of penetration fluctuation.

[0094] The detection efficiency impact coefficient is used to measure the impact of different ultrasonic frequencies on the detection speed and range during the fault diagnosis process. High-frequency ultrasound can perform more detailed detection due to its higher resolution, but it has a slower detection speed due to the need to process more complex signals, and is suitable for local fine-grained detection. However, it is not suitable for rapid scanning of a large area and has low efficiency. In contrast, low-frequency ultrasound has a larger detection range and is suitable for rapid scanning of large areas, especially when performing preliminary screening, which can increase the detection speed. However, due to its lower resolution, it does not perform as well as high-frequency ultrasound in detail detection. Therefore, the detection efficiency impact coefficient helps quantify and evaluate the performance of ultrasound at different frequencies in terms of detection speed and range in order to optimize the efficiency of fault diagnosis.

[0095] Analyzing the detection efficiency influence coefficient for evaluating the effect of the diagnostic method, thus solving the problem that the existing ultrasonic frequency setting is unclear, resulting in the inability to find a balance between penetration depth and resolution, which in turn affects the diagnostic effect of pasteurizer and sterilization temperature control equipment failures. It has the following advantages:

[0096] Improve diagnostic efficiency: The detection efficiency coefficient can quantify the impact of different frequencies on diagnostic speed, helping to select the most suitable frequency for rapid troubleshooting. Low-frequency ultrasound has a larger detection range and faster scanning speed, which is suitable for rapid screening of large areas, thereby significantly improving the efficiency of fault detection.

[0097] Balance detection accuracy and speed: By considering the trade-off between high-frequency and low-frequency ultrasound in detection accuracy and speed, the detection efficiency impact coefficient helps find a reasonable balance between speed and accuracy. When a quick preliminary diagnosis is required, low-frequency ultrasound can be selected, while when performing refined detection, high-frequency ultrasound can be selected, thus taking into account both diagnostic speed and accuracy.

[0098] Reduce diagnostic time and cost: By optimizing frequency selection, the most appropriate frequency can be used in different stages (such as preliminary screening and detailed analysis), reducing the time consumption of multiple tests and frequency switching, and improving overall diagnostic efficiency. This not only reduces the time cost during the detection process, but also reduces the downtime of equipment and improves production efficiency.

[0099] Improve the comprehensiveness of fault detection: Low-frequency ultrasound can quickly scan a large area, suitable for quickly detecting potential problems of the entire equipment, while high-frequency ultrasound can accurately capture tiny defects. By combining the advantages of different frequencies, the detection efficiency impact coefficient ensures more comprehensive and faster fault detection of equipment, avoiding missing potential fault points due to improper frequency selection.

[0100] Improve equipment operation stability: By optimizing the selection and adjustment of ultrasonic frequency, the fault diagnosis speed of equipment such as pasteurizers and sterilization temperature control equipment can be improved, the risk of equipment failure due to failure to detect failure in time can be reduced, and the stable operation of equipment and the continuity of the production line can be further guaranteed.

[0101] Optimize fault diagnosis strategy: The detection efficiency impact coefficient not only provides a quantitative evaluation of the effect of different frequencies on the detection speed, but also provides data support for the frequency selection in actual fault diagnosis. By properly selecting the frequency, problems can be quickly screened in the early stages of diagnosis, while refined detection can be performed in key areas, thereby improving the diagnosis effect and ensuring faster and more accurate fault location.

[0102] Enhanced fault prediction and prevention capabilities: By analyzing the coefficients affecting detection efficiency, potential equipment failures can be effectively predicted and prevented. Low-frequency ultrasound is suitable for early large-scale screening, while high-frequency ultrasound is suitable for precise fault location, which enhances the early warning and intervention capabilities of equipment failures and avoids the serious consequences caused by missed detection or delayed diagnosis.

[0103] In summary, the detection efficiency influence coefficient not only optimizes the selection of ultrasonic frequency, but also effectively solves the problem of unclear setting of ultrasonic frequency in the prior art by increasing the diagnosis speed, reducing time cost, and improving the comprehensiveness and accuracy of fault detection, thereby significantly improving the overall effect of fault diagnosis of pasteurizers and sterilization temperature control equipment.

