Analysis method and system for intelligent early warning of speed regulator oil pressing device

Abstract

The invention relates to the technical field of mechanical equipment monitoring and diagnosis, in particular to an analysis method and system for intelligent early warning of a speed regulator oil pressing device. The method comprises the following steps: collecting oil pressure, oil temperature and flow data of a speed regulator oil pressing device in real time; performing mode matching of real-time data by using a dynamic time warping method, detecting local abnormal data in combination with a local abnormal factor algorithm, identifying abnormal conditions of the device, and introducing a feedback loop to dynamically adjust a mode matching threshold value; according to a damped oscillation model principle, an oil pressure and flow dynamic model is constructed, relative changes of oil pressure and flow are analyzed, and device faults are analyzed in combination with system energy loss and a damping effect; and calculating a fault occurrence probability, carrying out grade division on risks, and generating early warning information. According to the intelligent early warning analysis method and system for the speed regulator oil pressing device, a dynamic time warping and local abnormal factor algorithm is introduced, and a damped oscillation model and a dynamic coupling model are combined to accurately early warn a fault.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical equipment monitoring and diagnosis, and in particular to an analysis method and system for intelligent early warning of a speed regulator oil pressure device. Background Art

[0002] An analysis method and system for intelligent early warning of a speed governor oil pressure device aims to improve the fault warning accuracy and response speed of the speed governor oil pressure device. By combining dynamic time warping with a local abnormal factor algorithm, key parameters such as oil pressure, oil temperature, and flow rate are monitored in real time and compared with historical data. The flexibility of control threshold adjustment and abnormal point detection is enhanced to timely identify potential risks before a fault occurs and automatically generate early warning information.

[0003] The existing analysis method and system of intelligent early warning for the speed governor oil pressure device usually find it difficult to accurately distinguish between normal fluctuations of the equipment and real fault signals. In addition, due to the complexity of the operating environment of the speed governor oil pressure device, such as changes in oil temperature and oil viscosity, and objective reasons such as aging and wear of mechanical parts, the traditional early warning method cannot effectively distinguish the difference between tiny oil pressure fluctuations and system failures in real time. Therefore, an analysis method and system of intelligent early warning for the speed governor oil pressure device are provided. Summary of the invention

[0004] The purpose of the present invention is to provide an analysis method and system for intelligent early warning of a speed regulator oil pressure device, so as to solve the problem raised in the above background technology that due to the complexity of the operating environment of the speed regulator oil pressure device, such as changes in oil temperature and oil viscosity, and aging and wear of mechanical parts, the traditional early warning method cannot effectively distinguish the difference between small oil pressure fluctuations and system failures in real time.

[0005] To achieve the above object, the present invention provides an analysis method for intelligent early warning of a speed governor oil pressure device, comprising:

[0006] S1, real-time collection of oil pressure, oil temperature and flow data of the speed governor oil pressure device, and pre-processing of the data;

[0007] S2. Use dynamic time warping method to perform pattern matching of real-time data, compare the similarity between real-time data and historical normal data, and combine local abnormal factor algorithm to detect local abnormal data, identify abnormal conditions of the device, and introduce feedback loop to dynamically adjust pattern matching thresholds and parameters;

[0008] S3. Based on the principle of damped oscillation model, a dynamic model of oil pressure and flow is constructed to analyze the relative changes of oil pressure and flow, and the device failure is analyzed in combination with system energy loss and damping effect;

[0009] S4. Based on the analysis results, calculate the probability of failure, classify the risks into levels and generate early warning information.

[0010] In S1, the oil pressure, oil temperature and flow rate data of the speed governor oil pressure device are collected in real time, and the data are pre-processed. The specific method steps are as follows:

[0011] S1.1. Deploy the sensor in the speed regulator oil pressure device to collect the oil pressure P of the speed regulator oil pressure device in real time. o , oil temperature T o and flow rate Q o and collects and transmits oil pressure P through PLC o , oil temperature T o and flow rate Q o data;

[0012] S1.2, collect the oil pressure P o , oil temperature T o and flow rate Q o Clean and standardize the data;

[0013] S1.3. Store the processed data in a time series database.

[0014] As a further improvement of the technical solution, the dynamic time warping method is a time series analysis method, which is used to measure the similarity between the historical data time series and the real-time data time series after alignment on the time axis to detect device anomalies;

[0015] The local anomaly factor algorithm is a density-based anomaly detection algorithm used to identify local anomalies in a data set to identify device anomalies.

