Predictive maintenance system for industrial internet equipment

Abstract

The invention relates to the technical field of data processing, in particular to a predictive maintenance system for industrial internet equipment, which comprises a data acquisition module used for acquiring operation data of the industrial internet equipment through a plurality of sensors; the data processing module comprises a data conversion unit for converting the preprocessed operation data; the prediction module comprises a model training unit for training a machine learning model according to the feature data; the storage module is used for storing the operation data, the feature data, the prediction result and the maintenance plan; the alarm module is used for giving an alarm when the industrial internet equipment breaks down; and the control module is used for adjusting the acquisition frequency of the operation data when it is judged that the predicted maintenance stability does not meet the requirement according to the variance of the fluctuation amplitude of the surface temperature. The present invention achieves an improvement in the stability of predictive maintenance.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a predictive maintenance system for industrial Internet equipment. Background Art

[0002] In the existing technology, with the development of industrial Internet, engineering manufacturing equipment has become more efficient and intelligent, but the maintenance of the equipment itself still faces huge challenges. Post-failure maintenance and regular maintenance in the traditional maintenance mode will affect production efficiency and product quality, and significantly increase the cost of manufacturers. With the maturity of technologies such as the Internet of Things, big data, cloud computing, machine learning and sensors, predictive maintenance technology has emerged.

[0003] Chinese Patent Publication No.: CN116681426A discloses a predictive maintenance system and method for industrial Internet equipment, including: an information collection module, an information screening module, a feature data calculation module, an equipment status database, a prediction module, and a comparison module; the information collection module collects information data of industrial Internet equipment; the information screening module performs effective information screening on the collected information data of industrial Internet equipment to obtain effective industrial Internet equipment data; the feature data calculation module performs feature calculation on the effective data of the screened industrial Internet equipment according to different types of data to obtain feature data of the current state; the equipment status database is used to store all reasonable status data of all industrial Internet equipment in normal operating conditions; the prediction module predicts and judges the status of the industrial Internet equipment according to the calculation results of the feature data to obtain the predicted status of the equipment; the comparison module compares the predicted status data of the current equipment obtained after processing by the prediction module with the corresponding reasonable status data of the equipment in the equipment status database, and makes a decision based on the comparison and judgment results, and further performs predictive maintenance of the industrial Internet equipment at a corresponding level. It can be seen from this that the industrial Internet equipment predictive maintenance system and method has the problem that the collected data is biased due to measurement errors in the sensors, which causes deviations in the distribution of data during the conversion process, and then causes problems in the data conversion process, thereby resulting in a decrease in the stability of predictive maintenance. Summary of the invention

[0004] To this end, the present invention provides a predictive maintenance system for industrial Internet equipment to overcome the problem in the prior art that measurement errors in sensors lead to deviations in the collected data, resulting in deviations in the distribution of data during the conversion process, and further leading to problems in the data conversion process, thereby resulting in a decrease in the stability of predictive maintenance.

[0005] To achieve the above-mentioned purpose, the present invention provides a predictive maintenance system for industrial Internet equipment, comprising: a data acquisition module, used to collect operating data of industrial Internet equipment through a number of sensors; a data processing module, which is connected to the data acquisition module, including a data conversion unit for converting preprocessed operating data to output feature data and a data retransmission unit connected to the data conversion unit for retransmitting when the feature data is lost; a prediction module, which is connected to the data processing unit, including a model training unit connected to the data conversion unit for training a machine learning model according to the feature data to output a prediction result and a plan generation unit for generating a maintenance plan according to the identified fault of the industrial Internet equipment; a storage module, which is respectively connected to the data processing unit, and the data conversion unit is connected to the data conversion unit. The data acquisition module, the data processing module and the prediction module are connected to each other for storing the operation data, the characteristic data, the prediction results and the maintenance plan respectively; the alarm module is connected to the prediction module for issuing an alarm when the industrial Internet device fails; the control module is respectively connected to the data acquisition module, the data processing module, the prediction module, the storage module and the alarm module for adjusting the collection frequency of the operation data or the data amount of a single training data according to the average number of alarms within several prediction cycles when it is determined that the stability of the predicted maintenance does not meet the requirements according to the variance of the fluctuation amplitude of the surface temperature, and adjusting the number of retransmissions of the lost characteristic data in combination with the average number of alarms within the several prediction cycles and the packet loss rate of the characteristic data.

