Industrial equipment data collection monitoring method, platform and electronic device

Abstract

The present application relates to the technical field of data processing, and more particularly to an industrial equipment data acquisition and monitoring method, platform and electronic device. The method comprises: constructing a running characteristic spectrum benchmark library of the industrial equipment; collecting monitoring data of the industrial equipment in real time and constructing an equipment running characteristic spectrum; constructing a deviation analysis model based on the running characteristic spectrum benchmark library and the equipment running characteristic spectrum to determine a characteristic deviation degree and a characteristic weight; comprehensively analyzing the characteristic deviation degree and the characteristic weight to determine a comprehensive anomaly index and formulate a graded early warning decision. The present application proposes the concept of equipment running characteristic spectrum, and based on this, constructs a multi-scale, self-adaptive benchmark dynamic comparison and comprehensive deviation degree evaluation method, achieving high specificity and high accuracy identification of the abnormal state of the equipment.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, platform and electronic equipment for data acquisition and monitoring of industrial equipment. Background Technology

[0002] Industrial equipment is a core component of the manufacturing production system, and its operational status directly impacts production efficiency, product quality, and production safety. With the rapid development of the Industrial Internet and intelligent manufacturing technologies, equipment monitoring is gradually shifting from traditional manual inspections and periodic maintenance to data-driven, intelligent early warning, and predictive maintenance models. Currently, equipment monitoring systems are evolving towards multi-source data fusion, real-time analysis, and intelligent decision-making. However, how to achieve comprehensive perception, accurate assessment, and timely early warning of equipment status under complex operating conditions remains a key issue that urgently needs to be addressed in the field of intelligent operation and maintenance of industrial equipment. Against this backdrop, this invention aims to construct an efficient and reliable equipment data acquisition and monitoring system through multi-dimensional, multi-scale feature modeling and dynamic deviation analysis, providing technical support for the intelligent management of industrial equipment.

[0003] Existing industrial equipment monitoring technologies largely rely on single-dimensional or limited features for status assessment, making it difficult to comprehensively reflect the true state of equipment under complex operating conditions. Traditional methods lack multi-scale fusion and dynamic weighting mechanisms in feature extraction and anomaly detection, resulting in insensitivity to early, subtle anomalies, high false alarm rates, and poor adaptability. Furthermore, existing early warning methods are mostly fixed-threshold alarms, lacking hierarchical decision-making and continuous situation assessment, failing to provide progressive and actionable operation and maintenance guidance, and thus failing to meet the needs of intelligent and refined management of modern industrial equipment. Summary of the Invention

[0004] The purpose of this invention is to provide a method, platform, and electronic device for data acquisition and monitoring of industrial equipment, so as to solve at least one of the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for data acquisition and monitoring of industrial equipment, comprising:

[0007] Construct a benchmark library of operational characteristic spectra for industrial equipment;

[0008] Real-time acquisition of monitoring data from industrial equipment, and construction of equipment operation characteristic spectrum;

[0009] A deviation analysis model is constructed based on the benchmark library of operating characteristic spectra and the operating characteristic spectra of equipment to determine the degree of characteristic deviation and the characteristic weight;

[0010] By comprehensively analyzing the deviation of features and the weight of features, a comprehensive anomaly index is determined and a graded early warning decision is made.

[0011] Furthermore, multi-scale feature analysis was performed on the collected operational data. Short-time scale features were extracted from the operational data in the power and vibration dimensions. The time window length was set to 1 second, and the asymmetry index of the three-phase current within the time window was used as the short-time scale feature. The expression for the asymmetry index is: AS=(|μ(Ia)-μ(I)|+|μ(Ib)-μ(I)|+|μ(Ic)-μ(I)|) / (3×μ(I)), where AS represents the asymmetry index, Ia, Ib, and Ic represent the three-phase currents, μ() represents the average value of the data in parentheses, and I represents the sum of the three-phase currents, I=Ia+Ib+Ic. Vector synthesis was performed on the triaxial acceleration signals to obtain the triaxial acceleration parameters. The expression for the triaxial acceleration parameters is: AC=sqrt(Ax 2 +Ay 2 +Az 2 In the formula, AC represents the triaxial acceleration parameters, Ax, Ay, and Az represent the triaxial acceleration signals, and sqrt() represents the calculation of the square root of the data in parentheses. The envelope signal is obtained by performing Hilbert transform on the triaxial acceleration parameters, and then Fourier transform is performed on the envelope signal. The ratio of the mean amplitude of the first 5 significant spectral peaks to the amplitude at the revolving frequency is extracted as the short-time scale feature.

[0012] Mid-time scale features are extracted from the operational data in the electrical and thermodynamic dimensions. The time window length is set to 1 minute. The fluctuation stability index of the total active power within the time window is used as the mid-time scale feature. The expression for the fluctuation stability index is: FP=1 / (1+σ(P) / μ(P)), where FP represents the fluctuation stability index, P represents the total active power, and σ() represents the standard deviation of the data in parentheses. The temperature difference stability index of the winding temperature and the cooling medium outlet temperature within the time window is used as the mid-time scale feature. The expression for the temperature difference stability index is: TS=1-(σ(Tw-Tc) / μ(Tw-Tc)), where TS represents the temperature difference stability index, Tw represents the winding temperature, and Tc represents the cooling medium outlet temperature.