[0016] As a further improvement of the technical solution, in S2, a dynamic time warping method is used to perform pattern matching of real-time data, compare the similarity between real-time data and historical normal data, and combine the local abnormal factor algorithm to detect local abnormal data, identify device abnormalities, and introduce a feedback loop to dynamically adjust the pattern matching threshold and parameters. The specific steps are as follows:

[0017] S2.1. Obtain oil pressure P from the time series database o , oil temperature T o and flow rate Q o Real-time data and historical normal data;

[0018] S2.2, set the DTW threshold, construct the real-time data sequence and the historical normal data sequence, use the dynamic time warping method to perform pattern matching on the real-time data sequence and the historical normal data sequence, calculate the DTW distance, and compare and analyze the DTW distance with the DTW threshold;

[0019] S2.3, set the LOF threshold, use the local anomaly factor algorithm to detect the local reachable density and local anomaly factor of each real-time data, and compare and analyze the local anomaly factor with the LOF threshold;

[0020] S2.4, based on the analysis results of S2.2 and S2.3, determine the abnormality of the abnormal point identification device;

[0021] S2.5. Introduce a feedback loop to dynamically adjust the DTW threshold and LOF threshold.

[0022] As a further improvement of the technical solution, in S2.2, a DTW threshold is set, a real-time data sequence and a historical normal data sequence are constructed, a dynamic time warping method is used to perform pattern matching on the real-time data sequence and the historical normal data sequence, the DTW distance is calculated, and the DTW distance is compared and analyzed with the DTW threshold. The specific method steps are as follows:

[0023] S2.2.1. Set the DTW threshold to α DTW , construct real-time data series and historical normal data series:

[0024] Real-time data series:

[0025] {P o (t),T o (t),Q o (t)},t=1,2,…,T;

[0026] Historical normal data series:

[0027] {P hist (t),T hist (t),Q hist (t)},t=1,2,…,T;

[0028] Where t is the time step; T is the length of the time series; P hist (t) is the historical normal oil pressure data; T hist (t) is the historical normal oil temperature data; Q hist (t) is the historical normal traffic data;

[0029] S2.2.2. Define the DTW distance D(t) as:

[0030]

[0031] S2.2.3. Constructing the cumulative distance matrix D i,j , to align the real-time data series and the historical normal data series:

[0032] D i,j =d(i,j)+min(D i-1,j,D i,j-1 ,D i-1,j-1 );

[0033] Where i represents the i-th time step; j represents the j-th time step; D i,j is the cumulative distance matrix, which represents the cumulative distance between the i-th time point in the real-time data sequence and the j-th time point in the historical normal data sequence. i-1,j ,D i,j-1 ,D i-1,j-1 And so on; d(i,j) is the local distance measure between the i-th point in the real-time data sequence and the j-th point in the historical normal data sequence;

[0034] S2.2.4. The DTW distance D(t) is finally obtained by recursive calculation. The DTW distance D(t) and the DTW threshold α are DTW For comparison:

[0035] If D(t)>α DTW , it means that the real-time data is significantly different from the historical normal data and there is an anomaly.

[0036] As a further improvement of the technical solution, in S2.3, the LOF threshold is set, the local anomaly factor algorithm is used to detect the local reachable density and local anomaly factor of each real-time data, and the local anomaly factor is compared and analyzed with the LOF threshold. The specific steps are as follows:

[0037] S2.3.1. Set the LOF threshold to α LOF ;

[0038] S2.3.2. Calculate the data point x in real-time data i With data point x j The reachable distance:

[0039] l k,reach (x i ,x j )=max(dist(x i ,x j ),dist(x j ,k th_nearest (x i )));

[0040] Among them, l k,reach (x i ,x j ) is the data point x i With data point x j The reachable distance; dist(x i ,x j ) is the data point x i With data point x jThe Euclidean distance of k is the data point x. i The number of nearest neighbors; k th_nearest (x i ) is the data point x i The kth nearest neighbor of ;

[0041] S2.3.3. Calculate the data point x i The local reachable density of is:

[0042]

[0043] Among them, lrd k (x i ) is the data point x i The local reachable density of N k (x i ) is the data point x i The nearest neighbor set of ;

[0044] S2.3.4. Using data point x i The local reachable density of the data point x is calculated i Local anomaly factor:

[0045]

[0046] Among them, LOF k (x i ) is the data point x i Local outlier factor; lrd k (x j ) is the data point x j The local reachable density of

[0047] S2.3.4. Compare and analyze the local anomaly factor with the LOF threshold:

[0048] If LOF k (x i )>α LOF , then it means that the data point x i There is an anomaly.

[0049] As a further improvement of the technical solution, in S2.4, the analysis results of S2.2 and S2.3 are combined to determine the abnormal situation of the abnormal point identification device. The specific method is as follows:

[0050] If D(t)>α DTW , and LOF k (x i )>α LOF , then the data point x i It is an outlier point;

[0051] When a data point is determined to be an abnormal point, the system will trigger an alarm mechanism, generate early warning information, and record the time and type of the abnormality;

[0052] S2.5. Introduce a feedback loop strategy to dynamically adjust the DTW threshold and LOF threshold. The feedback loop strategy is as follows:

[0053] Dynamically adjust the DTW threshold when abnormal patterns occur frequently in consecutive time steps:

[0054] α DTW,new =α DTW ×(1+δ);

[0055] Among them, δ is the DTW threshold adjustment coefficient; α DTW,new is the dynamically adjusted DTW threshold;

[0056] When the system detects large data fluctuations, it dynamically adjusts the neighborhood size k:

[0057] k new = k × (1 + λ);

[0058] Among them, λ is the neighborhood adjustment coefficient; k new is the dynamically adjusted neighborhood size.

