A digital operation and maintenance control system and method for electromechanical facilities

Abstract

The present invention relates to the technical field of digital operation and maintenance systems, in particular to a digital operation and maintenance control system and method for electromechanical facilities, including: a vibration data acquisition module, a data processing module communicatively connected to the vibration data acquisition module, a vibration feature analysis module communicatively connected to the data processing module, receiving vibration state parameters sent by the data processing module; based on the vibration state parameters, performing vibration data topological feature analysis and prediction system algorithms; generating prediction results; an operation and maintenance decision-making module communicatively connected to the vibration feature analysis module, receiving the prediction results sent by the vibration feature analysis module; based on the prediction results, generating operation and maintenance decision-making information; an execution control module communicatively connected to the operation and maintenance decision-making module, receiving the operation and maintenance decision-making information sent by the operation and maintenance decision-making module; based on the operation and maintenance decision-making information, performing corresponding operation and maintenance control operations, capable of capturing minor anomalies that may be overlooked by traditional methods, improving the accuracy and lead time of fault prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of digital operation and maintenance systems, and particularly to a digital operation and maintenance control system and method for electromechanical facilities. Background Art

[0002] With the advent of the Industrial 4.0 era, the digital operation and maintenance control of electromechanical facilities has become an indispensable part of modern industrial production. Traditional equipment maintenance methods mainly rely on regular inspections and experience-based judgments. This method often has difficulty in detecting potential faults in a timely manner, easily leading to unplanned equipment downtime and causing huge economic losses. In recent years, with the development of sensing technology and data analysis technology, predictive maintenance has gradually become the focus of the industry.

[0003] Currently, the closest prior arts mainly include threshold-based monitoring systems and predictive maintenance systems using machine learning. The threshold-based monitoring system triggers an alarm when the measured value exceeds a series of set thresholds for vibration amplitude and frequency. This method is simple to operate and has a fast response speed, but it has obvious limitations. First, it is difficult to capture complex fault patterns, especially those faults without obvious threshold breakthroughs in the early stage. Second, fixed thresholds often cannot adapt to the dynamic changes of equipment under different working conditions, easily resulting in a high false alarm rate or a high missed alarm rate.

[0004] The predictive maintenance system using machine learning trains a model with historical data and then classifies and predicts new data. This method has made significant progress compared with the traditional threshold method and can handle more complex data patterns. However, it still has some problems. First, machine learning models are often black-box and lack interpretability, which may be unacceptable in some high-risk scenarios. Second, the performance of machine learning models highly depends on the quality and quantity of training data. In the case of scarce data or changing data distributions, the model performance may drop significantly. Finally, traditional machine learning methods may have difficulty in capturing some complex structures and long-term dependencies in the data, which limits their application effects in complex systems.

[0005] In view of these limitations of the prior art, the industry urgently needs a more advanced and reliable digital operation and maintenance control system for electromechanical facilities. An ideal system should be able to deeply analyze the operating status of equipment from multiple dimensions, accurately predict potential faults, and at the same time have strong adaptability and interpretability. Summary of the Invention

[0006] A digital operation and maintenance control system and method for electromechanical facilities of the present invention are precisely designed for the above problems. The system of the present invention realizes the comprehensive perception, accurate analysis, and reliable prediction of the operating status of electromechanical facilities by integrating multidisciplinary knowledge and advanced algorithms.

[0007] The present invention provides a digital operation and maintenance control system for electromechanical facilities, including:

[0008] A vibration data acquisition module, configured to:

[0009] Acquire the vibration data of the electromechanical facilities;

[0010] Send the vibration data to the data processing module;

[0011] A data processing module, communicatively connected to the vibration data acquisition module, configured to:

[0012] Receive the vibration data sent by the vibration data acquisition module;

[0013] Perform time-frequency domain conversion on the vibration data to obtain a time-domain signal and a frequency-domain signal;

[0014] Generate vibration state parameters based on the time-domain signal and the frequency-domain signal;

[0015] A vibration feature analysis module, communicatively connected to the data processing module, configured to:

[0016] Receive the vibration state parameters sent by the data processing module;

[0017] Execute the vibration data topological feature analysis and prediction system algorithm based on the vibration state parameters;

[0018] Generate a prediction result;

[0019] An operation and maintenance decision-making module, communicatively connected to the vibration feature analysis module, configured to:

[0020] Receive the prediction result sent by the vibration feature analysis module;

[0021] Generate operation and maintenance decision-making information based on the prediction result;

[0022] An execution control module, communicatively connected to the operation and maintenance decision-making module, configured to:

[0023] Receive the operation and maintenance decision-making information sent by the operation and maintenance decision-making module;

[0024] Perform corresponding operation and maintenance control operations based on the operation and maintenance decision-making information.

[0025] Preferably, the vibration data acquisition module includes:

[0026] A first sensor, configured to acquire the vibration data of the installation support device of the electromechanical facilities;

[0027] A second sensor, configured to acquire the vibration data of the electromechanical facilities body;

[0028] A data aggregation unit, communicatively connected to the first sensor and the second sensor, for aggregating vibration data collected by the first sensor and the second sensor.

[0029] Preferably, the data processing module includes:

[0030] A time-frequency conversion unit, for performing time-frequency domain conversion on the vibration data;

[0031] A parameter generation unit, for generating vibration state parameters based on the time domain signal and the frequency domain signal;

[0032] Wherein, the vibration state parameters include a peak factor, a kurtosis index, an impact index, and a root mean square value.