[0006] Further, the control module is used to calculate the variance of the fluctuation amplitude of the surface temperature according to the surface temperature of the industrial Internet device in several prediction periods obtained, and to determine that the stability of the predictive maintenance does not meet the requirements when the variance of the fluctuation amplitude of the surface temperature satisfies the first variance condition or the second variance condition;

[0007] The first variance condition is that the variance of the fluctuation amplitude of the surface temperature is greater than the preset first variance and less than or equal to the preset second variance; the second variance condition is that the variance of the fluctuation amplitude of the surface temperature is greater than the preset second variance.

[0008] Furthermore, the control module preliminarily determines that the accuracy of the prediction result does not meet the requirements when the variance of the fluctuation amplitude of the surface temperature only satisfies the first variance condition, and makes a secondary determination on whether the accuracy of the prediction result meets the requirements based on the average number of alarms within several prediction cycles.

[0009] Further, the control module increases the frequency of collecting the operating data when the variance of the fluctuation amplitude of the surface temperature only satisfies the second variance condition.

[0010] Furthermore, the increase range of the collection frequency of the operating data is determined by the difference between the variance of the fluctuation amplitude of the surface temperature and a preset second variance.

[0011] Further, when the average alarm times in the plurality of prediction cycles meet the first alarm times condition or the second alarm times condition, the control module determines for the second time that the accuracy of the prediction result does not meet the requirement;

[0012] Among them, the first alarm number condition is that the average alarm number within a number of prediction cycles is greater than the preset first alarm number and less than or equal to the preset second alarm number; the second alarm number condition is that the average alarm number within a number of prediction cycles is greater than the preset second alarm number.

[0013] Furthermore, the control module preliminarily determines that the generation stability of the maintenance plan does not meet the requirements when the average number of alarms within the several prediction periods only meets the second alarm number condition, and makes a secondary determination on whether the generation stability of the maintenance plan meets the requirements based on the packet loss rate of the characteristic data.

[0014] Furthermore, the control module increases the data volume of the single training data when the average number of alarms in the plurality of prediction periods only meets the first number of alarms condition.

[0015] Furthermore, the increase in the data volume of the single training data is determined by the difference between the average number of alarms in a number of prediction cycles and a preset first number of alarms.

[0016] Furthermore, the control module secondarily determines that the stability of the generation of the maintenance plan does not meet the requirements when the packet loss rate of the characteristic data is greater than the preset packet loss rate, and increases the number of retransmissions of the lost characteristic data;

[0017] The increase in the number of retransmissions of the lost characteristic data is determined by the difference between the packet loss rate of the characteristic data and a preset packet loss rate.

[0018] Compared with the prior art, the beneficial effect of the present invention lies in that the system of the present invention adjusts the collection frequency of the operating data according to the variance of the fluctuation amplitude of the surface temperature by setting a data collection module, a data processing module, a prediction module, a storage module, an alarm module and a control module. The measurement error of the sensor causes the collected data to deviate, which causes the distribution of the data in the conversion process to deviate, and then causes problems in the data conversion process, resulting in a large fluctuation amplitude of the operating data. By increasing the collection frequency of the operating data, more details and changes can be captured, thereby reducing data deviations and fluctuation amplitudes. By adjusting the data volume of a single training data according to the average number of alarms in several prediction cycles, since some fault types are more common in the process of training the model, other types of faults are more common. The type of fault occurs relatively rarely, resulting in an unbalanced number of samples of different fault types in the data set, insufficient training data, problems in fault identification, and further false alarms when encountering normal conditions, which in turn leads to an increase in the number of alarms. By increasing the amount of single training data, the training process is accelerated and the number of false alarms is reduced. By adjusting the number of retransmissions of lost feature data according to the packet loss rate of feature data, some data packets are lost due to the influence of line instability during transmission of operating data, resulting in a decrease in data quality, leading to inaccurate judgment of equipment status, and thus inaccurate generation of maintenance plans for faults. By increasing the number of retransmissions of lost data, the probability of data loss is reduced, thereby improving the accuracy of equipment status judgment and achieving improved stability for predictive maintenance.