[0013] Long-term scale features are extracted from the acoustic dimension of the operating data. The time window length is set to 1 hour. The data within the time window is divided into frames of 1 second each. The Mel frequency cepstral coefficients of the device's operating sound signal in each frame are calculated. The coefficients of variation of the first 6 coefficients in all frames are taken to form a 6-dimensional temporal coefficient of variation vector, which is used as the long-term scale feature.

[0014] Furthermore, operational data of the equipment under different typical loads were collected and subjected to multi-scale feature analysis to obtain its short-term, medium-term, and long-term features. Gaussian mixture model (GMM) was used to fit the probability density distribution of the feature quantities in each scale feature. The distribution models of each feature quantity constituted the benchmark library of the operational feature spectrum of the industrial equipment.

[0015] Furthermore, multi-scale feature analysis is performed on the monitoring data to obtain a set of current feature values, in order to construct the equipment operation feature spectrum. The equipment operation feature spectrum is set as Fc={f1,f2,...,fn}, where n represents the total number of features.

[0016] Furthermore, the ratio of the average total active power in the monitoring data over one hour to the rated power of the equipment is used as the average load rate. The ratio of the average total active power in each group of data in the operating data to the rated power of the equipment is used as the benchmark average load rate. The distribution model of the characteristic quantity corresponding to the operating data that satisfies |average load rate of equipment - benchmark average load rate| < 0.05 is matched in the operating characteristic spectrum benchmark library. If the match is successful, the distribution model of the corresponding characteristic quantity is directly extracted as the matching benchmark. Otherwise, the matching benchmark is generated by interpolation.

[0017] Furthermore, for a feature quantity fi, when the corresponding matching benchmark follows an approximately Gaussian distribution, the historical mean μi and historical standard deviation σi corresponding to the feature quantity are extracted from the matching benchmark, and the feature deviation is analyzed. The expression for the feature deviation is set as: Di=|(fi-μi) / (k×σi)|, where Di represents the feature deviation and k represents the sensitivity adjustment coefficient.

[0018] For a feature quantity fi, when the corresponding matching benchmark follows a complex or multimodal distribution, its feature value is stored in the matching benchmark in the form of a Gaussian mixture model, and the feature deviation is analyzed. The expression for the feature deviation is set as: Di=-ln(P(fi|Mi)), where P(fi|Mi) represents the probability density of the current feature quantity under the corresponding Gaussian mixture model.

[0019] Furthermore, the feature weight is set to Wi. For short-term features in the electrical energy dimension, the initial value of the feature weight is set to 0.2. For short-term features in the vibration dimension, the initial value of the feature weight is set to 0.25. For medium-term features in the electrical energy dimension, the initial value of the feature weight is set to 0.2. For medium-term features in the thermodynamic dimension, the initial value of the feature weight is set to 0.15. For long-term features in the acoustic dimension, the average value of the feature deviation of the six coefficients is calculated, and the initial value of the feature weight is set to 0.2.

[0020] If the feature deviation of fi has consistently met the condition Di > 0.7 for the past 30 minutes, its corresponding feature weight is temporarily increased, so that the increased feature weight is W1i = (min(Wi × (1 + β), 0.4). At the same time, the other feature weights are decreased proportionally to the feature weights, so that the sum of the feature weights is 1. β represents the adjustment coefficient.

[0021] Furthermore, the feature deviation is smoothed and normalized using a sigmoid function. The normalized feature deviation is set as Norm(Di), where Norm(Di) = 1 / (1 + exp(-s × (Di - d0))), d0 represents the offset parameter, and s represents the slope parameter. The normalized feature deviation is then weighted and fused based on the feature weights to obtain a comprehensive anomaly index.

[0022] Set the sliding time window size to 10 minutes. Analyze one comprehensive anomaly index every minute within the sliding time window, calculate the average of all comprehensive anomaly indices within the window, and use this as the sliding index mean. Count the number of sample points whose comprehensive anomaly index exceeds the intermediate threshold within the sliding time window, divide this number by the total number of samples within the window, and use this as the proportion of persistently high values.

[0023] A tiered early warning decision is constructed based on a comprehensive anomaly index. When Ca < Thin and Rh < 0.1, the system is considered to be operating normally and there is no alarm. When Thin ≤ Ca < Thmid or Rh ≥ 0.1, an early warning is issued. When Ca ≥ Thmid and Rh ≥ 0.3, the system is considered to be in an abnormal state and a medium-level alarm is issued, and the three characteristic values ​​with the highest deviation are listed to the user. When Ca ≥ Thhigh and Rh ≥ 0.5 and the duration is greater than 3 minutes, the system is considered to be confirmed to be abnormal and a fault alarm is triggered. Ca represents the moving average, Rh represents the proportion of sustained high values, Thin represents the low-level threshold, Thmid represents the medium-level threshold, and Thhigh represents the high-level threshold.