[0033] Preferably, the vibration feature analysis module includes:

[0034] A topological mapping unit, for mapping the vibration state parameters to a high-dimensional topological space;

[0035] A group theory feature extraction unit, for performing group theory analysis on the features of the high-dimensional topological space;

[0036] A matrix embedding unit, for embedding group theory features into a special matrix and performing spectral analysis;

[0037] A number theory transformation unit, for performing number theory transformation and chaotic analysis on the eigenvalues;

[0038] A quantum probability prediction unit, for constructing a quantum probability prediction model based on the results of chaotic analysis.

[0039] Preferably, the topological mapping unit performs the following operations:

[0040] Preferably, the topological mapping unit performs the following operations:

[0041] ,

[0042] Wherein, is a topological feature vector, is the input vibration data set, is the k-dimensional Betti number, representing the number of k-dimensional topological features, is a weight coefficient, used to adjust the importance of Betti numbers in different dimensions, is a dimension index, is the highest dimension considered, represents summation.

[0043] Preferably, the group theory feature extraction unit performs the following operations:

[0044] Calculating group - theoretic characters ,

[0045] wherein, is the group - theoretic character, is the topological - feature vector, is the product symbol, is the index variable, is the dimension of the topological - feature vector, is 's symmetry group, is the i - th component of the topological - feature vector, represents the semi - direct product operation, is the direct product of m - dimensional binary cyclic groups.

[0046] Preferably, the matrix - embedding unit performs the following operations:

[0047] Calculating the matrix representation of the group ;

[0048] Calculating the set of eigenvalues ,

[0049] wherein, is the matrix representation of the group, is the group - theoretic character, is the matrix - exponential function, represents summation, is an element in the group G, is the representation matrix of the group element g, is 's set of eigenvalues, is the eigenvalue.

[0050] Preferably, the number - theoretic transformation unit performs the following operations:

[0051] Calculating the Zeta function of the eigenvalues ;

[0052] Calculating the Lyapunov exponent ,

[0053] wherein, is the Zeta function of the eigenvalues, is the set of eigenvalues, is the product symbol, is the index variable, is the number of eigenvalues, is the i - th eigenvalue, is the complex variable, is the Lyapunov exponent, is the limit symbol, is the variable tending to infinity, is the natural logarithm, is the absolute value symbol, is the i-th derivative of the Zeta function.

[0054] Preferably, the quantum probability prediction unit performs the following operations:

[0055] Construct a quantum state ;

[0056] Calculate the probability distribution ;

[0057] where, , is the quantum state, represents the summation symbol, is the index variable, is the total number of possible states of the system, is the probability amplitude, is the probability that the measurement result is i, is the i-th ground state of the system, is the natural constant, is the inverse temperature of the system, is the energy eigenvalue, is the summation index, is the Lyapunov exponent corresponding to the i-th state.

[0058] A digital operation and maintenance control method for electromechanical facilities based on the system includes the following steps:

[0059] Obtain the vibration data of the electromechanical facilities;

[0060] Perform time-frequency domain conversion on the vibration data to obtain a time-domain signal and a frequency-domain signal;

[0061] Generate vibration state parameters based on the time-domain signal and the frequency-domain signal;

[0062] Execute the vibration data topological feature analysis and prediction system algorithm, including:

[0063] Map the vibration state parameters to a high-dimensional topological space;

[0064] Perform group theory analysis on the characteristics of the high-dimensional topological space;

[0065] Embed the group theory characteristics into a special matrix and perform spectral analysis;

[0066] Perform number theory transformation and chaos analysis on the eigenvalues;

[0067] Construct a quantum probability prediction model based on the chaos analysis results;

[0068] Generate operation and maintenance decision-making information based on the output result of the quantum probability prediction model;

[0069] Execute corresponding operation and maintenance control operations according to the operation and maintenance decision-making information.

[0070] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0071] First of all, the system of the present invention adopts an innovative Vibration Data Topological Feature Analysis and Prediction System (VDTAPS) algorithm, which integrates advanced concepts from multiple mathematical branches such as topology, group theory, and matrix theory. This multi-dimensional data analysis method can capture minor anomalies that may be overlooked by traditional methods, greatly improving the accuracy and lead time of fault prediction. For example, in practical applications, this system can give an accurate early warning 72 hours before a fault occurs, which provides sufficient preparation time for maintenance personnel and effectively reduces the risk of unplanned downtime.

[0072] Secondly, the system of the present invention has strong adaptability. By introducing a quantum probability prediction model, the system can better process high-dimensional complex data and adapt to the dynamic changes of equipment under different working conditions. This solves the problem that traditional fixed-threshold methods are difficult to adapt to complex working conditions and significantly reduces the false alarm rate and missed alarm rate. In the test, the false alarm rate of this system is only 2.1%, and the missed alarm rate is only 1.5%, far superior to the prior art.

[0073] Thirdly, the system of the present invention also has significant advantages in interpretability. Different from black-box machine learning models, each step of the VDTAPS algorithm has clear physical or mathematical meanings. This not only helps users understand and trust the decisions of the system but also provides the possibility for further optimizing the system. In some high-risk industrial scenarios, this interpretability is particularly important.

[0074] Fourthly, the system of the present invention realizes high flexibility and scalability through modular design. Communication between each functional module is carried out through clear interfaces, and it can be flexibly configured and optimized according to specific application requirements. This enables the system to adapt to various types of electromechanical facilities and operating environments, greatly expanding its application scope.

[0075] Finally, the system of the present invention also performs excellently in terms of overall performance and efficiency. Despite using complex algorithms, the response time of the system remains around 0.5 seconds, fully meeting the requirements of real-time monitoring. At the same time, since the system can predict potential faults earlier and more accurately, it can help users optimize the maintenance plan, reduce unnecessary inspections and maintenance, thereby significantly reducing the operation and maintenance costs and improving the overall utilization efficiency of the equipment.

