Sensor damage detection method and system
By collecting and analyzing the vibration acceleration of mechanical equipment in real time, establishing a relationship network to screen suspicious sensor modules, and calculating abnormality index, the hysteresis and false alarm problems in sensor damage detection are solved, and efficient and intelligent damage detection is achieved.
Patent Information
- Application Number
- CN202411735490.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-29
- Publication Date
- 2025-03-04
AI Technical Summary
The prior art is difficult to effectively detect sensor damage, especially in the case of violent dismantling, and traditional methods have problems of lag and false alarms.
By collecting the operation log of the mechanical equipment, the vibration acceleration in the three-axis direction is collected in real time, the influence coefficients between sensor modules are analyzed, the relationship network is established, the suspicious sensor module is screened out, and whether it enters the alarm state is determined by the abnormal index.
It realizes rapid detection of sensor damage, reduces false alarm rate, and improves system reliability and user trust.
Smart Images
Figure CN119555134A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of damage detection, and in particular to a sensor damage detection method and system. Background Art
[0002] In the modern engineering machinery construction industry, sensor technology has been widely used in equipment status monitoring, work efficiency improvement, safety assurance and other aspects. However, in actual applications, sensor equipment may be subjected to violent dismantling by lawbreakers, such as knocking, prying, hammering, etc. to dismantle or directly damage it. These violent behaviors may not only cause damage to sensor equipment, but also affect the normal operation of mechanical equipment and even cause serious economic losses.
[0003] At present, in terms of sensor damage detection, the occurrence of abnormal events is usually detected by judging whether the business data is abnormal or whether the vibration change breaks through the range. This method has certain problems. On the one hand, when the sensor is violently dismantled, the business data collected by it may not have shown abnormal changes, so it is impossible to rely on these business data to infer the violent dismantling event. This lag makes the traditional detection method based on business data anomalies invalid. On the other hand, mechanical equipment will generate a large amount of large-amplitude vibrations during normal operation, and these vibrations may have certain similarities with the vibration characteristics during violent dismantling. If the normal working vibration and the violent dismantling vibration cannot be effectively distinguished, a large number of false alarms will result, reducing the reliability of the system and the trust of users. Therefore, at this stage, an efficient and intelligent sensor damage detection technology solution is needed to solve the above problems. Summary of the invention
[0004] The object of the present invention is to provide a sensor damage detection method and system to solve the problems raised in the above background technology.
[0005] In order to solve the above technical problems, the present invention provides a sensor damage detection method, comprising the following steps:
[0006] S100, collecting the operation log of the mechanical equipment, and collecting the vibration acceleration in three-axis directions in real time through the sensor module installed on the mechanical equipment.
[0007] S200, analyzing the parameter records of each sensor module in the operation log, calculating the influence coefficients between the sensor modules and analyzing whether there is an influence relationship, establishing a relationship network based on the influence relationship and screening out suspicious sensor modules.
[0008] S300, calculating an abnormality index through the vibration acceleration of the suspicious sensor module, judging whether to enter an alarm state according to changes in the abnormality index, and quickly guiding and giving abnormality alarms for the data in the sensor module in the alarm state.
[0009] S400, the cloud platform displays the data information uploaded by the alarm device in real time through a large visual screen, and notifies the management personnel of the occurrence of abnormal events.
[0010] In S100, mechanical equipment refers to construction machinery used for building construction. The operation log refers to the parameter record of each sensor module during the operation of the mechanical equipment. Each sensor module corresponds to a parameter record, including a code and a vibration acceleration relationship diagram. The code is used to distinguish different sensor modules, and the vibration acceleration relationship diagram represents the vibration acceleration in the three-axis direction collected by the sensor module at different times. The three-axis direction refers to the direction defined along three mutually perpendicular coordinate axes in three-dimensional space; the sensor module refers to a functional module composed of different types of sensors.
[0011] Sensor modules are installed at different locations of mechanical equipment, and various types of sensors are used to collect data at pre-set frequencies. Each sensor module includes at least one triaxial acceleration sensor. These sensor modules are connected to a smart terminal, and all sensor modules transmit the collected data to the smart terminal, which then uploads it to the cloud platform. Each mechanical equipment has at least one smart terminal.
[0012] In S200, the specific steps are as follows:
[0013] S201, obtain all parameter records in the operation log and set u time nodes, mark the u time nodes at the same time in the vibration acceleration relationship diagram in each parameter record, and analyze the vibration acceleration in the three-axis direction at each time node. Count the number of all parameter records n, and associate these parameter records two by two to obtain (n 2 -n)÷2 pairs of parameter records. Analyze the vibration acceleration of each pair of parameter records in the three axes at each time node, and associate the vibration acceleration of different parameter records at the same time node and in the same axis direction.