[0024] On the other hand, the present invention also provides an industrial equipment data acquisition and monitoring platform, comprising:

[0025] The benchmark construction module is used to build a benchmark library of operating characteristic spectra for industrial equipment.

[0026] The feature extraction module is used to collect monitoring data of industrial equipment in real time and construct the equipment operation feature spectrum;

[0027] The deviation analysis module is used to build a deviation analysis model based on the operating characteristic spectrum benchmark library and the equipment operating characteristic spectrum in order to determine the characteristic deviation degree and characteristic weight;

[0028] The graded early warning module is used to comprehensively analyze feature deviation and feature weight to determine the comprehensive anomaly index and formulate graded early warning decisions.

[0029] On the other hand, the present invention also provides an electronic device, characterized in that the electronic device comprises:

[0030] One or more processors;

[0031] Storage device for storing one or more programs;

[0032] When the one or more programs are executed by the one or more processors, the one or more processors implement the industrial equipment data acquisition and monitoring method as described above.

[0033] The beneficial effects of this invention are as follows: By collecting and modeling multi-dimensional and multi-scale features, a benchmark library of equipment operation feature spectra is constructed. Combined with real-time data matching and deviation analysis models, refined monitoring and intelligent early warning of equipment status are realized. By integrating multi-source data such as electrical energy, vibration, acoustics, and thermodynamics, the operating status of equipment is comprehensively covered, improving the systematicness and completeness of monitoring. Multi-scale feature analysis and Gaussian mixture modeling are adopted to enhance the robustness and adaptability of status representation. Through dynamic feature weights and hierarchical early warning mechanisms, early detection and accurate location of anomalies are achieved, significantly improving the initiative and intelligence level of equipment operation and maintenance. This solution has good scalability and engineering applicability and can be widely applied to the health management and predictive maintenance of various industrial equipment. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart of the industrial equipment data acquisition and monitoring method in this embodiment.

[0036] Figure 2 This is a flowchart of the method for constructing the deviation analysis model in this embodiment.

[0037] Figure 3 This is a schematic diagram of the industrial equipment data acquisition and monitoring platform in this embodiment.

[0038] Figure 4 This is a schematic diagram of the electronic device in this embodiment. Detailed Implementation

[0039] The industrial equipment data acquisition and monitoring method, platform, and electronic device disclosed in this invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered isolated; they can be combined with each other to achieve better technical effects. In the accompanying drawings of the following embodiments, the same reference numerals appearing in each drawing represent the same features or components, which can be applied to different embodiments. Therefore, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0040] It should be noted that the structures, proportions, sizes, etc., illustrated in the accompanying drawings are merely for illustrative purposes and to aid those skilled in the art in understanding and reading the invention. They are not intended to limit the conditions under which the invention can be implemented. Any modifications to the structure, changes in proportions, or adjustments to size, provided they do not affect the effectiveness or purpose of the invention, should fall within the scope of the technical content disclosed in the invention. The scope of the preferred embodiments of the present invention includes other implementations, wherein functions may be performed not in the order stated or discussed, including substantially simultaneously or in reverse order, depending on the functions involved. This should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0041] Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0042] In the description of the embodiments of this application, " / " means "or", and "and / or" is used to describe the relationship between related objects, indicating that there can be three relationships. For example, "A and / or B" means: A and B exist alone, B exists alone, and A and B exist simultaneously. In the description of the embodiments of this application, "multiple" refers to two or more embodiments.

[0043] Please see Figure 1 As shown, this is the industrial equipment data acquisition and monitoring method of this embodiment, including:

[0044] Step S1: Construct a benchmark library of operating characteristic spectra for industrial equipment.

[0045] Specifically, in step S1 of this embodiment, during the normal operation phase when the equipment is confirmed to be fault-free, its operating data is collected synchronously. The operating data includes data in multiple dimensions, including electrical energy, vibration, acoustics, and thermodynamics. The operating data in the electrical energy dimension includes three-phase current and total active power, and its acquisition frequency is set to 100Hz. The operating data in the vibration dimension is the vibration acceleration signal, which is collected by vibration sensors installed at key bearing seats or housings of the equipment, and the acquisition frequency is 10kHz. The operating data in the acoustic dimension is the operating sound signal of the equipment, which is collected by acoustic sensors arranged in the near field of the equipment, and the acquisition frequency is 20kHz. The operating data in the thermodynamic dimension includes winding temperature and cooling medium outlet temperature, which are temperature data of key parts of the equipment collected, and the acquisition frequency is 1Hz.