[0076] Generally speaking, through innovative algorithm design and system architecture, the present invention effectively addresses the deficiencies of existing technologies in terms of accuracy, adaptability, interpretability, and efficiency. It can not only significantly improve the reliability and safety of electromechanical facilities but also assist users in optimizing resource allocation and enhancing overall operational efficiency. Against the backdrop of the industrial digital transformation, the present invention provides a comprehensive, reliable, and efficient solution for the intelligent operation and maintenance of electromechanical facilities, holding important theoretical significance and practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is the logical block diagram of the overall system of the present invention.

[0078] Figure 2 It is the logical block diagram of the vibration data acquisition module of the present invention.

[0079] Figure 3 It is the logical block diagram of the data processing module of the present invention.

[0080] Figure 4 It is the logical block diagram of the vibration feature analysis module of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0081] Referring to Figures 1-4 , the present invention provides a digital operation and maintenance control system and method for electromechanical facilities, aiming to improve the operation and maintenance efficiency and reliability of electromechanical facilities. The following is a detailed description of the system of the present invention:

[0082] The digital operation and maintenance control system of the present invention includes a vibration data acquisition module 1, a data processing module 2, a vibration feature analysis module 3, an operation and maintenance decision-making module 4, and an execution control module 5. These modules work together to achieve a complete closed-loop from data acquisition to decision execution.

[0083] The vibration data acquisition module 1 is the data input end of the system and is used to obtain the vibration data of electromechanical facilities. Preferably, this module can collect various vibration parameters, such as displacement, velocity, and acceleration, etc. In an embodiment of the present invention, the vibration data acquisition module 1 uses a high-precision piezoelectric acceleration sensor, and the sampling frequency can reach 20 kHz to ensure capturing high-frequency vibration signals. After obtaining the data, the vibration data acquisition module 1 will first perform a data validity check. This step is to ensure the quality of the collected data and avoid incorrect data caused by sensor failures, signal interference, etc. from affecting subsequent analysis. Specifically, the system will check the integrity of the data, whether it is within the expected range, and whether there are abnormal fluctuations, etc. If the data is determined to be invalid, the system will re-collect the data; if the data is valid, it will be transmitted to the data processing module 2 for the next step of processing.

[0084] The data acquired by the vibration data acquisition module 1 is then sent to the data processing module 2. The data processing module 2 is the core processing unit of the system and is responsible for the preliminary processing and analysis of the original vibration data. Specifically, this module first performs time-frequency domain transformation on the vibration data to obtain the time-domain signal and the frequency-domain signal. Time-frequency domain transformation usually adopts the fast Fourier transform (FFT) algorithm. In this invention, an improved wavelet packet transform algorithm is preferably used, which can better process non-stationary signals.

[0085] Based on the obtained time-domain signal and frequency-domain signal, the data processing module 2 further generates vibration state parameters. These parameters usually include statistical features such as root mean square value, peak factor, kurtosis, margin, etc., as well as characteristic frequency components in the frequency spectrum. For example, for bearing fault diagnosis, the characteristic frequency of the bearing and its harmonics may be particularly concerned.

[0086] The vibration state parameters generated by the data processing module 2 are then sent to the vibration feature analysis module 3. This module is the core innovative part of this invention and executes an innovative algorithm called the Vibration Data Topological Analysis and Prediction System (VDTAPS) algorithm, aiming to extract deeper features from the vibration data and perform prediction analysis.

[0087] The first step of the VDTAPS algorithm is to map the vibration state parameters to a high-dimensional topological space. This step uses persistent homology theory, which can capture the essential features in the data.

[0088] Next, the algorithm uses group theory methods to extract deeper algebraic features. This step can effectively capture the symmetry and invariance in the vibration data, which is very useful for identifying certain specific types of mechanical faults.

[0089] The subsequent steps of the VDTAPS algorithm include matrix embedding and spectral analysis, number theory transformation and chaos analysis, and the construction of a quantum probability prediction model. These steps further extract the deep features of the vibration data and finally generate prediction results.

[0090] Based on the prediction results generated by the vibration feature analysis module 3, the operation and maintenance decision-making module 4 will generate corresponding operation and maintenance decision-making information. These decision-making information may include the health status assessment of the equipment, fault early warning, maintenance suggestions, etc. For example, if the prediction results show that there is an 80% probability that a certain bearing will fail within the next 100 hours, the operation and maintenance decision-making module 4 may generate a decision to replace the bearing during the next planned shutdown. After generating the decision-making information, the operation and maintenance decision-making module 4 will make a key judgment: whether maintenance of the electromechanical facilities is required. This judgment is based on multiple factors, including but not limited to the predicted failure probability, the severity of potential failures, the importance of the equipment, maintenance costs, etc. If the system determines that maintenance is required, the execution control module 5 will be activated to start the corresponding maintenance process. If it is determined that maintenance is not required at present, the system will continue to monitor the equipment status and restart the data collection process. This dynamic decision-making mechanism can maximize the equipment usage efficiency while ensuring the reliable operation of the equipment and avoiding unnecessary maintenance costs.

[0091] Finally, the execution control module 5 performs corresponding operation and maintenance control operations based on the operation and maintenance decision-making information. This may include automatically adjusting equipment parameters, sending maintenance instructions, triggering the alarm system, etc. In some cases, if signs of a serious fault are detected, the execution control module 5 may trigger an emergency shutdown procedure to prevent greater losses.

[0092] The system of the present invention also includes some optimized designs. For example, the vibration data acquisition module 1 includes a first sensor 11 and a second sensor 12. The first sensor 11 is used to obtain the vibration data of the installation support device of the electromechanical facilities, while the second sensor 12 is used to obtain the vibration data of the electromechanical facilities body. This dual-sensing design can more comprehensively capture the vibration state of the electromechanical facilities, which helps to distinguish the vibrations from the equipment itself and the external environment.