[0014] The vibration acceleration relationship diagram is a line chart of the relationship between vibration acceleration and time. Each vibration acceleration relationship diagram can simultaneously display the vibration acceleration of the sensor module in the three-axis direction over time. By analyzing the vibration acceleration relationship diagram, the vibration acceleration in the three-axis direction at different times can be obtained. The vibration acceleration relationship diagram is drawn in real time based on the data collected by the three-axis acceleration sensor.
[0015] S202, group the associated vibration accelerations according to the axial direction and establish a training set for each group. The associated vibration accelerations in the same group are respectively used as independent variables and dependent variables and packaged as samples, and these u samples are placed in the training set corresponding to the group to which they belong. Establish a linear regression model, substitute the independent variables of the samples in each training set as input values into the linear regression model expression, calculate the difference between the output result of each sample and the dependent variable, sum the differences of all samples in the training set as the gap coefficient, and obtain the minimum gap coefficient PL by adjusting the intercept F and the regression coefficient I in the expression. min , thus obtaining the final expression:
[0016] E=F+IL
[0017] In the formula, E is the dependent variable and L is the independent variable.
[0018] S203. Each pair of parameter records corresponds to three expressions. Calculate the standard deviation F of the intercepts in these three expressions. sd , standard deviation of regression coefficient I sd and the standard deviation of the minimum gap coefficient PL sd , substitute into the formula to calculate the influence coefficient PX; set the coefficient threshold PX th , the influence coefficient is greater than PX th The encoding of each pair of parameter records is associated with the corresponding sensor module, and the associated sensor modules are marked with the influence relationship. The influence coefficient calculation formula is as follows:
[0019]
[0020] The influence coefficient indicates the similarity of the three expressions corresponding to each pair of parameter records. When the standard deviation of the intercept, regression coefficient and minimum gap coefficient in these expressions is smaller, the similarity of the corresponding expressions is higher. The two sensor modules corresponding to this pair of parameter records have a mutual influence relationship in terms of vibration acceleration.
[0021] S204, establish a relationship network, connect each pair of associated sensor modules with an impact relationship. Count the number of connections between each sensor module and other sensor modules in the relationship network, sort the sensor modules in descending order according to the number of connections; obtain the sensor module MZ in order according to the sorting order one The vibration acceleration in the three-axis direction collected in real time is substituted into the MZ one Associated sensor module MZ two In the final expression between, the expected vibration acceleration in the three-axis direction is calculated, and the sensor module MZ is calculated. two The difference between the vibration acceleration in the three-axis direction collected in real time and the expected vibration acceleration is set to the deviation threshold g. When the difference in the three-axis direction is greater than g, the sensor module MZ istwo Marks an exception.
[0022] S205, when with MZ one After all associated sensor modules are calculated, MZ is obtained in order of arrangement. one Subsequent sensor modules calculate the difference in vibration acceleration and mark the anomaly until all sensor modules in the arrangement order have completed the calculation and marked the anomaly. Then, the number of times each sensor module in the relationship network is marked as anomaly is counted, and the sensor module with the most number of marked anomalies and a number greater than M is selected as the suspicious sensor module.
[0023] Each sensor module may have an influence relationship with multiple other sensor modules, and the established relationship network may also have multiple influence relationship chains. In the influence chain, when the sensor module with problems is used as calculation data, all other sensor modules that have influence relationships but have no problems will be marked as abnormal. When each sensor module without problems is used as calculation data, the sensor module with problems will be marked as abnormal multiple times. By counting the number of times each sensor module is marked as abnormal, the suspicious sensor module with real problems can be found.
[0024] In S300, the specific steps are as follows:
[0025] S301. Mark the sensor modules that have an influencing relationship with the suspicious sensor modules in the relationship network, count the number k of all marked sensor modules, obtain the vibration acceleration in the three-axis direction collected by each marked sensor module in real time, and bring them into the relationship formula between the suspicious sensor module and the suspicious sensor module to calculate the expected vibration acceleration. Collect the vibration acceleration in the three-axis direction of the suspicious sensor module and bring it into the formula to calculate the abnormal index of the suspicious sensor module:
[0026]
[0027] In the formula, RT is the abnormality index, va i is the vibration acceleration of the suspicious sensor module in the i-th axis direction, The estimated vibration acceleration of the suspicious sensor module in the i-th axis direction is calculated for the h-th marked sensor module.