[0059] As a further improvement of the technical solution, in S3, according to the principle of the damped oscillation model, an oil pressure flow dynamic model is constructed to analyze the relative changes of oil pressure and flow, and the device failure is analyzed in combination with the system energy loss and damping effect. The specific method steps are as follows:

[0060] S3.1. Construct the oil pressure flow dynamic model based on the damped oscillation model principle:

[0061] Damped Oscillation Model:

[0062]

[0063] Among them, in the original damped oscillation model, m is the system mass; c is the damping coefficient; k′ is the stiffness coefficient; x(t1) is the system displacement; F(t1) is the external force; t1 is the time in the original damped oscillation model;

[0064] Use oil pressure P o (t) replaces the system displacement x(t1), the flow rate Q o (t) replaces the external force F(t1) and constructs the oil pressure flow dynamic model:

[0065] Oil pressure flow dynamic model:

[0066]

[0067] S3.2, using the dynamic coupling model to describe the coupling relationship between the two:

[0068]

[0069] Among them, α1 is the dynamic adjustment parameter of oil pressure; α2 is the dynamic adjustment parameter of flow rate; β1 and β2 are the mutual coupling coefficients between flow rate and oil pressure;

[0070] S3.3. Calculate the energy loss of the system:

[0071]

[0072] Where ΔE(t-1,t) is the energy loss of the system from time step t-1 to time step t;

[0073] S3.4. In the oil pressure flow dynamic model, monitor the change of the damping coefficient c:

[0074] If the damping coefficient c changes greatly, it indicates that the device is faulty;

[0075] In the dynamic coupling model, the changes of the mutual coupling coefficients β1 and β2 between flow rate and oil pressure are monitored;

[0076] Monitor the change of energy loss ΔE(t-1,t):

[0077] When the energy loss ΔE(t-1,t) increases, it indicates that the energy consumption of the system increases, and there are problems such as increased flow resistance, increased fluid viscosity, and pipe blockage;

[0078] When the energy loss ΔE(t-1,t) decreases, it indicates that the efficiency of the system has decreased, and there are problems with pump damage and poor sealing.

[0079] As a further improvement of the technical solution, in S4, the probability of failure is calculated according to the analysis results, and the risks are graded and early warning information is generated. The specific method steps are as follows:

[0080] S4.1. Calculate the failure probability based on local anomaly factor detection:

[0081]

[0082] Among them, P fault (t) is the failure probability based on local abnormal factor detection;

[0083] S4.2. Setting energy loss threshold α E , calculate the energy loss ΔE(t-1,t) exceeds the energy loss threshold α E Probability of failure:

[0084]

[0085] Among them, P(ΔE(t-1,t)>α E ) is the energy loss ΔE(t-1,t) exceeding the energy loss threshold α E The failure probability of E is the energy loss threshold; μ ΔE is the expected value of energy loss; σ ΔE is the standard deviation of energy loss; Φ is the cumulative distribution function of the standard normal distribution;

[0086] S4.3. Failure probability P based on local abnormality factor detection fault (t) and the energy loss ΔE(t-1,t) exceeds the energy loss threshold α E The failure probability P(ΔE(t-1,t)>α E ), determine the risk level and issue an early warning:

[0087] when The device is judged to be low risk, and the warning message is "the device is operating normally";

[0088] If the device is judged to be of medium risk, the warning message is "The device status is abnormal, there is a certain risk, please check regularly";

[0089] when The device is judged to be high risk, and the warning message is "The risk of equipment failure is high, please check the equipment immediately and take safety measures."

[0090] On the other hand, the present invention provides an analysis system for intelligent early warning of a speed regulator oil pressure device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned analysis method for intelligent early warning of a speed regulator oil pressure device.

[0091] Compared with the prior art, the present invention has the following beneficial effects:

[0092] 1. The analysis method and system of the intelligent early warning of the speed regulator oil pressure device, based on the combination of dynamic time warping and local abnormal factor algorithm, can accurately identify small abnormal fluctuations in multiple real-time monitoring parameters such as oil pressure, oil temperature, and flow, and compare and analyze with historical data to distinguish the normal operating status of the equipment from potential failure modes.