[0093] The data aggregation unit 13 is communicatively connected to the first sensor 11 and the second sensor 12, and is used to aggregate the vibration data collected by the two sensors. The data aggregation unit 13 adopts an advanced data fusion algorithm, which can effectively integrate the vibration data from different sources and improve the reliability and integrity of the data.

[0094] The data processing module 2 includes a time-frequency conversion unit 21 and a parameter generation unit 22. The time-frequency conversion unit 21 is responsible for converting the vibration data into the time-frequency domain, while the parameter generation unit 22 generates vibration state parameters based on the time-domain signal and the frequency-domain signal. This modular design improves the flexibility and maintainability of the system.

[0095] The vibration state parameters generated by the parameter generation unit 22 include the peak factor, kurtosis index, shock index, and root mean square value. The selection of these parameters is based on the experience and theory of electromechanical equipment fault diagnosis. For example, the kurtosis index is particularly effective in detecting early bearing faults, while the shock index can reflect the gear meshing condition.

[0096] The system of the present invention realizes the intelligence and precision of the operation and maintenance of electromechanical facilities by integrating multidisciplinary knowledge and advanced algorithms. It can not only detect potential faults in a timely manner but also predict the future state of the equipment, providing a solid foundation for preventive maintenance. This method can significantly reduce the unplanned downtime of the equipment, improve production efficiency, reduce maintenance costs, and extend the service life of the equipment.

[0097] The vibration feature analysis module 3 of the present invention is the core innovative part of the system, including a topological mapping unit 31, a group theory feature extraction unit 32, a matrix embedding unit 33, a number theory transformation unit 34, and a quantum probability prediction unit 35. These units together constitute a complete implementation of the vibration data topological feature analysis and prediction system (VDTAPS) algorithm.

[0098] In electromechanical facilities, vibration data often contains complex non-linear structures. Through persistent homology, invariant features in these structures can be captured, that is, the geometric shape changes at different time scales. For example, sensors installed on the blades of a wind turbine will collect a large amount of data that fluctuates with time and the environment. Persistent homology can help identify which fluctuations are caused by equipment aging or potential faults rather than external factors such as wind speed changes.

[0099] The topological mapping unit 31 is responsible for mapping the vibration state parameters to a high-dimensional topological space. The core of this step is to calculate the topological feature vector of the vibration data. Specifically, the topological mapping unit 31 performs the following operations:

[0100] ,

[0101] In this formula is the topological feature vector, representing a high-dimensional vector extracted from the vibration data that can reflect the structural characteristics of the system, is the input vibration data set, is the k-dimensional Betti number, representing the number of k-dimensional topological features, is the weight coefficient, used to adjust the importance of different-dimensional Betti numbers, is the dimension index, is the highest dimension considered, represents the summation. In the preferred embodiment of the present invention, usually takes the value of 3, so that the key topological features of most electromechanical systems can be captured. The weight coefficient The selection can be adjusted according to specific application scenarios. For example, in the case of rotating machinery, more attention may be paid to 1D and 2D topological features. Therefore, and the values can be increased.

[0102] This method can effectively distinguish noise from meaningful information, thereby improving the accuracy of fault detection. It is particularly useful for early detection of minor anomalies because even small changes may indicate impending failures.

[0103] The vibration data acquisition module 1 collects high-frequency vibration signals from multiple locations on the wind turbine blade. These data are sent to the data processing module 2 for preliminary processing, including denoising, normalization, etc. Then, the time-frequency conversion unit 21 converts the time-domain signal into a frequency-domain signal, and the parameter generation unit 22 calculates various statistical features. Finally, these features are passed as inputs to the topological mapping unit 31, which calculates the topological feature vector using the above formula for subsequent analysis.

[0104] The normal operation of electromechanical equipment is usually accompanied by specific periodicity and symmetry. For example, the rotational motion of a gear system generates regular vibration patterns. If this pattern changes, such as becoming asymmetric or having additional frequency components, it may be due to gear wear or other mechanical problems. Through group theory analysis, these symmetry changes can be quantified and used as an indicator for fault diagnosis.

[0105] Based on the topological features, the group theory feature extraction unit 32 further extracts algebraic features. The core operations of this step are as follows:

[0106] ,

[0107] In this formula, is the group theory feature, is the topological feature vector, is the product symbol, is the index variable, is the dimension of the topological feature vector, is 's symmetry group, is the i-th component of the topological feature vector, represents the semi-direct product operation, which is a way to combine two groups where one group acts on the other to form a new group. Here it is used to combine different symmetry groups to enhance the feature expression ability. is the direct product of m-dimensional binary cyclic groups, consisting of m Formed by the direct product of (i.e., the additive group modulo 2), it represents a discrete symmetric operation, which can be understood as a flipping or inversion operation. This group - theoretic analysis method can effectively capture the symmetry and invariance in vibration data, which is very useful for identifying certain specific types of mechanical faults. For example, in a gear system, the vibration data during normal operation usually has a certain periodic symmetric structure, while this symmetry may be broken when a fault occurs. m is the dimension of the topological feature vector, that is, the number of symmetric groups participating in the semi - direct product operation, corresponding to the number of vibration state parameters.

[0108] This method helps to identify subtle changes that are not easily detectable but are crucial for the performance of the equipment. Especially in large and complex systems, traditional threshold - based methods may not be able to capture these problems in a timely manner, while group - theoretic features provide a more sensitive and accurate tool.