[0019] Furthermore, the system of the present invention adjusts the collection frequency of operating data by setting a preset first variance and a preset second variance. Since the measurement error of the sensor causes deviation in the collected data, the distribution of the data in the conversion process is biased, which in turn causes problems in the data conversion process, resulting in a large fluctuation in the operating data. By increasing the collection frequency of the operating data, more details and changes can be captured, thereby reducing data deviation and fluctuation, and further achieving improved stability for predictive maintenance.

[0020] Furthermore, the system of the present invention adjusts the data volume of a single training data by setting a preset first alarm number and a preset second alarm number. During the process of training the model, some fault types are more common and other types of fault types occur relatively less frequently, resulting in an imbalance in the number of samples of different fault types in the data set, insufficient training data, problems in fault identification, and further false alarms when encountering normal conditions, thereby increasing the number of alarms. By increasing the data volume of a single training data, the training process is accelerated, the number of false alarms is reduced, and the stability of predictive maintenance is further improved.

[0021] Furthermore, the system of the present invention adjusts the number of retransmissions of lost characteristic data by setting a preset packet loss rate. Since some data packets are lost due to the influence of line instability during the transmission of operating data, the quality of data is reduced, resulting in inaccurate judgment of equipment status, and thus inaccurate generation of maintenance plans for faults. By increasing the number of retransmissions of lost data, the probability of data loss is reduced, thereby improving the accuracy of equipment status judgment and further achieving improved stability of predictive maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is an overall structural block diagram of a predictive maintenance system for industrial Internet devices according to an embodiment of the present invention;

[0023] Figure 2 This is a logic flow chart of a predictive maintenance system for industrial Internet devices according to an embodiment of the present invention;

[0024] Figure 3 This is a specific structural block diagram of a data processing module of a predictive maintenance system for industrial Internet devices according to an embodiment of the present invention;

[0025] Figure 4 This is a specific structural block diagram of the prediction module of the predictive maintenance system of the industrial Internet equipment according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0027] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0028] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0029] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0030] See also Figure 1 , Figure 2 , Figure 3 as well as Figure 4 As shown, they are respectively an overall structural block diagram, a logic flow chart, a specific structural block diagram of a data processing module, and a specific structural block diagram of a prediction module of a predictive maintenance system for industrial Internet devices according to an embodiment of the present invention. A predictive maintenance system for industrial Internet devices according to the present invention comprises:

[0031] The data acquisition module is used to collect the operating data of industrial Internet equipment through a number of sensors;

[0032] A data processing module connected to the data acquisition module, comprising a data conversion unit for converting the pre-processed operation data to output characteristic data and a data retransmission unit connected to the data conversion unit for retransmitting the characteristic data when it is lost;

[0033] A prediction module, which is connected to the data processing unit, and includes a model training unit connected to the data conversion unit for training a machine learning model according to the feature data to output a prediction result, and a plan generation unit for generating a maintenance plan according to the identified fault of the industrial Internet device;

[0034] A storage module, which is connected to the data acquisition module, the data processing module and the prediction module respectively, and is used to store the operation data, the characteristic data, the prediction result and the maintenance plan respectively;

[0035] An alarm module, which is connected to the prediction module and is used to issue an alarm when a failure occurs in the industrial Internet device;

[0036] A control module, which is respectively connected to the data acquisition module, the data processing module, the prediction module, the storage module and the alarm module, and is used to adjust the collection frequency of operating data or the data volume of a single training data according to the average number of alarms in several prediction cycles when it is determined that the stability of the predicted maintenance does not meet the requirements based on the variance of the surface temperature fluctuation amplitude, and to adjust the number of retransmissions of lost feature data in combination with the average number of alarms in the several prediction cycles and the packet loss rate of the feature data.