[0046] Specifically, in step S1 of this embodiment, multi-scale feature analysis is performed on the collected operational data. Short-time scale features are extracted from the operational data in the power and vibration dimensions. The time window length is set to 1 second, and the asymmetry index of the three-phase current within the time window is used as the short-time scale feature. The expression for the asymmetry index is: AS=(|μ(Ia)-μ(I)|+|μ(Ib)-μ(I)|+|μ(Ic)-μ(I)|) / (3×μ(I)), where AS represents the asymmetry index, Ia, Ib, and Ic represent the three-phase currents, μ() represents the average value of the data in parentheses, and I represents the sum of the three-phase currents, I=Ia+Ib+Ic; the triaxial acceleration signal is vector synthesized to obtain the triaxial acceleration parameters. The expression for the triaxial acceleration parameters is: AC=sqrt(Ax 2 +Ay 2 +Az 2 In the formula, AC represents the triaxial acceleration parameters, Ax, Ay, and Az represent the triaxial acceleration signals, and sqrt() represents the calculation of the square root of the data in parentheses. The envelope signal is obtained by performing Hilbert transform on the triaxial acceleration parameters, and then Fourier transform is performed on the envelope signal. The ratio of the mean amplitude of the first 5 significant spectral peaks to the amplitude at the revolving frequency is extracted as the short-time scale feature.

[0047] Mid-time scale features are extracted from the operational data in the electrical and thermodynamic dimensions. The time window length is set to 1 minute. The fluctuation stability index of the total active power within the time window is used as the mid-time scale feature. The expression for the fluctuation stability index is: FP=1 / (1+σ(P) / μ(P)), where FP represents the fluctuation stability index, P represents the total active power, and σ() represents the standard deviation of the data in parentheses. The temperature difference stability index of the winding temperature and the cooling medium outlet temperature within the time window is used as the mid-time scale feature. The expression for the temperature difference stability index is: TS=1-(σ(Tw-Tc) / μ(Tw-Tc)), where TS represents the temperature difference stability index, Tw represents the winding temperature, and Tc represents the cooling medium outlet temperature.

[0048] Long-term scale features are extracted from the acoustic dimension of the operating data. The time window length is set to 1 hour. The data within the time window is divided into frames of 1 second each. The Mel frequency cepstral coefficients of the device's operating sound signal in each frame are calculated. The coefficients of variation of the first 6 coefficients in all frames are taken to form a 6-dimensional temporal coefficient of variation vector, which is used as the long-term scale feature.

[0049] Specifically, in step S1 of this embodiment, operational data of the equipment under different typical loads is collected and subjected to multi-scale feature analysis to obtain its short-term, medium-term, and long-term features. Gaussian mixture models (GMMs) are used to fit the probability density distribution of the feature quantities at each scale. The distribution models of each feature quantity constitute the operational feature spectrum benchmark library of the industrial equipment. The typical load refers to 30%, 60%, and 90% of the rated load, and different data in the operational feature spectrum benchmark library are correlated with its load rate.

[0050] Specifically, in step S1 of this embodiment, normal operation data of the equipment is collected from multiple dimensions, and short-term, medium-term, and long-term features are extracted. A Gaussian mixture model is then used to construct a feature spectrum benchmark library. This method can comprehensively capture the operating status characteristics of the equipment under typical loads, providing a reliable benchmark for subsequent anomaly detection. Its beneficial effect lies in improving the comprehensiveness and stability of the state representation through multi-scale feature fusion, avoiding the limitations of single features being easily affected by operating condition fluctuations, and providing a solid data foundation for equipment health assessment.

[0051] Please continue reading. Figure 1 As shown, the industrial equipment data acquisition and monitoring method further includes:

[0052] Step S2: Real-time acquisition of monitoring data from industrial equipment and construction of equipment operation characteristic spectrum. The monitoring data is the operating data of the current time period acquired synchronously with the same dimensions, frequency and window scale as in step S1 above.

[0053] Specifically, in step S2 of this embodiment, multi-scale feature analysis is performed on the monitoring data to obtain a set of current feature values ​​to construct the equipment operation feature spectrum. The equipment operation feature spectrum is set as Fc={f1,f2,...,fn}, where n represents the total number of features. Each feature in the equipment operation feature spectrum corresponds to the feature in step S1 above, such as f1 representing the asymmetry index, f2 representing the ratio of the mean amplitude of the significant spectral peak to the amplitude at the frequency transition, etc.

[0054] Specifically, in step S2 of this embodiment, monitoring data is collected in real time using the same dimensions, frequency, and window scale as the benchmark library, and corresponding feature quantities are extracted to construct the equipment operation feature spectrum. This method ensures the consistency between real-time data and benchmark data at the feature level, facilitating subsequent comparative analysis. Its beneficial effects lie in realizing a dynamic and structured expression of equipment operating status, providing standardized feature inputs for real-time monitoring and trend analysis, and enhancing the system's sensitivity and responsiveness to changes in equipment status.

[0055] Please continue reading. Figure 1 As shown, the industrial equipment data acquisition and monitoring method further includes:

[0056] Step S3: Construct a deviation analysis model based on the operating feature spectrum benchmark library and the equipment operating feature spectrum to determine the feature deviation degree and feature weight.

[0057] Please see Figure 2 As shown, this is a method for constructing a deviation analysis model, including:

[0058] Step S31: Based on the monitoring data, match the distribution model of the feature quantity in the feature spectrum benchmark library to obtain the matching benchmark.