[0109] The vibration data collected from sensors at different positions inside the gearbox is pre - processed and then used to construct vibration state parameters. The group - theoretic feature extraction unit 32 uses these parameters to calculate group - theoretic features. By comparing the differences in group - theoretic features between the normal state and the current state, the health status of the equipment can be evaluated and potential fault points can be predicted.

[0110] The vibration characteristics in an electromechanical system can be expressed in matrix form. For example, unbalance of the motor rotor will cause specific vibration modes, which can be represented by a set of complex matrices. By performing spectral analysis on these matrices, the dynamic behavior of the system, such as resonance frequency, damping ratio, etc., can be understood. In addition, the eigenvalues also provide important information about the system's stability and response speed.

[0111] The matrix embedding unit 33 embeds the group - theoretic features into a special matrix and performs spectral analysis. This step includes two main operations:

[0112] First, calculate the matrix representation of the group:

[0113] ,

[0114] Second, calculate the set of eigenvalues:

[0115] ,

[0116] In these formulas, is the matrix representation of the group, which maps group elements to the matrix space through a linear transformation, so that the operations of the group can be represented by matrix multiplication. is the group - theoretic feature, which is a homomorphism from the group to the matrix group. It maps each element in the group into a corresponding matrix, retaining the algebraic structure of the group. is the matrix exponential function. For a matrix A, is defined as , which is an infinite series that, when applied to a matrix, produces a new matrix. denotes the sum, is an element in the group G, is the representation matrix of the group element g, is the set of eigenvalues of is an eigenvalue, and eigenvalues are the intrinsic values of the matrix and they determine the main behaviors of the matrix, such as stretching, rotation, etc. is the number of eigenvalues, equal to the order of the matrix, i.e., the number of rows or columns of the matrix. This matrix embedding method can transform complex group - theoretic features into computable numerical features for subsequent analysis and processing.

[0117] Through matrix embedding and spectral analysis, not only can a deeper understanding be obtained, but also efficient control strategies can be developed. For example, when designing an active vibration control system, it is very important to understand the natural frequencies and modes of the system, which helps to select appropriate controller parameters to achieve the best vibration reduction effect.

[0118] During the operation of the motor, the interaction between the stator and the rotor generates complex electromagnetic and mechanical forces, which cause vibrations. The vibration data acquisition module 1 is responsible for monitoring these vibrations and transmitting the data to the data processing module 2. After a series of processing steps, including time - frequency conversion and feature extraction, the obtained vibration state parameters are used as the input of the matrix embedding unit 33. This unit calculates the matrix representation and its eigenvalues, providing a basis for subsequent fault diagnosis and preventive maintenance.

[0119] Electromechanical systems sometimes exhibit chaotic behavior, especially under high loads or extreme working conditions. For example, when a hydraulic pump operates under high pressure, the internal fluid flow may become unstable, generating irregular vibrations. By introducing the Zeta function and calculating the Lyapunov exponent, the degree of this chaos can be quantified, and it can be predicted whether the system will enter an unstable state.

[0120] The number - theoretic transform unit 34 performs number - theoretic transform and chaos analysis on the eigenvalues. This step first calculates the Zeta function of the eigenvalues:

[0121] ,

[0122] Then, calculate the Lyapunov exponent:

[0123] ,

[0124] In these formulas, The Zeta function with eigenvalues, which is a complex variable function used to study the distribution law of eigenvalues, especially the behavior at infinity, is the set of eigenvalues, is the product symbol, is the index variable, is the number of eigenvalues, is the i-th eigenvalue, is the complex variable, is the Lyapunov exponent, which measures the stability of the system as it evolves over time. A positive value indicates that the system tends to be unstable, while a negative value indicates stability, is the limit symbol, is the variable tending to infinity, is the natural logarithm, used to calculate the logarithm of the absolute value of the derivative. This is to ensure that even if the derivative is negative, a positive value can be obtained, thus not affecting the positive or negative judgment of the Lyapunov exponent, is the absolute value symbol, is the i-th derivative of the Zeta function. The derivative describes the rate of change of the Zeta function, and higher-order derivatives provide more detailed change information. The introduction of the Zeta function enables the study of the distribution of eigenvalues from a number theory perspective, while the Lyapunov exponent reflects the degree of chaos of the system. In a mechatronic system, a sudden increase in the Lyapunov exponent may mean that the system is about to enter an unstable state.

[0125] This is crucial for taking preventive measures in advance to avoid the occurrence of serious failures. For example, when the Lyapunov exponent suddenly increases, it means that the system is approaching the critical point. At this time, the warning mechanism should be immediately activated to notify the operator to check the equipment and avoid further deterioration.

[0126] The pressure sensor and flow sensor inside the hydraulic pump continuously monitor the operating state of the system. These data are integrated into the vibration feature analysis module 3, where the number theory transformation unit 34 uses the Zeta function and the Lyapunov exponent to evaluate the stability of the system. Once an abnormal situation is detected, the operation and maintenance decision-making module 4 will generate corresponding maintenance suggestions according to the analysis results, such as adjusting the working parameters or arranging an emergency repair.

[0127] In some cases, the future behavior of mechatronic systems has a high degree of uncertainty, and traditional statistical models are difficult to accurately predict. Using a quantum probability prediction model can simulate this uncertainty and provide a probability distribution closer to the actual situation. For example, for the safety assessment of nuclear power plants, the influence of various uncertain factors such as material aging and environmental changes needs to be considered. The quantum probability model can help engineers better understand and manage these risks.