[0093] 2. In the analysis method and system of the intelligent early warning of the speed regulator oil pressure device, the changes in oil pressure and flow are analyzed by combining the damped oscillation model with the dynamic coupling model, so as to realize real-time monitoring and prediction based on the dynamic response characteristics of the system, accurately identify abnormal fluctuations in oil pressure and flow, and timely warn of potential faults through the model coupling relationship. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 The figure is a flow chart of the overall method of the present invention. DETAILED DESCRIPTION

[0095] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0096] Embodiment 1:

[0097] See also Figure 1 As shown, this embodiment provides an analysis method for intelligent early warning of a speed regulator oil pressure device, comprising the following steps:

[0098] S1, real-time collection of oil pressure, oil temperature and flow data of the speed governor oil pressure device, and pre-processing of the data;

[0099] In this embodiment S1, the oil pressure, oil temperature and flow rate data of the speed governor oil pressure device are collected in real time, and the data are pre-processed. The specific method steps are as follows:

[0100] S1.1. Deploy the sensor in the speed regulator oil pressure device to collect the oil pressure P of the speed regulator oil pressure device in real time. o , oil temperature T o and flow rate Q o and collects and transmits oil pressure P through PLC o , oil temperature T o and flow rate Q o data;

[0101] In this embodiment, the sensors include pressure sensors, temperature sensors and flow sensors, which are respectively installed at the oil pump, pressure regulating valve, hydraulic cylinder and oil pipe position of the speed governor oil pressure device; PLC is an industrial automation computer control system used to monitor, control and automate various industrial processes, mechanical operations and production lines. PLC uses programmed logic to achieve precise control of equipment, machines and production processes; the data collection frequency is usually between a few seconds and a few minutes, which is set according to system requirements;

[0102] S1.2, collect the oil pressure P o , oil temperature T o and flow rate Q o Clean and standardize the data;

[0103] In this embodiment, an outlier removal method is used to filter out noise and invalid data; a normalization method is used to standardize the collected data to eliminate the dimensional differences between different sensors and different units;

[0104] S1.3. Store the processed data in a time series database.

[0105] In this embodiment, a time series database is a database used to store and query time series data. It is a data set arranged in chronological order, which is continuous in time, and each piece of data is associated with a specific timestamp; each piece of data has a clear timestamp indicating the time when the data occurred. The timestamp is usually the primary key to ensure the order and uniqueness of the data.

[0106] S2. Use dynamic time warping method to perform pattern matching of real-time data, compare the similarity between real-time data and historical normal data, and combine local abnormal factor algorithm to detect local abnormal data, identify abnormal conditions of the device, and introduce feedback loop to dynamically adjust pattern matching thresholds and parameters;

[0107] The dynamic time warping method is a time series analysis method used to measure the similarity between the time series of historical data and the time series of real-time data after alignment on the time axis to detect device anomalies;

[0108] The local anomaly factor algorithm is a density-based anomaly detection algorithm used to identify local anomalies in a data set to identify device anomalies.

[0109] In this embodiment, the dynamic time warping method is combined with the local anomaly factor algorithm to more comprehensively discover the failure risk of the speed regulator oil pressure device. The dynamic time warping method can identify patterns with high similarity through global comparison of time series, but cannot accurately locate local anomalies, especially when the anomaly only manifests as a small fluctuation at a certain point in time. The local anomaly factor algorithm can effectively identify local anomalies through anomaly detection based on local density; the dynamic time warping method provides global pattern matching, and the local anomaly factor algorithm captures local anomalies to improve the accuracy of overall anomaly recognition.

[0110] In this embodiment S2, a dynamic time warping method is used to perform pattern matching of real-time data, compare the similarity between real-time data and historical normal data, and combine the local abnormal factor algorithm to detect local abnormal data, identify device abnormalities, and introduce a feedback loop to dynamically adjust the pattern matching threshold and parameters. The specific steps are as follows:

[0111] S2.1. Obtain oil pressure P from the time series database o , oil temperature T o and flow rate Q oReal-time data and historical normal data;

[0112] S2.2, set the DTW threshold, construct the real-time data sequence and the historical normal data sequence, use the dynamic time warping method to perform pattern matching on the real-time data sequence and the historical normal data sequence, calculate the DTW distance, and compare and analyze the DTW distance with the DTW threshold;

[0113] S2.3, set the LOF threshold, use the local anomaly factor algorithm to detect the local reachable density and local anomaly factor of each real-time data, and compare and analyze the local anomaly factor with the LOF threshold;

[0114] S2.4, based on the analysis results of S2.2 and S2.3, determine the abnormality of the abnormal point identification device;

[0115] S2.5. Introduce a feedback loop to dynamically adjust the DTW threshold and LOF threshold.