[0059] Specifically, in step S31 of this embodiment, the ratio of the average total active power in the monitoring data within 1 hour to the rated power of the equipment is used as the average load rate. The ratio of the average total active power in each group of data in the operating data to the rated power of the equipment is used as the benchmark average load rate. The distribution model of the characteristic quantity corresponding to the operating data that satisfies |average load rate of equipment - benchmark average load rate| < 0.05 is matched in the operating characteristic spectrum benchmark library. If the match is successful, the distribution model of the corresponding characteristic quantity is directly extracted as the matching benchmark. Otherwise, the matching benchmark is generated by interpolation.

[0060] Specifically, in step S31 of this embodiment, when generating a matching benchmark by interpolation, it is assumed that the benchmark library contains Llow and Lhigh, and satisfies the distribution model of Llow < L < Lhigh. For each feature fi, the parameters of its benchmark distribution will be obtained by linear interpolation. For example, the generated mean matching benchmark is μi(L) = μ(Llow) + [ (μi(Lhigh)-μi(Llow)) / (Lhigh-Llow)]×(L-Llow), where L represents the average monitoring load rate.

[0061] Please continue reading. Figure 2 As shown, the method for constructing the deviation analysis model further includes:

[0062] Step S32: Analyze the feature deviation based on the matching benchmark and the equipment operation feature spectrum.

[0063] Specifically, in step S32 of this embodiment, when the matching benchmark for the feature quantity fi follows an approximately Gaussian distribution, the historical mean μi and historical standard deviation σi corresponding to the feature quantity are extracted from the matching benchmark, and the feature deviation is analyzed. The expression for the feature deviation is set as: Di=|(fi-μi) / (k×σi)|, where Di represents the feature deviation, k represents the sensitivity adjustment coefficient, and i is the number that distinguishes different feature quantities in the equipment operation feature spectrum, i∈N. + And i≤n, the sensitivity adjustment coefficient is based on statistical content, k is 3 based on the statistical 3σ principle, that is, it is assumed that about 99.73% of the normal data falls within the (μ-3σ,μ+3σ) range, Di≥1 indicates that the current value has exceeded 1 / 3 of the normal fluctuation range and has begun to attract attention, k can be adjusted according to the criticality of the equipment, for particularly critical equipment, k can be tightened to k=2 to improve sensitivity.

[0064] Specifically, in step S32 of this embodiment, when the matching benchmark corresponding to the feature quantity fi follows a complex or multimodal distribution, its feature value is stored in the matching benchmark in the form of a Gaussian mixture model, and the feature deviation is analyzed. The expression for the feature deviation is set as: Di=-ln(P(fi|Mi)), where P(fi|Mi) represents the probability density of the current feature quantity under the corresponding Gaussian mixture model.

[0065] Please continue reading. Figure 2 As shown, the method for constructing the deviation analysis model further includes:

[0066] Step S33: Assign feature weights to each feature quantity in the equipment operation feature spectrum.

[0067] Specifically, in step S33 of this embodiment, the feature weight is set to Wi. For short-timescale features in the electrical energy dimension, the initial value of the feature weight is set to 0.2; for short-timescale features in the vibration dimension, the initial value of the feature weight is set to 0.25; for medium-timescale features in the electrical energy dimension, the initial value of the feature weight is set to 0.2; for medium-timescale features in the thermodynamic dimension, the initial value of the feature weight is set to 0.15; and for long-timescale features in the acoustic dimension, the average value of the feature deviation of the six coefficients is calculated, and its feature weight is assigned an initial value of 0.2. The sum of the feature weights of all feature quantities should be equal to 1.

[0068] Specifically, in step S33 of this embodiment, when the feature deviation of fi continuously satisfies Di>0.7 for the past 30 minutes, the corresponding feature weight is temporarily increased so that the increased feature weight is W1i=(min(Wi×(1+β),0.4), while other feature weights are decreased according to the feature weight ratio so that the sum of the feature weights is 1. β represents the adjustment coefficient, and in this embodiment, β=0.3 is set.

[0069] Specifically, in step S3 of this embodiment, the deviation of each feature is calculated by matching the feature distribution model in the benchmark library, and dynamic weights are assigned according to the importance of the features. This method combines statistical distribution with real-time operating conditions to achieve quantitative assessment of feature anomalies. Its beneficial effect is that it can identify the contribution degree of different features in anomalies, and enhance the sensitivity to key anomaly features through a dynamic weight adjustment mechanism, thereby improving the accuracy and targeting of anomaly identification and avoiding false alarms and missed alarms.

[0070] Please continue reading. Figure 1 As shown, the industrial equipment data acquisition and monitoring method further includes:

[0071] Step S4: Comprehensively analyze the feature deviation and feature weight to determine the comprehensive anomaly index and formulate a graded early warning decision.