[0037] Specifically, the data processing module also includes a preprocessing unit for preprocessing the operating data.

[0038] Specifically, the preprocessing operations include cleaning, denoising, and feature extraction.

[0039] Specifically, the data conversion unit is connected to the preprocessing unit.

[0040] Specifically, the prediction module also includes a fault identification unit connected to the model training unit for identifying the fault of the industrial Internet device according to the prediction result.

[0041] Specifically, the plan generation unit is connected to the fault identification unit.

[0042] Specifically, the sensors include temperature sensors, pressure sensors, and vibration sensors.

[0043] Specifically, industrial Internet devices include smart instruments, CNC machine tools, and smart surveillance cameras.

[0044] Specifically, the surface temperature of the Industrial Internet equipment is the surface temperature of the CNC machine tool.

[0045] Specifically, the operating data includes the temperature of the CNC machine tool, the operating voltage of the intelligent instrument, and the vibration intensity of the CNC machine tool.

[0046] Specifically, the characteristic data include the average temperature of CNC machine tools, the fluctuation range of the working voltage of smart instruments, and the average working current of smart surveillance cameras.

[0047] Specifically, the prediction results include the probability of failure, the time when failure occurs, and the remaining life of the equipment.

[0048] Specifically, the maintenance plan includes the equipment name, maintenance priority order, and maintenance plan execution time.

[0049] Specifically, the machine learning model can be a decision tree, a random forest, or a neural network.

[0050] Specifically, the amount of single training data refers to the amount of feature data used by the machine learning model in a training process.

[0051] Specifically, when it is detected that the characteristic data is discontinuous data, it is determined that the characteristic data is lost.

[0052] In implementation, the system of the present invention is provided with a data acquisition module, a data processing module, a prediction module, a storage module, an alarm module and a control module, and the collection frequency of the operating data is adjusted according to the variance of the fluctuation amplitude of the surface temperature. The measurement error of the sensor causes the deviation of the collected data, which causes the deviation of the distribution of the data in the conversion process, and then causes problems in the data conversion process, resulting in a large fluctuation amplitude of the operating data. By increasing the collection frequency of the operating data, more details and changes can be captured, thereby reducing data deviation and fluctuation amplitude. By adjusting the data volume of a single training data according to the average number of alarms in several prediction cycles, since in the process of training the model, some fault types are more common, and other types of fault types are relatively less common. , resulting in an imbalance in the number of samples of different fault types in the data set, insufficient training data, problems in fault identification, and further false alarms when encountering normal conditions, which in turn leads to an increase in the number of alarms. By increasing the amount of single training data, the training process is accelerated and the number of false alarms is reduced. By adjusting the number of retransmissions of lost feature data according to the packet loss rate of the feature data, some data packets are lost due to the influence of line instability during the transmission of the operating data, resulting in a decrease in data quality, leading to inaccurate judgment of the equipment status, and thus inaccurate generation of maintenance plans for faults. By increasing the number of retransmissions of lost data, the probability of data loss is reduced, thereby improving the accuracy of equipment status judgment and achieving improved stability for predictive maintenance.

[0053] Specifically, the control module is used to calculate the variance of the fluctuation amplitude of the surface temperature according to the surface temperature of the industrial Internet device in several prediction periods obtained, and to determine that the stability of the predictive maintenance does not meet the requirements when the variance of the fluctuation amplitude of the surface temperature satisfies the first variance condition or the second variance condition;

[0054] The first variance condition is that the variance of the fluctuation amplitude of the surface temperature is greater than the preset first variance and less than or equal to the preset second variance; the second variance condition is that the variance of the fluctuation amplitude of the surface temperature is greater than the preset second variance.