[0116] In S2.2, the DTW threshold is set, the real-time data sequence and the historical normal data sequence are constructed, the dynamic time warping method is used to perform pattern matching on the real-time data sequence and the historical normal data sequence, the DTW distance is calculated, and the DTW distance is compared and analyzed with the DTW threshold. The specific method steps are as follows:

[0117] S2.2.1. Set the DTW threshold to α DTW , construct real-time data series and historical normal data series:

[0118] Real-time data series:

[0119] {P o (t),T o (t),Q o (t)},t=1,2,…,T;

[0120] Historical normal data series:

[0121] {P hist (t),T hist (t),Q hist (t)},t=1,2,…,T;

[0122] Where t is the time step; T is the length of the time series; P hist (t) is the historical normal oil pressure data; T hist (t) is the historical normal oil temperature data; Q hist (t) is the historical normal traffic data;

[0123] S2.2.2. Define the DTW distance D(t) as:

[0124]

[0125] S2.2.3. Constructing the cumulative distance matrix D i,j , to align the real-time data series and the historical normal data series:

[0126] D i,j =d(i,j)+min(D i-1,j ,D i,j-1 ,D i-1,j-1 );

[0127] Where i represents the i-th time step; j represents the j-th time step; D i,j is the cumulative distance matrix, which represents the cumulative distance between the i-th time point in the real-time data sequence and the j-th time point in the historical normal data sequence. i-1,j ,D i,j-1 ,D i-1,j-1 And so on; d(i,j) is the local distance measure between the i-th point in the real-time data sequence and the j-th point in the historical normal data sequence;

[0128] In this embodiment, the cumulative distance matrix D i,j Represents the cumulative distance matrix element in dynamic time warping, which is used to record the minimum cumulative distance from i to j when matching two time series;

[0129] S2.2.4. The DTW distance D(t) is finally obtained by recursive calculation. The DTW distance D(t) and the DTW threshold α are DTW For comparison:

[0130] If D(t)>α DTW , it means that the real-time data is significantly different from the historical normal data and there is an anomaly.

[0131] In S2.3, the LOF threshold is set, and the local anomaly factor algorithm is used to detect the local reachable density and local anomaly factor of each real-time data. The local anomaly factor is compared and analyzed with the LOF threshold. The specific steps are as follows:

[0132] S2.3.1. Set the LOF threshold to α LOF ;

[0133] S2.3.2. Calculate the data point x in real-time data i With data point x j The reachable distance:

[0134] l k,reach (x i ,x j )=max(dist(x i ,x j ),dist(x j,k th_nearest (x i )));

[0135] Among them, l k,reach (x i ,x j ) is the data point x i With data point x j The reachable distance; dist(x i ,x j ) is the data point x i With data point x j The Euclidean distance of k is the data point x. i The number of nearest neighbors; k th_nearest (x i ) is the data point x i The kth nearest neighbor of ;

[0136] S2.3.3. Calculate the data point x i The local reachable density of is:

[0137]

[0138] Among them, lrd k (x i ) is the data point x i The local reachable density of N k (x i ) is the data point x i The nearest neighbor set of ;

[0139] S2.3.4. Using data point x i The local reachable density of the data point x is calculated i Local anomaly factor:

[0140]

[0141] Among them, LOF k (x i ) is the data point x i Local outlier factor; lrd k (x j ) is the data point x j The local reachable density of

[0142] S2.3.4. Compare and analyze the local anomaly factor with the LOF threshold:

[0143] If LOF k (x i )>α LOF , then it means that the data point x i There is an anomaly.

[0144] S2.4. Based on the analysis results of S2.2 and S2.3, determine the abnormality of the abnormal point identification device. The specific method is as follows:

[0145] If D(t)>α DTW , and LOF k (x i )>α LOF , then the data point x i It is an outlier point;

[0146] When a data point is determined to be an abnormal point, the system will trigger an alarm mechanism, generate early warning information, and record the time and type of the abnormality;

[0147] S2.5. Introduce a feedback loop strategy to dynamically adjust the DTW threshold and LOF threshold. The feedback loop strategy is as follows:

[0148] Dynamically adjust the DTW threshold when abnormal patterns occur frequently in consecutive time steps:

[0149] α DTW,new =α DTW ×(1+δ);

[0150] Among them, δ is the DTW threshold adjustment coefficient; α DTW,new is the dynamically adjusted DTW threshold;

[0151] When the system detects large data fluctuations, it dynamically adjusts the neighborhood size k:

[0152] k new = k × (1 + λ);

[0153] Among them, λ is the neighborhood adjustment coefficient; k new is the dynamically adjusted neighborhood size.

[0154] S3. Based on the principle of damped oscillation model, a dynamic model of oil pressure and flow is constructed to analyze the relative changes of oil pressure and flow, and the device failure is analyzed in combination with system energy loss and damping effect;

[0155] In this embodiment S3, according to the damped oscillation model principle, an oil pressure flow dynamic model is constructed to analyze the relative changes of oil pressure and flow, and the device failure is analyzed in combination with system energy loss and damping effect. The specific method steps are as follows:

[0156] S3.1. Construct the oil pressure flow dynamic model based on the damped oscillation model principle:

[0157] Damped Oscillation Model:

[0158]

[0159] Among them, in the original damped oscillation model, m is the system mass; c is the damping coefficient; k′ is the stiffness coefficient; x(t1) is the system displacement; F(t1) is the external force; t1 is the time in the original damped oscillation model;