[0072] Specifically, in step S4 of this embodiment, the feature deviation is smoothed and normalized using a sigmoid function. The normalized feature deviation is set as Norm(Di), where Norm(Di) = 1 / (1 + exp(-s × (Di - d0))), d0 represents the offset parameter, and s represents the slope parameter. The normalized feature deviation is then weighted and fused based on the feature weights to obtain a comprehensive anomaly index. The expression for the comprehensive anomaly index is: In the formula, C(t) represents the comprehensive anomaly index at time t. In this embodiment, the offset parameter is set to 0.5, which means that when Di=0.5, Norm(Di) is about 0.5, which is in the middle state. The slope parameter is set to 5, which controls the steepness of the curve, so that when Di is in the normal range, Norm(Di) is close to 0; when Di is significantly abnormal, Norm(Di) rapidly approaches 1.

[0073] Specifically, in step S4 of this embodiment, the sliding time window is set to 10 minutes. Within the sliding time window, one comprehensive anomaly index is analyzed every minute, and the average of all comprehensive anomaly indices within the window is calculated and used as the sliding index mean. The number of sample points whose comprehensive anomaly index exceeds the intermediate threshold within the sliding time window is counted, divided by the total number of samples within the window, and this number is used as the proportion of persistently high values. The size of the sliding time window should be greater than the shortest working cycle of the equipment and should cover a brief fault development process. Within this window, one comprehensive anomaly index can also be analyzed every second to increase the analysis accuracy.

[0074] Specifically, in step S4 of this embodiment, a hierarchical early warning decision is constructed based on the comprehensive anomaly index. When Ca < Thin and Rh < 0.1, the system is determined to be operating normally without alarms. When Thin ≤ Ca < Thmid or Rh ≥ 0.1, an early warning is issued. When Ca ≥ Thmid and Rh ≥ 0.3, the system is determined to be in an abnormal state, and a medium-level alarm is issued, listing the three feature values ​​with the highest deviation from the target value to the user. When Ca ≥ Thhigh and Rh ≥ 0.5 and the duration is greater than 3 minutes, the system is determined to be confirmed to be abnormal, and a fault alarm is triggered. Ca represents the moving average, Rh represents the proportion of sustained high values, Thin represents the low-level threshold, Thmid represents the medium-level threshold, and Thhigh represents the high-level threshold. In this embodiment, the low-level threshold is set to 0.2, the medium-level threshold is set to 0.4, and the high-level threshold is set to 0.65. The low-level threshold should be set to [0.15, 0.25], the medium-level threshold should be set to [0.35, 0.45], and the high-level threshold should be set to [0.6, 0.7]. The above-mentioned graded early warning decision is made by judging one by one in descending order of high level, triggering the corresponding higher-level early warning. The order of early warning levels from low to high is: no alarm, early warning, medium alarm, and fault alarm.

[0075] Specifically, in step S4 of this embodiment, a comprehensive anomaly index is calculated by normalizing and weighting the feature deviation, and then combined with sliding window statistics and threshold judgment to achieve hierarchical early warning. This method forms a closed-loop decision-making process from multi-feature fusion to overall state assessment. Its beneficial effect lies in the fact that by combining multi-level thresholds with the proportion of persistently high values, a progressive judgment from early warning to fault confirmation is achieved, improving the timeliness and reliability of early warning and providing users with clear and operable decision support.

[0076] Please see Figure 3 As shown, this is the industrial equipment data acquisition and monitoring platform of this embodiment, including:

[0077] The benchmark construction module is used to build a benchmark library of operating characteristic spectra for industrial equipment.

[0078] The feature extraction module is used to collect monitoring data of industrial equipment in real time and construct the equipment operation feature spectrum;

[0079] The deviation analysis module is used to build a deviation analysis model based on the operating characteristic spectrum benchmark library and the equipment operating characteristic spectrum in order to determine the characteristic deviation degree and characteristic weight;

[0080] The graded early warning module is used to comprehensively analyze feature deviation and feature weight to determine the comprehensive anomaly index and formulate graded early warning decisions.

[0081] Please see Figure 4 As shown, it is a structural schematic diagram of an electronic device in this embodiment. The electronic device 60 in this embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), wearable electronic devices, etc., as well as fixed terminals such as digital TVs, desktop computers, smart home devices, etc. Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0082] like Figure 4As shown, the electronic device 60 may include a processing unit 61, which can perform various appropriate actions and processes to implement the methods of the embodiments described herein, based on a program stored in ROM 62 or a program loaded from storage device 68 into RAM 63. RAM 63 also stores various programs and data required for the operation of the electronic device 60. The processing unit 61, ROM 62, and RAM 63 are interconnected via bus 64. I / O interface 65 is also connected to bus 64. Typically, the following devices can be connected to I / O interface 65: input devices 66 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 67 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 68 including, for example, magnetic tapes, hard disks, etc.; and communication devices 69. Communication device 69 allows the electronic device 60 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 An electronic device 60 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0083] Specifically, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts, thereby implementing the methods as described above. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 69, or installed from a storage device 68, or installed from a ROM 62. When the computer program is executed by the processing device 61, it performs the functions defined in the methods of the embodiments of the present invention.