[0055] It can be understood that the three intervals corresponding to the preset first variance Q1 and the preset second variance Q2 correspond to three situations respectively. The first interval is Q≤Q1, corresponding to the situation where the stability of predictive maintenance meets the requirements; the second interval is Q1<Q≤Q2, corresponding to the situation where some fault types are more common and other types of fault types are relatively rare in the process of training the model, resulting in an imbalance in the number of samples of different fault types in the data set, resulting in insufficient training data; the third interval is Q>Q2, corresponding to the situation where the sensor has a measurement error, resulting in a deviation in the collected data, resulting in a deviation in the distribution of the data during the conversion process, and then causing problems in the data conversion process. In practice, Q1 is generally selected in the range of [4℃ 2 , 6℃ 2 ], Q1 is generally selected in the range of [7℃ 2 , 9℃ 2 ].

[0056] Preferably, the first variance Q1 is preset to be 5°C. 2 The preferred embodiment of the preset second variance Q2 is 8°C 2 .

[0057] Specifically, the variance of the fluctuation amplitude of the surface temperature is denoted as Q.

[0058] Specifically, the variance of the surface temperature fluctuation amplitude is the variance of the surface temperature fluctuation amplitude of the industrial Internet equipment within several prediction periods. The calculation method of the variance of the surface temperature fluctuation amplitude is a conventional technical means well known to technical personnel in this field. Therefore, the calculation process of the variance of the surface temperature fluctuation amplitude will not be repeated here.

[0059] Specifically, the calculation formula for the fluctuation amplitude of the surface temperature is:

[0060] G=T1-T2

[0061] Among them, G is the fluctuation amplitude of the surface temperature, T1 is the maximum surface temperature of the industrial Internet equipment in a single operating cycle, and T2 is the minimum surface temperature of the industrial Internet equipment in a single operating cycle.

[0062] Specifically, when a sensor has a measurement error, the measurement error can be discovered by comparing the measurement data at different time points.

[0063] In implementation, the system of the present invention determines the stability of predictive maintenance by setting a preset first variance and a preset second variance, thereby reducing the impact of decreased accuracy of predictive maintenance due to inaccurate judgment of the stability of predictive maintenance, and further improving the stability of predictive maintenance.

[0064] Specifically, the control module preliminarily determines that the accuracy of the prediction result does not meet the requirements when the variance of the fluctuation amplitude of the surface temperature only satisfies the first variance condition, and makes a secondary determination on whether the accuracy of the prediction result meets the requirements based on the average number of alarms within several prediction cycles.

[0065] Specifically, the control module increases the frequency of collecting the operating data when the variance of the fluctuation amplitude of the surface temperature only satisfies the second variance condition.

[0066] Specifically, the increase range of the collection frequency of the operating data is determined by the difference between the variance of the fluctuation amplitude of the surface temperature and a preset second variance.

[0067] Specifically, the calculation formula for the increased collection frequency of operating data is:

[0068]

[0069] Wherein, V' is the increased collection frequency of operating data, V is the current collection frequency of operating data, A is the minimum unit variance for adjusting the collection frequency of operating data, and v is the collection frequency of operating data when the variance of the fluctuation amplitude of the surface temperature is the minimum unit variance.

[0070] In practice, the value of A is generally taken as 9°C 2 ~11℃ 2 , preferably, the preferred embodiment of A is 10°C 2 .

[0071] In implementation, the optional range of v is [3 times / minute, 5 times / minute]. Preferably, when the collection frequency of operating data is less than 4 times / minute, problems may occur in the data conversion process. Therefore, the preferred embodiment of v is 4 times / minute.

[0072] Specifically, when the value of V' is a decimal, the value of V' is automatically rounded up to an integer.