[0160] Use oil pressure P o (t) replaces the system displacement x(t1), the flow rate Q o (t) replaces the external force F(t1) and constructs the oil pressure flow dynamic model:

[0161] Oil pressure flow dynamic model:

[0162]

[0163] S3.2, using the dynamic coupling model to describe the coupling relationship between the two:

[0164]

[0165]

[0166] Among them, α1 is the dynamic adjustment parameter of oil pressure; α2 is the dynamic adjustment parameter of flow rate; β1 and β2 are the mutual coupling coefficients between flow rate and oil pressure;

[0167] S3.3. Calculate the energy loss of the system:

[0168]

[0169] Where ΔE(t-1,t) is the energy loss of the system from time step t-1 to time step t;

[0170] In this embodiment, the energy loss of the system is in the form of energy dissipation caused by the damping effect;

[0171] S3.4. In the oil pressure flow dynamic model, monitor the change of the damping coefficient c:

[0172] If the damping coefficient c changes greatly, it indicates that the device is faulty;

[0173] In this embodiment, if the damping coefficient c becomes larger, it indicates that the device has a fault such as partial blockage in the pipeline or valve, or increased fluid viscosity; if the damping coefficient c becomes smaller, it indicates that the device has a fault such as decreased pumping capacity, system leakage, and poor sealing;

[0174] In the dynamic coupling model, the changes of the mutual coupling coefficients β1 and β2 between flow rate and oil pressure are monitored;

[0175] The increase of coupling coefficient β1 indicates that the influence of flow rate change on oil pressure increases, and there is a valve or pipeline wear fault;

[0176] The coupling coefficient β1 decreases, indicating that the oil pressure's ability to control the flow rate is weakened, and there are problems such as reduced pump efficiency or poor system sealing;

[0177] Monitor the change of energy loss ΔE(t-1,t):

[0178] When the energy loss ΔE(t-1,t) increases, it indicates that the energy consumption of the system increases, and there are problems such as increased flow resistance, increased fluid viscosity, and pipe blockage;

[0179] When the energy loss ΔE(t-1,t) decreases, it indicates that the efficiency of the system has decreased, and there are problems with pump damage and poor sealing.

[0180] S4. Calculate the probability of failure based on the analysis results, classify the risks and generate early warning information;

[0181] In this embodiment S4, according to the analysis results, the probability of failure occurrence is calculated, the risks are graded and early warning information is generated. The specific method steps are as follows:

[0182] S4.1. Calculate the failure probability based on local anomaly factor detection:

[0183]

[0184] Among them, P fault (t) is the failure probability based on local abnormal factor detection;

[0185] S4.2. Setting energy loss threshold α E , calculate the energy loss ΔE(t-1,t) exceeds the energy loss threshold α E Probability of failure:

[0186]

[0187] Among them, P(ΔE(t-1,t)>α E ) is the energy loss ΔE(t-1,t) exceeding the energy loss threshold α E The failure probability of E is the energy loss threshold; μ ΔE is the expected value of energy loss; σ ΔE is the standard deviation of energy loss; Φ is the cumulative distribution function of the standard normal distribution;

[0188] S4.3. Failure probability P based on local abnormality factor detection fault (t) and the energy loss ΔE(t-1,t) exceeds the energy loss threshold α E The failure probability P(ΔE(t-1,t)>α E ), determine the risk level and issue an early warning:

[0189] when The device is judged to be low risk, and the warning message is "the device is operating normally";

[0190] If the device is judged to be of medium risk, the warning message is "The device status is abnormal, there is a certain risk, please check regularly";

[0191] when The device is judged to be high risk, and the warning message is "The risk of equipment failure is high, please check the equipment immediately and take safety measures."

[0192] Embodiment 2:

[0193] This embodiment provides an analysis system for intelligent early warning of a speed regulator oil pressure device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the above-mentioned analysis method for intelligent early warning of a speed regulator oil pressure device.

[0194] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only preferred examples of the present invention, and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. An analysis method for intelligent early warning of a speed governor oil pressure device, characterized in that: The following steps are involved: S1, real-time collection of oil pressure, oil temperature and flow data of the speed governor oil pressure device, and pre-processing of the data; S2. Use dynamic time warping method to perform pattern matching of real-time data, compare the similarity between real-time data and historical normal data, and combine local abnormal factor algorithm to detect local abnormal data, identify abnormal conditions of the device, and introduce feedback loop to dynamically adjust pattern matching thresholds and parameters; S3. Based on the principle of damped oscillation model, a dynamic model of oil pressure and flow is constructed to analyze the relative changes of oil pressure and flow, and the device failure is analyzed in combination with system energy loss and damping effect; S4. Based on the analysis results, calculate the probability of failure, classify the risks into levels and generate early warning information.

2. The analysis method of intelligent early warning of the speed governor oil pressure device according to claim 1 is characterized by: In S1, the oil pressure, oil temperature and flow rate data of the speed regulator oil pressure device are collected in real time, and the data are pre-processed. The specific method steps are as follows: S1.