[0084] Specifically, the computer-readable medium described in this embodiment may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0085] Specifically, in this embodiment, the computer-readable medium carries one or more programs. When the electronic device executes one or more of these programs, the electronic device causes the following: to construct a benchmark library of operating characteristic spectra for industrial equipment; to collect monitoring data of industrial equipment in real time and construct an operating characteristic spectrum for the equipment; to construct a deviation analysis model based on the benchmark library of operating characteristic spectra and the operating characteristic spectrum for the equipment to determine the degree of characteristic deviation and the weight of the characteristics; and to comprehensively analyze the degree of characteristic deviation and the weight of the characteristics to determine a comprehensive anomaly index and formulate a graded early warning decision.

[0086] In the above description, the disclosure of this invention is not intended to limit itself to these aspects. Rather, within the scope of the objectives of this disclosure, components can be selectively and operationally combined in any number. Furthermore, terms such as “comprising,” “encompassing,” and “having” should be interpreted by default as inclusive or open-ended, rather than exclusive or closed, unless explicitly defined as such. All technical, scientific, or other terms are to be understood by those skilled in the art, unless defined as such. Public terms found in dictionaries should not be interpreted in the context of the relevant technical documents in an overly idealistic or impractical manner, unless explicitly defined as such in this disclosure. Any modifications or alterations made by those skilled in the art based on the foregoing disclosure are within the scope of the claims.

Claims

1. A method for data acquisition and monitoring of industrial equipment, characterized in that, include: Construct a benchmark library of operational characteristic spectra for industrial equipment; Real-time acquisition of monitoring data from industrial equipment, and construction of equipment operation characteristic spectrum; A deviation analysis model is constructed based on the benchmark library of operating characteristic spectra and the operating characteristic spectra of equipment to determine the degree of characteristic deviation and the characteristic weight; By comprehensively analyzing the deviation of features and the weight of features, a comprehensive anomaly index is determined and a graded early warning decision is made.

2. The industrial equipment data acquisition and monitoring method according to claim 1, characterized in that, Multi-scale feature analysis was performed on the collected operational data. Short-time scale features were extracted from the operational data in the power and vibration dimensions. The time window length was set to 1 second, and the asymmetry index of the three-phase current within the time window was used as the short-time scale feature. The expression for the asymmetry index is: AS=(|μ(Ia)-μ(I)|+|μ(Ib)-μ(I)|+|μ(Ic)-μ(I)|) / (3×μ(I)), where AS represents the asymmetry index, Ia, Ib, and Ic represent the three-phase currents, μ() represents the average value of the data in parentheses, and I represents the sum of the three-phase currents, I=Ia+Ib+Ic. Vector synthesis was performed on the triaxial acceleration signals to obtain the triaxial acceleration parameters. The expression for the triaxial acceleration parameters is: AC=sqrt(Ax 2 +Ay 2 +Az 2 In the formula, AC represents the triaxial acceleration parameters, Ax, Ay, and Az represent the triaxial acceleration signals, and sqrt() represents the calculation of the square root of the data in parentheses. The envelope signal is obtained by performing Hilbert transform on the triaxial acceleration parameters, and then Fourier transform is performed on the envelope signal. The ratio of the mean amplitude of the first 5 significant spectral peaks to the amplitude at the revolving frequency is extracted as the short-time scale feature. Mid-time scale features are extracted from the operational data in the electrical and thermodynamic dimensions. The time window length is set to 1 minute. The fluctuation stability index of the total active power within the time window is used as the mid-time scale feature. The expression for the fluctuation stability index is: FP=1 / (1+σ(P) / μ(P)), where FP represents the fluctuation stability index, P represents the total active power, and σ() represents the standard deviation of the data in parentheses. The temperature difference stability index of the winding temperature and the cooling medium outlet temperature within the time window is used as the mid-time scale feature. The expression for the temperature difference stability index is: TS=1-(σ(Tw-Tc) / μ(Tw-Tc)), where TS represents the temperature difference stability index, Tw represents the winding temperature, and Tc represents the cooling medium outlet temperature. Long-term scale features are extracted from the acoustic dimension of the operating data. The time window length is set to 1 hour. The data within the time window is divided into frames of 1 second each. The Mel frequency cepstral coefficients of the device's operating sound signal in each frame are calculated. The coefficients of variation of the first 6 coefficients in all frames are taken to form a 6-dimensional temporal coefficient of variation vector, which is used as the long-term scale feature.

3. The industrial equipment data acquisition and monitoring method according to claim 2, characterized in that, Operational data of the equipment under different typical loads were collected and subjected to multi-scale feature analysis to obtain its short-term, medium-term, and long-term features. The probability density distribution of the feature quantities in each scale was fitted by Gaussian mixture model (GMM). The distribution models of each feature quantity constituted the benchmark library of the operational feature spectrum of the industrial equipment.

4. The industrial equipment data acquisition and monitoring method according to claim 3, characterized in that, Multi-scale feature analysis is performed on the monitoring data to obtain a set of current feature values, in order to construct the equipment operation feature spectrum. The equipment operation feature spectrum is set as Fc={f1,f2,...,fn}, where n represents the total number of features.