[0073] For example, when using the predictive maintenance system of industrial Internet equipment to perform predictive maintenance on industrial equipment, V = 3 times / minute, Q = 12°C 2 , calculated

[0074] In implementation, the system of the present invention adjusts the collection frequency of operating data by setting a preset first variance and a preset second variance. Measurement errors in the sensor lead to deviations in the collected data, which in turn lead to problems in the data conversion process, resulting in large fluctuations in the operating data. By increasing the collection frequency of the operating data, more details and changes can be captured, thereby reducing data deviations and fluctuations, and further improving the stability of predictive maintenance.

[0075] Specifically, when the average number of alarms in the plurality of prediction cycles meets the first alarm number condition or the second alarm number condition, the control module determines that the accuracy of the prediction result does not meet the requirement;

[0076] Among them, the first alarm number condition is that the average alarm number within a number of prediction cycles is greater than the preset first alarm number and less than or equal to the preset second alarm number; the second alarm number condition is that the average alarm number within a number of prediction cycles is greater than the preset second alarm number.

[0077] It can be understood that the three intervals corresponding to the preset first alarm number P1 and the preset second alarm number P2 correspond to three situations respectively. The first interval is P≤P1, corresponding to the situation where the accuracy of the prediction result meets the requirements; the second interval is P1<P≤P2, corresponding to the situation that in the process of training the model, some fault types are more common, and other types of fault types occur relatively less, resulting in an unbalanced number of samples of different fault types in the data set, resulting in insufficient training data; the third interval is P>P2, corresponding to the situation that the operating data is affected by the unstable line during the transmission process, resulting in the loss of some data packets, resulting in a decrease in data quality, and resulting in inaccurate judgment of the equipment status. In practice, the range of P1 is generally selected from [5 times, 7 times], and the range of P2 is generally selected from [8 times, 10 times].

[0078] Preferably, the preferred embodiment of the preset first alarm number P1 is 6 times, and the preferred embodiment of the preset second alarm number P2 is 9 times.

[0079] Specifically, the average number of alarms within a number of prediction cycles is denoted as P.

[0080] Specifically, the calculation formula for the average number of alarms within several prediction periods is:

[0081]

[0082] Among them, P is the average number of alarms in several prediction cycles, X a is the number of alarms in the ath prediction period, n is the number of prediction periods, and n is a natural number greater than or equal to 1.

[0083] In implementation, the system of the present invention performs a secondary judgment on the accuracy of the prediction results by setting a preset first alarm number and a preset second alarm number, thereby reducing the impact of the decreased stability of predictive maintenance due to inaccurate secondary judgment on the accuracy of the prediction results, and further improving the stability of predictive maintenance.

[0084] Specifically, the control module preliminarily determines that the generation stability of the maintenance plan does not meet the requirements when the average number of alarms within the several prediction cycles only meets the second alarm number condition, and makes a secondary determination on whether the generation stability of the maintenance plan meets the requirements based on the packet loss rate of the characteristic data.

[0085] Specifically, the control module increases the data volume of the single training data when the average number of alarms in the prediction cycles only meets the first number of alarms condition.

[0086] Specifically, the increase in the data volume of the single training data is determined by the difference between the average number of alarms in a number of prediction cycles and the preset first number of alarms.

[0087] Specifically, the calculation formula for the amount of data for a single training data after the increase is:

[0088]

[0089] Among them, H' is the data volume of the single training data after the increase, H is the data volume of the current single training data, B is the minimum unit alarm number for adjusting the data volume of the single training data, and h is the data volume of the single training data when the average alarm number within several prediction cycles is the minimum unit alarm number.

[0090] In practice, the value of B is generally 11 to 13 times, and preferably, the preferred embodiment of B is 12 times.

[0091] In implementation, the optional value range of h is [450Byte, 550Byte]. Preferably, when the amount of single training data is less than 500Byte, it may cause insufficient training data. Therefore, the preferred embodiment of h is 500Byte.

[0092] For example, when using the predictive maintenance system of industrial Internet equipment to perform predictive maintenance on industrial equipment, where H = 400 Byte, P = 15 times, the calculation is

[0093] Specifically, by analyzing the operating data, operating parameters and historical records, it is determined whether there is a fault. When an alarm is issued but the data analysis results show that the equipment is operating normally, it is a false alarm.