1. Deploy the sensor in the speed regulator oil pressure device to collect the oil pressure P of the speed regulator oil pressure device in real time. o , oil temperature T o and flow rate Q o and collects and transmits oil pressure P through PLC o , oil temperature T o and flow rate Q o data; S1.2, collect the oil pressure P o , oil temperature T o and flow rate Q o Clean and standardize the data; S1.

3. Store the processed data in a time series database.

3. The analysis method of intelligent early warning of the speed governor oil pressure device according to claim 2 is characterized by: The dynamic time warping method is a time series analysis method used to measure the similarity between the historical data time series and the real-time data time series after alignment on the time axis to detect device anomalies; The local anomaly factor algorithm is a density-based anomaly detection algorithm used to identify local anomalies in a data set to identify device anomalies.

4. The analysis method of intelligent early warning of the speed governor oil pressure device according to claim 3 is characterized by: In S2, a dynamic time warping method is used to perform pattern matching of real-time data, compare the similarity between real-time data and historical normal data, and combine the local abnormal factor algorithm to detect local abnormal data, identify device abnormalities, and introduce a feedback loop to dynamically adjust the pattern matching threshold and parameters. The specific steps are as follows: S2.

1. Obtain oil pressure P from the time series database o , oil temperature T o and flow rate Q o Real-time data and historical normal data; S2.2, set the DTW threshold, construct the real-time data sequence and the historical normal data sequence, use the dynamic time warping method to perform pattern matching on the real-time data sequence and the historical normal data sequence, calculate the DTW distance, and compare and analyze the DTW distance with the DTW threshold; S2.3, set the LOF threshold, use the local anomaly factor algorithm to detect the local reachable density and local anomaly factor of each real-time data, and compare and analyze the local anomaly factor with the LOF threshold; S2.4, based on the analysis results of S2.2 and S2.3, determine the abnormality of the abnormal point identification device; S2.

5. Introduce a feedback loop to dynamically adjust the DTW threshold and LOF threshold.

5. The analysis method of intelligent early warning of the speed governor oil pressure device according to claim 4 is characterized in that: In S2.2, a DTW threshold is set, a real-time data sequence and a historical normal data sequence are constructed, a dynamic time warping method is used to perform pattern matching on the real-time data sequence and the historical normal data sequence, a DTW distance is calculated, and the DTW distance is compared and analyzed with the DTW threshold. The specific method steps are as follows: S2.2.

1. Set the DTW threshold to α DTE , construct real-time data series and historical normal data series: Real-time data series: {P o (t),T o (t),Q o (t)},t=1,2,…,T; Historical normal data series: {P hist (t),T hist (t),Q hist (t)},t=1,2,…,T; Where t is the time step; T is the length of the time series; P hist (t) is the historical normal oil pressure data; T hist (t) is the historical normal oil temperature data; Q hist (t) is the historical normal traffic data; S2.2.

2. Define the DTW distance D(t) as: S2.2.

3. Constructing the cumulative distance matrix D i,j , to align the real-time data series and the historical normal data series: D i,j =d(i,j)+min(D i-1,j ,D i,j-1 ,D i-1,j-1 ); Where i represents the i-th time step; j represents the j-th time step; D i,j is the cumulative distance matrix, which represents the cumulative distance between the i-th time point in the real-time data sequence and the j-th time point in the historical normal data sequence. i-1,j ,D i,j-1 ,D i-1,j-1 And so on; d(i,j) is the local distance measure between the i-th point in the real-time data sequence and the j-th point in the historical normal data sequence; S2.2.

4. The DTW distance D(t) is finally obtained by recursive calculation. The DTW distance D(t) and the DTW threshold α are DTW For comparison: If D(t)>α DTW , it means that the real-time data is significantly different from the historical normal data and there is an anomaly.

6. The analysis method of intelligent early warning of the speed governor oil pressure device according to claim 5 is characterized by: In S2.3, the LOF threshold is set, and the local anomaly factor algorithm is used to detect the local reachable density and local anomaly factor of each real-time data, and the local anomaly factor is compared and analyzed with the LOF threshold. The specific steps are as follows: S2.3.

1. Set the LOF threshold to α LOF ; S2.3.

2. Calculate the data point x in real-time data i With data point x j The reachable distance: l k,reach (x i ,x j )=max(dist(x i ,x j ),dist(x j ,k th_nearest (x i ))); Among them, l k,reach (x i ,x j ) is the data point x i With data point x j The reachable distance; dist(x i ,x j ) is the data point x i With data point x j The Euclidean distance of k is the data point x. i The number of nearest neighbors; k th_nearest (x i ) is the data point x i The kth nearest neighbor of ; S2.3.

3. Calculate the data point x i The local reachable density of is: Among them, lrd k (x i ) is the data point x i The local reachable density of N k (x i ) is the data point x i The nearest neighbor set of ; S2.3.