[0104] The specific method for obtaining the detection efficiency influence coefficient is as follows:

[0105] Based on the diagnostic equipment fault types of fault diagnosis at different ultrasonic frequencies in history, a diagnostic target set of fault diagnosis at different ultrasonic frequencies is obtained, wherein the diagnostic target set of fault diagnosis at different ultrasonic frequencies includes each equipment fault type of fault diagnosis and the diagnostic efficiency of each equipment fault type at different ultrasonic frequencies;

[0106] Assigning preset priority weights to various diagnostic equipment fault types within a diagnostic target set for fault diagnosis at different ultrasonic frequencies;

[0107] A regression model is constructed based on the diagnostic target set and corresponding priority weights of fault diagnosis under different ultrasonic frequencies to generate the detection efficiency influence coefficient.

[0108] The specific calculation formula of the detection efficiency influence coefficient is as follows:

[0109]

[0110] Where, χ is the detection efficiency influence coefficient, μ1…μ n are the preset priority weights corresponding to the diagnostic targets of fault diagnosis under different ultrasonic frequencies, ω1…ω n is the diagnostic efficiency of each equipment fault type, YC n is the type of the diagnostic target of the fault diagnosis under different ultrasonic frequencies, and n is the number of the diagnostic target of the fault diagnosis under different ultrasonic frequencies.

[0111] This embodiment improves diagnostic accuracy and efficiency by optimizing frequency selection. The steps include data analysis, random adjustment of ultrasonic frequency, signal echo analysis, and data collection of diagnostic accuracy and efficiency. In particular, the echo resolution influence coefficient and the penetration fluctuation abnormal coefficient are introduced to optimize the resolution and penetration ability of ultrasonic frequency in equipment fault detection respectively. Through this method, the frequency can be flexibly adjusted according to the different materials, structures and working environments of the equipment to improve the accuracy and efficiency of fault detection. This method can not only accurately identify different types of faults, but also effectively reduce the misdiagnosis rate and missed diagnosis rate.

[0112] Example 2, S5, constructing an ultrasonic frequency evaluation model based on the bp neural network according to the diagnostic effect data and the diagnostic efficiency data.

[0113] The ultrasonic frequency evaluation model is constructed based on the bp neural network according to the diagnostic effect data and the diagnostic efficiency data, specifically:

[0114] The acquired echo resolution influence coefficient, penetration fluctuation abnormal coefficient and detection efficiency influence coefficient are used to construct an ultrasonic frequency evaluation model to generate an ultrasonic frequency evaluation coefficient;

[0115] The specific calculation formula of the ultrasonic frequency evaluation coefficient is as follows:

[0116]

[0117] Wherein, CSB is the ultrasonic frequency evaluation coefficient, δ1 is the preset proportional coefficient of the echo resolution influence coefficient, δ2 is the preset proportional coefficient of the penetration fluctuation abnormal coefficient, δ3 is the preset proportional coefficient of the detection efficiency influence coefficient, α is the echo resolution influence coefficient, β is the penetration fluctuation abnormal coefficient, and χ is the detection efficiency influence coefficient.

[0118] S6, construct an ultrasonic frequency-diagnostic effect display model according to different test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model, and perform data analysis to obtain the screened ultrasonic frequency, which is applied to the fault diagnosis of dairy beverage machinery and equipment.

[0119] The ultrasonic frequency-diagnostic effect display model is constructed according to different test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model, specifically:

[0120] The test ultrasonic frequency is taken as the x-axis of the rectangular coordinate system, and the output of the ultrasonic frequency evaluation model corresponding to the fault diagnosis under the test ultrasonic frequency is taken as the y-axis of the rectangular coordinate system;

[0121] Mapping a number of test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model onto a rectangular coordinate system to obtain a number of data points;

[0122] Several data points are connected using smooth curves to construct an ultrasonic frequency-diagnostic effect display model.

[0123] The data analysis is performed to obtain the screened ultrasonic frequency, which is applied to the fault diagnosis of dairy beverage machinery and equipment, specifically:

[0124] Establish corresponding mathematical expressions based on the ultrasonic frequency-diagnostic effect display model;

[0125] The peak values ​​in the ultrasonic frequency-diagnostic effect display model are obtained, and the secondary derivatives of the peak values ​​are performed based on the mathematical expression. The test ultrasonic frequency corresponding to the peak value with the smallest secondary derivative result is used as the screened ultrasonic frequency and applied to the fault diagnosis of dairy beverage machinery and equipment.