[0094] In implementation, the system of the present invention adjusts the data volume of a single training data by setting a preset first alarm number and a preset second alarm number. During the process of training the model, some fault types are more common and other types of fault types occur relatively less frequently, resulting in an imbalance in the number of samples of different fault types in the data set, insufficient training data, problems in fault identification, and further false alarms when encountering normal conditions, thereby increasing the number of alarms. By increasing the data volume of a single training data, the training process is accelerated, the number of false alarms is reduced, and the stability of predictive maintenance is further improved.

[0095] Specifically, when the packet loss rate of the characteristic data is greater than a preset packet loss rate, the control module secondarily determines that the generation stability of the maintenance plan does not meet the requirements, and increases the number of retransmissions of the lost characteristic data.

[0096] It can be understood that the two intervals corresponding to the preset packet loss rate Y0 correspond to two situations. The first interval is Y≤Y0, corresponding to the situation where the stability of the maintenance plan generation meets the requirements; the second interval is Y>Y0, corresponding to the situation where the operation data is affected by the unstable line during the transmission process, resulting in the loss of some data packets, resulting in the degradation of data quality and inaccurate judgment of the equipment status. In practice, Y0 is generally selected in the range of [6%, 8%].

[0097] Preferably, the preferred embodiment of the preset packet loss rate Y0 is 7%.

[0098] Specifically, the packet loss rate of the feature data is recorded as Y.

[0099] In implementation, the system of the present invention performs a secondary judgment on the generation stability of the maintenance plan by setting a preset packet loss rate, thereby reducing the impact of the decreased stability of the predictive maintenance due to inaccurate secondary judgment on the generation stability of the maintenance plan, and further improving the stability of the predictive maintenance.

[0100] Specifically, the increase range of the number of retransmissions of the lost characteristic data is determined by the difference between the packet loss rate of the characteristic data and the preset packet loss rate.

[0101] Specifically, the calculation formula for the increased number of retransmissions of lost characteristic data is:

[0102]

[0103] Among them, L' is the increased number of retransmissions of the lost characteristic data, L is the current number of retransmissions of the lost characteristic data, C is the minimum unit packet loss rate for adjusting the number of retransmissions of the lost characteristic data, and l is the number of retransmissions of the lost characteristic data when the packet loss rate of the characteristic data is the minimum unit packet loss rate.

[0104] In practice, the value of C is generally 9% to 11%. Preferably, the preferred embodiment of C is 10%.

[0105] In implementation, the optional range of the value of l is [4 times, 6 times]. Preferably, when the number of retransmissions of the lost characteristic data is less than 5 times, the quality of the data may be reduced. Therefore, the preferred embodiment of l is 5 times.

[0106] Specifically, when the value of L' is a decimal, the value of L' is automatically rounded up to an integer.

[0107] For example, when using the predictive maintenance system of industrial Internet equipment to perform predictive maintenance on industrial equipment, where L = 3 times and Y = 12%, the calculation is

[0108] In implementation, the system of the present invention adjusts the number of retransmissions of lost characteristic data by setting a preset packet loss rate. Since some data packets are lost due to line instability during transmission of operating data, the quality of data is reduced, resulting in inaccurate judgment of equipment status, and thus inaccurate generation of maintenance plans for faults. By increasing the number of retransmissions of lost data, the probability of data loss is reduced, thereby improving the accuracy of equipment status judgment and further achieving improved stability for predictive maintenance.