[0128] Finally, the quantum probability prediction unit 35 constructs a quantum probability prediction model based on the chaos analysis results. This step first constructs the quantum state:

[0129] ,

[0130] Then it calculates the probability distribution:

[0131] ,

[0132] where, , is the quantum state, represents the summation symbol, is the index variable, is the total number of possible states of the system, is the probability amplitude, and the square of the probability amplitude gives the probability that the measurement result is in this state, is the probability that the measurement result is i, calculated according to the Boltzmann distribution, reflecting the possibility that the system is in a specific state, is the i-th ground state of the system, representing a basic state of the system, and all possible states are linear combinations of these ground states, is the natural constant, is the inverse temperature of the system, which is the reciprocal of the thermodynamic temperature i.e., , where is the Boltzmann constant, is the energy eigenvalue, is the summation index, is the Lyapunov exponent corresponding to the i-th state.

[0133] This prediction model based on quantum probability has the advantage of dealing with high-dimensional complex data and can capture subtle correlations that traditional probability models may overlook. In the operation and maintenance of electromechanical facilities, this model can provide more accurate fault prediction and remaining useful life estimation.

[0134] The advantage of this model is that it can handle high-dimensional complex data and capture subtle correlations that traditional probability models may overlook. This makes it very suitable for long-term prediction and remaining useful life estimation, providing strong support for formulating scientific and reasonable maintenance plans.

[0135] Various sensors in the nuclear power plant (such as temperature, pressure, radiation level, etc.) continuously collect data, and after preprocessing, these data form vibration state parameters. The quantum probability prediction unit 35 uses these parameters to construct the quantum state and predicts the future fault risk by calculating the probability distribution. The operation and maintenance decision-making module 4 can optimize the maintenance strategy according to the prediction results to ensure the safe and stable operation of the nuclear power plant.

[0136] The VDTAPS algorithm of the present invention realizes the in-depth mining and prediction of vibration data of electromechanical facilities through this series of complex mathematical transformations and analyses. Compared with traditional methods, VDTAPS can better process non-linear and non-stationary vibration signals, capture tiny anomalies that may be overlooked, thereby improving the accuracy of fault diagnosis and prediction.

[0137] In practical applications, the various parameters of the VDTAPS algorithm can be optimized according to the specific types of electromechanical facilities and operating environments. For example, for high-speed rotating equipment, it may be necessary to increase the weight of high-frequency vibration components; while for large static equipment, more attention may be paid to low-frequency vibration characteristics. Through this flexible parameter adjustment, the system of the present invention can adapt to various different industrial scenarios and provide precise operation and maintenance support for various electromechanical facilities.

[0138] The present invention also provides a digital operation and maintenance control method for electromechanical facilities corresponding to the above system. This method is a further elaboration and refinement of the system functions, aiming to provide an operation and maintenance solution for electromechanical facilities with strong operability and remarkable effects.

[0139] First of all, this method starts by acquiring the vibration data of electromechanical facilities. Preferably, this step can use high-precision vibration sensors, such as piezoelectric acceleration sensors or fiber optic sensors. In an embodiment of the present invention, the acquisition frequency of vibration data can reach 20 kHz to ensure the capture of high-frequency vibration signals. At the same time, in order to comprehensively reflect the operating state of electromechanical facilities, this method can also collect auxiliary parameters such as temperature, pressure, and current.

[0140] After acquiring the vibration data, this method performs time-frequency domain conversion on these data to obtain time-domain signals and frequency-domain signals. This step usually uses the fast Fourier transform (FFT) algorithm, but in the preferred embodiment of the present invention, an improved wavelet packet transform algorithm is used. Compared with the traditional FFT, the wavelet packet transform has obvious advantages in processing non-stationary signals and can better capture transient characteristics and local characteristics.

[0141] Next, based on the obtained time-domain signals and frequency-domain signals, this method generates vibration state parameters. These parameters usually include time-domain statistical characteristics (such as root mean square value, peak factor, kurtosis, margin, etc.) and frequency-domain characteristics (such as characteristic frequencies and their amplitudes). In an embodiment of the present invention, envelope spectrum analysis is also introduced, which is particularly effective for diagnosing early faults of components such as bearings and gears.

[0142] After generating the vibration state parameters, this method executes the vibration data topology feature analysis and prediction system (VDTAPS) algorithm. This is the core innovative step of the present invention, which includes a series of complex mathematical transformation and analysis processes.

[0143] The VDTAPS algorithm first maps the vibration state parameters to a high-dimensional topological space. This step utilizes the persistent homology theory, which can capture the essential topological features in the data. For example, for bearing faults, different types of faults (such as inner race faults, outer race faults, rolling element faults) may exhibit different characteristic structures in the topological space.

[0144] Subsequently, the algorithm conducts a group theory analysis on the features of the high-dimensional topological space. This step can effectively capture the symmetries and invariances in the vibration data. In practical applications, this is very useful for identifying certain specific types of mechanical faults. For example, a gear system usually exhibits a certain periodic symmetric structure during normal operation, and this symmetry may be broken when a fault occurs.

[0145] Next, the algorithm embeds the group theory features into a special matrix and performs spectral analysis. This step transforms the complex group theory features into computable numerical features, facilitating subsequent analysis and processing. The distribution and variation of the eigenvalues can reflect the overall dynamic characteristics of the electromechanical system.

[0146] Then, the algorithm conducts number theory transformation and chaos analysis on the eigenvalues. This step introduces the Zeta function and Lyapunov exponents, which can deeply analyze the dynamic behavior of the system from a mathematical perspective. In particular, the Lyapunov exponent can effectively quantify the chaos degree of the system and is of great significance for predicting the long-term behavior of the system.

[0147] Finally, the algorithm constructs a quantum probability prediction model based on the results of chaos analysis. This prediction model based on the principles of quantum mechanics can better handle high-dimensional complex data and capture subtle correlation relationships compared to traditional probability models. In the operation and maintenance of electromechanical facilities, this model can provide more accurate fault prediction and remaining useful life estimation.