[0126] This embodiment uses a BP neural network to construct an ultrasonic frequency evaluation model, uses diagnostic effect data and diagnostic efficiency data to evaluate the ultrasonic frequency, and optimizes the fault diagnosis process of dairy beverage machinery and equipment. Specifically, the ultrasonic frequency evaluation coefficient is generated through the echo resolution influence coefficient, the penetration fluctuation abnormal coefficient and the detection efficiency influence coefficient, and then the most suitable ultrasonic frequency for fault diagnosis is obtained. On this basis, an ultrasonic frequency-diagnostic effect display model is constructed, and its data is smoothed to further screen out effective ultrasonic frequencies. By taking the secondary derivative of the peak value, the optimal frequency is accurately identified, which improves the accuracy and efficiency of equipment fault diagnosis.

[0127] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0128] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0129] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0130] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0131] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0132] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for diagnosing dairy beverage machinery faults based on data analysis, characterized in that: The steps include: Obtain equipment data of dairy beverage machinery and equipment, and set the initial ultrasonic frequency and ultrasonic frequency range according to the equipment data; Performing random number adjustments on the initial ultrasonic frequency several times to obtain several test ultrasonic frequencies within the ultrasonic frequency range; Several ultrasonic waves of test ultrasonic frequencies are respectively transmitted to the components to be diagnosed of the pasteurizer and the sterilization temperature control equipment, and the intensity and spectrum of the corresponding signal echo are analyzed to perform fault diagnosis; Obtain the diagnostic accuracy data and diagnostic efficiency data of fault diagnosis under different test ultrasonic frequencies respectively; According to the diagnostic effect data and diagnostic efficiency data, an ultrasonic frequency evaluation model is constructed based on the BP neural network; According to different test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model, an ultrasonic frequency-diagnostic effect display model is constructed, and data analysis is performed to obtain the screened ultrasonic frequency, which is applied to the fault diagnosis of dairy beverage machinery and equipment.

2. The method for diagnosing dairy beverage machinery and equipment faults based on data analysis according to claim 1, characterized in that: The device data of the dairy beverage machine is obtained, and the initial ultrasonic frequency and ultrasonic frequency range are set according to the device data, specifically: The equipment data includes equipment material, wall thickness, temperature, working pressure and surface condition; Perform data analysis on the relationship between each device data and the ultrasonic frequency to obtain the constraint conditions of each device data on the ultrasonic frequency, and set the ultrasonic frequency range according to the constraint conditions of each device data on the ultrasonic frequency; Set the initial ultrasonic frequency based on historical data.

3. The method for diagnosing dairy beverage machinery and equipment faults based on data analysis according to claim 2, characterized in that: The initial ultrasonic frequency is adjusted by random numbers several times to obtain several test ultrasonic frequencies within the ultrasonic frequency range, specifically: Generate a number of random numbers by computer; The initial ultrasonic frequency is uniformly and randomly distributed according to each random number to generate a number of test ultrasonic frequencies; The test ultrasonic frequencies outside the ultrasonic frequency range are eliminated.

4. The method for diagnosing dairy beverage machinery equipment faults based on data analysis according to claim 3, characterized in that: The intensity and spectrum of the corresponding signal echo are analyzed to perform fault diagnosis, specifically: Fit the echo intensity curve based on the least square method, calculate the deviation from the preset standard echo intensity curve, and classify the possibility of faults according to the deviation. The possibility classification includes weak faults and serious faults. The echo signal is subjected to Fourier transformation to obtain a signal spectrum in the frequency domain, and spectrum features are extracted from the signal spectrum to perform spectrum anomaly classification, which includes spectrum narrowing and spectrum shift.

5. The method for diagnosing dairy beverage machinery equipment faults based on data analysis according to claim 4, characterized in that: The accurate diagnostic data includes the echo resolution influence coefficient and the penetration fluctuation abnormal coefficient; the specific method for obtaining the echo resolution influence coefficient is as follows: Extracting time amplitude data from ultrasonic signal echoes; Preprocessing the ultrasonic echo and calculating the amplitude distribution characteristics of the ultrasonic echo, wherein the amplitude distribution characteristics include the echo mean and the echo variance; The spectrum density of the ultrasonic echo is obtained based on the amplitude distribution characteristics, the high-frequency and low-frequency self-entropy of the ultrasonic echo is obtained according to the spectrum density, and the detail resolution entropy between the ultrasonic echoes is calculated based on the self-entropy of the ultrasonic echo; The minimum comparison value of the echo mean and the echo standard deviation is obtained, and the minimum comparison value is added and combined with the detail resolution entropy to calculate the echo resolution influence coefficient.