[0109] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A predictive maintenance system for industrial Internet equipment, characterized in that: include: The data acquisition module is used to collect the operating data of industrial Internet equipment through a number of sensors; A data processing module, which is connected to the data acquisition module, and includes a data conversion unit for converting the pre-processed operation data to output characteristic data and a data retransmission unit connected to the data conversion unit for retransmitting the characteristic data when it is lost; A prediction module connected to the data processing unit, comprising a model training unit connected to the data conversion unit for training a machine learning model according to the feature data to output a prediction result and a plan generation unit for generating a maintenance plan according to the identified fault of the industrial Internet device; A storage module, which is connected to the data acquisition module, the data processing module and the prediction module respectively, and is used to store the operation data, the characteristic data, the prediction result and the maintenance plan respectively; An alarm module, which is connected to the prediction module and is used to issue an alarm when a failure occurs in the industrial Internet device; A control module, which is respectively connected to the data acquisition module, the data processing module, the prediction module, the storage module and the alarm module, and is used to adjust the collection frequency of operating data or the data volume of a single training data according to the average number of alarms in several prediction cycles when it is determined that the stability of the predicted maintenance does not meet the requirements based on the variance of the surface temperature fluctuation amplitude, and to adjust the number of retransmissions of lost feature data in combination with the average number of alarms in the several prediction cycles and the packet loss rate of the feature data.

2. The predictive maintenance system for industrial Internet equipment according to claim 1, characterized in that: The control module is used to calculate the variance of the fluctuation amplitude of the surface temperature according to the surface temperature of the industrial Internet device in several prediction periods obtained, and to determine that the stability of the predictive maintenance does not meet the requirements when the variance of the fluctuation amplitude of the surface temperature satisfies the first variance condition or the second variance condition; The first variance condition is that the variance of the fluctuation amplitude of the surface temperature is greater than the preset first variance and less than or equal to the preset second variance; the second variance condition is that the variance of the fluctuation amplitude of the surface temperature is greater than the preset second variance.

3. The predictive maintenance system for industrial Internet equipment according to claim 1, characterized in that: The control module preliminarily determines that the accuracy of the prediction result does not meet the requirements when the variance of the fluctuation amplitude of the surface temperature only meets the first variance condition, and makes a secondary determination on whether the accuracy of the prediction result meets the requirements based on the average number of alarms within several prediction cycles.

4. The predictive maintenance system for industrial Internet equipment according to claim 1, characterized in that: The control module increases the frequency of collecting the operating data when the variance of the fluctuation amplitude of the surface temperature only satisfies the second variance condition.

5. The predictive maintenance system for industrial Internet equipment according to claim 4, characterized in that: The increase range of the collection frequency of the operating data is determined by the difference between the variance of the fluctuation amplitude of the surface temperature and a preset second variance.

6. The predictive maintenance system for industrial Internet equipment according to claim 5, characterized in that: When the average number of alarms in the plurality of prediction cycles meets the first alarm number condition or the second alarm number condition, the control module determines that the accuracy of the prediction result does not meet the requirement; Among them, the first alarm number condition is that the average alarm number within a number of prediction cycles is greater than the preset first alarm number and less than or equal to the preset second alarm number; the second alarm number condition is that the average alarm number within a number of prediction cycles is greater than the preset second alarm number.

7. The predictive maintenance system for industrial Internet equipment according to claim 6, characterized in that: The control module preliminarily determines that the generation stability of the maintenance plan does not meet the requirements when the average number of alarms within the several prediction cycles only meets the second alarm number condition, and makes a secondary determination on whether the generation stability of the maintenance plan meets the requirements based on the packet loss rate of the characteristic data.

8. The predictive maintenance system for industrial Internet equipment according to claim 7, characterized in that: The control module increases the data volume of the single training data when the average number of alarms in the prediction cycles only meets the first alarm number condition.

9. The predictive maintenance system for industrial Internet equipment according to claim 8, characterized in that: The increase in the data volume of the single training data is determined by the difference between the average number of alarms in a number of prediction cycles and the preset first number of alarms.

10. The predictive maintenance system for industrial Internet equipment according to claim 9, characterized in that: The control module secondarily determines that the stability of the generation of the maintenance plan does not meet the requirements when the packet loss rate of the characteristic data is greater than the preset packet loss rate, and increases the number of retransmissions of the lost characteristic data; The increase in the number of retransmissions of the lost characteristic data is determined by the difference between the packet loss rate of the characteristic data and a preset packet loss rate.

Images