[0148] Based on the output results of the VDTAPS algorithm, this method generates operation and maintenance decision-making information. These information may include the health status assessment of the equipment, fault warnings, maintenance suggestions, etc. For example, if the prediction model shows that there is an 80% probability that a certain key component will fail within the next 100 hours, the system may generate a decision to replace the component during the next planned shutdown.

[0149] Finally, this method executes corresponding operation and maintenance control operations according to the operation and maintenance decision-making information. This may include automatically adjusting equipment parameters, sending maintenance instructions, triggering the alarm system, etc. In some cases, if signs of a serious fault are detected, the system may trigger an emergency shutdown procedure to prevent greater losses.

[0150] The method of the present invention realizes the intelligentization and precision of the operation and maintenance of electromechanical facilities by integrating multidisciplinary knowledge and advanced algorithms. It can not only detect potential faults in a timely manner, but also predict the future state of equipment, providing a solid foundation for preventive maintenance. This method can significantly reduce the unplanned downtime of equipment, improve production efficiency, reduce maintenance costs, and extend the service life of equipment.

[0151] In practical applications, each step of this method can be optimized and adjusted according to the specific type of electromechanical facilities and operating environment. For example, for high-speed rotating equipment, it may be necessary to increase the weight of high-frequency vibration components; while for large static equipment, more attention may be paid to low-frequency vibration characteristics. Through this flexible parameter adjustment, the method of the present invention can adapt to various industrial scenarios and provide precise operation and maintenance support for various electromechanical facilities.

[0152] To verify the superiority of the present invention, a typical industrial scenario was selected for example testing and comparative analysis. Specifically, a key rotating equipment in a large petrochemical plant - a centrifugal compressor unit was selected as the test object. The reliability of this type of equipment directly affects the stability and efficiency of the entire production line, so it is of great significance to carry out precise operation and maintenance control on it.

[0153] Example 1: Adopting the digital operation and maintenance control system and method of the present invention

[0154] In this example, high-precision vibration sensors were installed on the centrifugal compressor unit, and the sampling frequency was set to 20 kHz. The system uses the VDTAPS algorithm of the present invention for data analysis and prediction. In particular, the parameters in the algorithm were optimized: in the topological mapping unit, the highest dimension d was set to 3, and the weight coefficient was adjusted according to the characteristics of the compressor. In the quantum probability prediction unit, the inverse temperature β of the system was adjusted according to historical data.

[0155] Comparative Example 1: Traditional threshold-based monitoring system

[0156] This is a commonly used method in the industrial field. The system sets a series of thresholds for vibration amplitude and frequency, and triggers an alarm when the measured value exceeds these thresholds.

[0157] Comparative Example 2: Predictive maintenance system using machine learning

[0158] This system uses the support vector machine (SVM) algorithm to classify and predict the equipment state. It trains the model through historical data, and then classifies new data to predict possible faults.

[0159] These three methods were tested for 6 months, mainly focusing on the following indicators:

[0160] 1. Fault prediction accuracy rate: The percentage of the number of times a fault is correctly predicted out of the total number of predictions.

[0161] 2. Prediction lead time: How long before an actual fault occurs the system can predict the fault.

[0162] 3. False alarm rate: The percentage of the number of times a fault is reported incorrectly out of the total number of reports.

[0163] 4. Missed alarm rate: The percentage of the number of times an actual fault that occurs is not detected out of the total number of faults.

[0164] 5. System response time: The time required from data input to generating a prediction result.

[0165] The detection methods and criteria are as follows:

[0166] 1. Fault prediction accuracy rate: Calculated by comparing the system's prediction result with the actual fault that occurs.

[0167] 2. Prediction lead time: Record the time when the system predicts a fault and the time when the actual fault occurs, and calculate the time difference between the two.

[0168] 3. False alarm rate: Record the number of times the system reports a fault but no fault actually occurs, and divide it by the total number of reports.

[0169] 4. Missed alarm rate: Record the number of faults that actually occur but are not detected by the system, and divide it by the total number of faults.

[0170] 5. System response time: Measure the time from data input to output result using a high-precision timer.

[0171] The test results are shown in the following table:

[0172]

[0173] From the test results, it can be seen that the system of the present invention has shown obvious advantages in multiple key indicators. Especially in the two indicators of fault prediction accuracy rate and prediction lead time, the performance of the present invention far exceeds that of the other two methods. This means that by using the system of the present invention, the factory can predict potential faults earlier and more accurately, thus having more sufficient time for maintenance planning and avoiding sudden shutdowns.

[0174] The system of the present invention also performs well in terms of false alarm rate and missed alarm rate. A low false alarm rate means a reduction in unnecessary inspections and maintenance, which can save a large amount of human and material resources. A low missed alarm rate ensures that the system can capture almost all potential faults, greatly improving the reliability and safety of the equipment.

[0175] Although in terms of system response time, the system of the present invention is slightly inferior to the traditional method, it is still within an acceptable range and much faster than the machine learning method. Considering that the prediction lead time provided by the system of the present invention is much greater than that of other methods, this difference in response time can be ignored.

[0176] These excellent test results are mainly due to the VDTAPS algorithm of the present invention. By integrating advanced concepts from multiple mathematical branches such as topology, group theory, and matrix theory, this algorithm can extract deeper features from vibration data. In particular, it can capture minor anomalies that may be overlooked by traditional methods and general machine learning methods, and these anomalies are often important indicators of early faults.