6. The method for diagnosing dairy beverage machinery equipment faults based on data analysis according to claim 5, characterized in that: The specific method for obtaining the abnormal coefficient of penetration fluctuation is as follows: Real-time acquisition of ultrasonic signal penetration depth and signal attenuation rate corresponding to the same time interval within a preset time range when ultrasonic fault detection is performed, and numbering the ultrasonic signal penetration depth and signal attenuation rate; The standard deviation and the average value of the penetration force are calculated by the penetration depth and the signal attenuation rate of the ultrasonic signal corresponding to the same time interval within the preset time range when the ultrasonic fault detection is performed; The coefficient of variation of penetration force was calculated based on the standard deviation of penetration force and the mean penetration force; The penetration force fluctuation anomaly coefficient is calculated based on the penetration force variation coefficient.

7. The method for diagnosing dairy beverage machinery faults based on data analysis according to claim 6, characterized in that: The diagnostic efficiency data includes a detection efficiency influence coefficient; the specific method for obtaining the detection efficiency influence coefficient is as follows: Based on the diagnostic equipment fault types of fault diagnosis at different ultrasonic frequencies in history, a diagnostic target set of fault diagnosis at different ultrasonic frequencies is obtained, wherein the diagnostic target set of fault diagnosis at different ultrasonic frequencies includes each equipment fault type of fault diagnosis and the diagnostic efficiency of each equipment fault type at different ultrasonic frequencies; Assigning preset priority weights to various diagnostic equipment fault types within a diagnostic target set for fault diagnosis at different ultrasonic frequencies; A regression model is constructed based on the diagnostic target set and corresponding priority weights of fault diagnosis under different ultrasonic frequencies to generate the detection efficiency influence coefficient.

8. The method for diagnosing dairy beverage machinery equipment faults based on data analysis according to claim 7, characterized in that: The ultrasonic frequency-diagnostic effect display model is constructed according to different test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model, specifically: The test ultrasonic frequency is taken as the x-axis of the rectangular coordinate system, and the output of the ultrasonic frequency evaluation model corresponding to the fault diagnosis under the test ultrasonic frequency is taken as the y-axis of the rectangular coordinate system; Mapping a number of test ultrasonic frequencies and the output of the corresponding ultrasonic frequency evaluation model onto a rectangular coordinate system to obtain a number of data points; Several data points are connected using smooth curves to construct an ultrasonic frequency-diagnostic effect display model.

9. The method for diagnosing dairy beverage machinery faults based on data analysis according to claim 8, characterized in that: The data analysis is performed to obtain the screened ultrasonic frequency, which is applied to the fault diagnosis of dairy beverage machinery and equipment, specifically: Establish corresponding mathematical expressions based on the ultrasonic frequency-diagnostic effect display model; The peak values ​​in the ultrasonic frequency-diagnostic effect display model are obtained, and the secondary derivatives of the peak values ​​are performed based on the mathematical expression. The test ultrasonic frequency corresponding to the peak value with the smallest secondary derivative result is used as the screened ultrasonic frequency and applied to the fault diagnosis of dairy beverage machinery and equipment.

10. The method for diagnosing dairy beverage machinery equipment faults based on data analysis according to claim 9, characterized in that: The specific calculation formula of the self-entropy of the ultrasonic echo is as follows: H(P i )=-Σ x P i ×logP i ; The specific calculation formula of the detail resolution entropy between ultrasonic echoes is as follows: H(P i ,P j )=-∑ x P i ×logP j ; The specific calculation formula of the echo resolution influence coefficient is as follows: In the formula, H(P i ) is the ultrasonic echo P i The entropy of x is the ultrasonic echo value, H(P i , P j ) is the ultrasonic echo P i and P j The details of the parsing entropy, is the total number of 2 echoes selected from n echoes, α is the echo resolution influence coefficient, n is the total number of ultrasonic echoes, min(V ij ) is the minimum value of the difference between the echo mean and the echo variance, V ij is the sum of the difference between the echo mean and the echo variance.

Images