4. Using data point x i The local reachable density of the data point x is calculated i Local anomaly factor: Among them, LOF k (x i ) is the data point x i Local outlier factor; lrd k (x j ) is the data point x j The local reachable density of S2.3.

4. Compare and analyze the local anomaly factor with the LOF threshold: If LOF k (x i )>α LOF , then it means that the data point x i There is an anomaly.

7. The analysis method for intelligent early warning of the speed governor oil pressure device according to claim 6 is characterized by: In S2.4, the analysis results of S2.2 and S2.3 are combined to determine the abnormal situation of the abnormal point identification device. The specific method is as follows: If D(t)>α DTW , and LOF k (x i )>α LOF , then the data point x i It is an outlier point; When a data point is determined to be an abnormal point, the system will trigger an alarm mechanism, generate early warning information, and record the time and type of the abnormality; S2.

5. Introduce a feedback loop strategy to dynamically adjust the DTW threshold and LOF threshold. The feedback loop strategy is as follows: Dynamically adjust the DTW threshold when abnormal patterns occur frequently in consecutive time steps: a DTW,new =a DTW ×(1+δ); Among them, δ is the DTW threshold adjustment coefficient; α DTW,new is the dynamically adjusted DTW threshold; When the system detects large data fluctuations, it dynamically adjusts the neighborhood size k: k new =k×(1+λ); Among them, λ is the neighborhood adjustment coefficient; k new is the dynamically adjusted neighborhood size.

8. The analysis method for intelligent early warning of the speed governor oil pressure device according to claim 7 is characterized by: In S3, according to the damped oscillation model principle, an oil pressure flow dynamic model is constructed to analyze the relative changes of oil pressure and flow, and the device failure is analyzed in combination with system energy loss and damping effect. The specific method steps are as follows: S3.

1. Construct the oil pressure flow dynamic model based on the damped oscillation model principle: Damped Oscillation Model: Among them, in the original damped oscillation model, m is the system mass; c is the damping coefficient; k′ is the stiffness coefficient; x(t1) is the system displacement; F(t1) is the external force; t1 is the time in the original damped oscillation model; Use oil pressure P o (t) replaces the system displacement x(t1), the flow rate Q o (t) replaces the external force F(t1) and constructs the oil pressure flow dynamic model: Oil pressure flow dynamic model: S3.2, using the dynamic coupling model to describe the coupling relationship between the two: Among them, α1 is the dynamic adjustment parameter of oil pressure; α2 is the dynamic adjustment parameter of flow rate; β1 and β2 are the mutual coupling coefficients between flow rate and oil pressure; S3.

3. Calculate the energy loss of the system: Where ΔE(t-1,t) is the energy loss of the system from time step t-1 to time step t; S3.

4. In the oil pressure flow dynamic model, monitor the change of the damping coefficient c: If the damping coefficient c changes greatly, it indicates that the device is faulty; In the dynamic coupling model, the changes of the mutual coupling coefficients β1 and β2 between flow rate and oil pressure are monitored; Monitor the change of energy loss ΔE(t-1,t): When the energy loss ΔE(t-1,t) increases, it indicates that the energy consumption of the system increases, and there are problems such as increased flow resistance, increased fluid viscosity, and pipe blockage; When the energy loss ΔE(t-1,t) decreases, it indicates that the efficiency of the system has decreased, and there are problems with pump damage and poor sealing.

9. The analysis method for intelligent early warning of the speed governor oil pressure device according to claim 8 is characterized by: In S4, according to the analysis results, the probability of failure is calculated, the risks are graded and early warning information is generated. The specific method steps are as follows: S4.

1. Calculate the failure probability based on local anomaly factor detection: Among them, P fault (t) is the failure probability based on local abnormal factor detection; S4.

2. Setting energy loss threshold α E , calculate the energy loss ΔE(t-1,t) exceeds the energy loss threshold α E Probability of failure: Among them, P(ΔE(t-1,t)>α E ) is the energy loss ΔE(t-1,t) exceeding the energy loss threshold α E The failure probability of E is the energy loss threshold; μ ΔE is the expected value of energy loss; σ ΔE is the standard deviation of energy loss; v is the cumulative distribution function of the standard normal distribution; S4.

3. Failure probability P based on local abnormality factor detection fault (t) and the energy loss ΔE(t-1,t) exceeds the energy loss threshold α E The failure probability P(ΔE(t-1,t)>α E ), determine the risk level and issue an early warning: when The device is judged to be low risk, and the warning message is "the device is operating normally"; If the device is judged to be of medium risk, the warning message is "device status is abnormal, there is a certain risk, please check regularly"; when The device is judged to be high risk, and the warning message is "The risk of equipment failure is high, please check the equipment immediately and take safety measures." 10. An analysis system for intelligent early warning of a speed governor oil pressure device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes a computer program to implement the analysis method for intelligent early warning of the speed regulator oil pressure device as described in any one of claims 1-9.

Images