[0177] In addition, the quantum probability prediction model in the VDTAPS algorithm demonstrates powerful prediction capabilities. It can not only accurately predict the occurrence of faults but also provide a relatively long prediction lead time. This provides sufficient preparation time for the maintenance personnel in the factory, enabling them to perform maintenance at the most appropriate time and minimizing the impact on production.

[0178] Generally speaking, these test results fully prove the superiority of the present invention in the field of digital operation and maintenance control of electromechanical facilities. It can not only improve the reliability and safety of equipment but also help factories optimize maintenance strategies, improve production efficiency, and reduce operation and maintenance costs. In the context of the current Industry 4.0 and intelligent manufacturing, the present invention provides a powerful solution for the intelligent operation and maintenance of electromechanical equipment.

[0179] It should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A digital operation and maintenance control system for electromechanical facilities, characterized in that: include: Vibration data acquisition module for: Obtain vibration data of electromechanical facilities; Sending the vibration data to a data processing module; A data processing module is connected to the vibration data acquisition module for: Receiving vibration data sent by the vibration data acquisition module; Performing time-frequency domain conversion on the vibration data to obtain a time-domain signal and a frequency-domain signal; Generate a vibration state parameter based on the time domain signal and the frequency domain signal; A vibration characteristic analysis module is connected in communication with the data processing module and is used to: Receiving the vibration state parameters sent by the data processing module; Based on the vibration state parameters, executing vibration data topological feature analysis and prediction system algorithm; Generate prediction results; The operation and maintenance decision module is in communication with the vibration characteristic analysis module and is used to: Receiving the prediction result sent by the vibration characteristic analysis module; Based on the prediction results, generate operation and maintenance decision information; An execution control module is connected in communication with the operation and maintenance decision module and is used to: Receiving operation and maintenance decision information sent by the operation and maintenance decision module; Based on the operation and maintenance decision information, perform corresponding operation and maintenance control operations; The vibration characteristic analysis module comprises: A topological mapping unit, used for mapping the vibration state parameters to a high-dimensional topological space; A group theory feature extraction unit, used for performing group theory analysis on the features of the high-dimensional topological space; Matrix embedding unit, used to embed group theory features into special matrices and perform spectral analysis; A number theory transformation unit, used for performing number theory transformation and chaos analysis on eigenvalues; A quantum probability prediction unit, used to construct a quantum probability prediction model based on chaos analysis results; The topology mapping unit performs the following operations: , in, is the topological eigenvector, is the input vibration data set, is the k-dimensional Betti number, which indicates the number of k-dimensional topological features, is the weight coefficient, which is used to adjust the importance of Betti numbers in different dimensions. is the dimension indicator, is the highest dimension considered, It means sum; The group theory feature extraction unit performs the following operations: Computational group theory features , in, For group theory characteristics, is the topological eigenvector, is the multiplication symbol, is the indicator variable, is the dimension of the topological eigenvector, for The symmetry group of is the i-th component of the topological eigenvector, represents the semi-direct product operation, is the direct product of the m-dimensional binary cyclic groups; The matrix embedding unit performs the following operations: Compute the matrix representation of the group ; Calculate the eigenvalue set , in, is the matrix representation of the group, For group theory characteristics, is the matrix exponential function, It means summation, is an element in group G, is the representation matrix of the group element g, for The set of eigenvalues ​​of is the characteristic value; The number theory transformation unit performs the following operations: Zeta function to calculate eigenvalues ; Calculate Lyapunov exponent , in, is the Zeta function of the eigenvalue, is the set of eigenvalues, is the multiplication symbol, is the indicator variable, is the number of eigenvalues, is the i-th eigenvalue, is a complex variable, is the Lyapunov exponent, is the limit symbol, For a variable that tends to infinity, is the natural logarithm, is the absolute value symbol, is the i-th order derivative of the Zeta function; The quantum probability prediction unit performs the following operations: Constructing quantum states ; Calculating probability distributions ; in, , is a quantum state, represents the summation sign, is the indicator variable, is the total number of possible states of the system, is the probability amplitude, is the probability that the measurement result is i, is the i-th ground state of the system, is a natural constant, is the inverse temperature of the system, is the energy eigenvalue, For the summation index, is the Lyapunov index corresponding to the i-th state.

2. The system according to claim 1, characterized in that The vibration data acquisition module comprises: A first sensor is used to obtain vibration data of a support device for mounting an electromechanical facility; A second sensor is used to obtain vibration data of the electromechanical facility body; The data aggregation unit is connected to the first sensor and the second sensor for aggregating the vibration data collected by the first sensor and the second sensor.

3. The system according to claim 1, characterized in that The data processing module comprises: A time-frequency conversion unit, used for converting the vibration data into a time-frequency domain; A parameter generating unit, configured to generate a vibration state parameter based on the time domain signal and the frequency domain signal; The vibration state parameters include peak factor, kurtosis index, impact index and root mean square value.

4. A digital operation and maintenance control method for electromechanical facilities based on the system according to any one of claims 1 to 3, characterized in that: The following steps are involved: Obtain vibration data of electromechanical facilities; Performing time-frequency domain conversion on the vibration data to obtain a time-domain signal and a frequency-domain signal; Generate a vibration state parameter based on the time domain signal and the frequency domain signal; Execute vibration data topology feature analysis and prediction system algorithms, including: Mapping the vibration state parameters to a high-dimensional topological space; Performing group theory analysis on the characteristics of the high-dimensional topological space; Embed group theory features into special matrices and perform spectral analysis; Perform number-theoretic transformation and chaos analysis on eigenvalues; Construct a quantum probability prediction model based on chaos analysis results; Generate operation and maintenance decision information based on the output results of the quantum probability prediction model; According to the operation and maintenance decision information, corresponding operation and maintenance control operations are performed.

Images