5. The industrial equipment data acquisition and monitoring method according to claim 4, characterized in that, The ratio of the average total active power in the monitoring data over one hour to the rated power of the equipment is used as the average load rate. The ratio of the average total active power in each group of data in the operating data to the rated power of the equipment is used as the benchmark average load rate. The distribution model of the characteristic quantity corresponding to the operating data that satisfies |average load rate of equipment - benchmark average load rate| < 0.05 is matched in the operating characteristic spectrum benchmark library. If the match is successful, the distribution model of the corresponding characteristic quantity is directly extracted as the matching benchmark. Otherwise, the matching benchmark is generated by interpolation.

6. The industrial equipment data acquisition and monitoring method according to claim 5, characterized in that, For a feature quantity fi, when the corresponding matching benchmark follows an approximately Gaussian distribution, the historical mean μi and historical standard deviation σi corresponding to the feature quantity are extracted from the matching benchmark, and the feature deviation is analyzed. The expression for the feature deviation is set as: Di=|(fi-μi) / (k×σi)|, where Di represents the feature deviation and k represents the sensitivity adjustment coefficient. For a feature quantity fi, when the corresponding matching benchmark follows a complex or multimodal distribution, its feature value is stored in the matching benchmark in the form of a Gaussian mixture model, and the feature deviation is analyzed. The expression for the feature deviation is set as: Di=-ln(P(fi|Mi)), where P(fi|Mi) represents the probability density of the current feature quantity under the corresponding Gaussian mixture model.

7. The industrial equipment data acquisition and monitoring method according to claim 6, characterized in that, The feature weights are set to Wi. For short-term features in the electrical energy dimension, the initial value of the feature weights is set to 0.

2. For short-term features in the vibration dimension, the initial value of the feature weights is set to 0.

25. For medium-term features in the electrical energy dimension, the initial value of the feature weights is set to 0.

2. For medium-term features in the thermodynamic dimension, the initial value of the feature weights is set to 0.

15. For long-term features in the acoustic dimension, the average value of the feature deviation of the six coefficients is calculated, and the initial value of the feature weights is set to 0.

2. If the feature deviation of fi has consistently met the condition Di > 0.7 for the past 30 minutes, its corresponding feature weight is temporarily increased, so that the increased feature weight is W1i = (min(Wi × (1 + β), 0.4). At the same time, the other feature weights are decreased proportionally to the feature weights, so that the sum of the feature weights is 1. β represents the adjustment coefficient.

8. The industrial equipment data acquisition and monitoring method according to claim 7, characterized in that, The feature deviation is smoothed and normalized using a sigmoid function. The normalized feature deviation is set as Norm(Di), where Norm(Di) = 1 / (1 + exp(-s × (Di - d0))), d0 represents the offset parameter, and s represents the slope parameter. The normalized feature deviation is then weighted and fused based on the feature weights to obtain a comprehensive anomaly index. Set the sliding time window size to 10 minutes, analyze a comprehensive anomaly index every minute within the sliding time window, calculate the average of all comprehensive anomaly indices within the window, and use it as the sliding index mean. The number of sample points whose comprehensive abnormal index exceeds the intermediate threshold within the sliding time window is counted, then divided by the total number of samples within the window, and this number is used as the proportion of persistently high values. A tiered early warning decision is constructed based on a comprehensive anomaly index. When Ca < Thlow and Rh < 0.1, the system is judged to be operating normally and there is no alarm. When Think ≤ Ca < Thmid or Rh ≥ 0.1, an early warning is issued; when Ca ≥ Thmid and Rh ≥ 0.3, the system is determined to be in an abnormal state, and a medium-level alarm is issued, listing the three feature values ​​with the highest deviation from the target value to the user; when Ca ≥ Thhigh and Rh ≥ 0.5 and the duration is greater than 3 minutes, the system is determined to be abnormal, and a fault alarm is triggered. Ca represents the moving average, Rh represents the proportion of sustained high values, Think represents the low-level threshold, Thmid represents the medium-level threshold, and Thhigh represents the high-level threshold.

9. An industrial equipment data acquisition and monitoring platform, applied to the industrial equipment data acquisition and monitoring method as described in any one of claims 1-8, characterized in that, include: The benchmark construction module is used to build a benchmark library of operating characteristic spectra for industrial equipment. The feature extraction module is used to collect monitoring data of industrial equipment in real time and construct the equipment operation feature spectrum; The deviation analysis module is used to build a deviation analysis model based on the operating characteristic spectrum benchmark library and the equipment operating characteristic spectrum in order to determine the characteristic deviation degree and characteristic weight; The graded early warning module is used to comprehensively analyze feature deviation and feature weight to determine the comprehensive anomaly index and formulate graded early warning decisions.

10. An electronic device, characterized in that, The electronic device includes: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the industrial equipment data acquisition and monitoring method as described in any one of claims 1-8.

Images