[0028] S302, set the sampling frequency SF and the interval time, and calculate the abnormal index of the suspicious sensor module according to the sampling frequency SF. Set the index threshold RT th , when the abnormal index of the suspicious sensor module is greater than RT thWhen the duration is longer than time, the alarm state is entered. The frequency of data collection of the sensor module in the alarm state is increased and the data is transmitted to the smart terminal first. The smart terminal uploads the sensor module data in the alarm state to the cloud platform and issues an abnormal alarm.
[0029] The sampling frequency and interval duration are set in advance by the management personnel. The vibration acceleration of the suspicious sensor module is collected multiple times in succession and the abnormal index is calculated. The dual factors of the sampling interval duration and the size of the abnormal index are used to comprehensively determine whether to issue an abnormal alarm and prevent false alarms.
[0030] The cloud platform uses a large visual screen to display the mechanical equipment and specific sensor modules in alarm status in real time, collects and verifies the data uploaded by the alarm status sensor module, and pushes abnormal alarm information to management personnel for decision-making.
[0031] A sensor damage detection system comprises a data acquisition module, an abnormality analysis module, a damage detection module and an intelligent early warning module.
[0032] The data acquisition module collects the operation log of the mechanical equipment, and collects the vibration acceleration in the three-axis direction in real time through the sensor module installed on the mechanical equipment. The abnormal analysis module analyzes the parameter records of each sensor module in the operation log, calculates the influence coefficient between the sensor modules and analyzes whether there is an influence relationship, establishes a relationship network and screens out suspicious sensor modules. The damage detection module calculates the abnormal index through the vibration acceleration of the suspicious sensor module, determines whether to enter the alarm state according to the change of the abnormal index, and quickly guides and alarms the data in the sensor module in the alarm state. The intelligent early warning module displays the data information uploaded by the alarm device in real time through a large visual screen, and notifies the management personnel of the occurrence of abnormal events.
[0033] The data collection module includes a parameter information collection unit and an operation log collection unit.
[0034] The parameter information acquisition unit collects vibration acceleration in three-axis directions in real time through the sensor module installed on the mechanical equipment.
[0035] The operation log collection unit is used to collect parameter records of each sensor module. The parameter record includes a code and a vibration acceleration relationship diagram. The code is used to distinguish different sensor modules, and the vibration acceleration relationship diagram represents the vibration acceleration in the three-axis direction collected by the sensor module at different times. The three-axis direction refers to the direction defined along three mutually perpendicular coordinate axes in three-dimensional space.
[0036] The anomaly analysis module includes a relationship analysis unit and an anomaly screening unit.
[0037] The relationship analysis unit is used to analyze whether there is an influence relationship between sensor modules.
[0038] First, set u time nodes and mark these u time nodes at the same time in the vibration acceleration relationship diagram in each parameter record. Count the number of all parameter records n, and associate these parameter records two by two to obtain (n 2 -n)÷2 pairs of parameter records. Analyze the vibration acceleration of each pair of parameter records in the three axes at each time node, and associate the vibration acceleration of different parameter records at the same time node and in the same axis direction.
[0039] Secondly, group the associated vibration accelerations according to the axial direction and establish a training set for each group. The associated vibration accelerations in the same group are used as independent variables and dependent variables and packaged as samples. These u samples are placed in the training set corresponding to the group to which they belong. Establish a linear regression model, substitute the independent variables of the samples as input values into the linear regression model expression, calculate the difference between the output result of each sample and the dependent variable, and sum the differences of all samples in the training set as the gap coefficient. The minimum gap coefficient PL is obtained by adjusting the intercept F and the regression coefficient I in the expression. min , thus obtaining the final expression: E=F+IL, where E is the dependent variable and L is the independent variable.
[0040] Finally, calculate the standard deviation F of the intercepts in the three expressions corresponding to each pair of parameter records sd , standard deviation of regression coefficient I sd and the standard deviation of the minimum gap coefficient PL sd , according to the formula Calculate the influence coefficient; set the coefficient threshold PX th , the influence coefficient is greater than PX th The encoding of each pair of parameter records is associated with the corresponding sensor module, and the associated sensor modules are marked with the influence relationship.
[0041] The abnormal screening unit is used to screen suspicious sensor modules.
[0042] First, a relationship network is established, and each pair of associated sensor modules with an impact relationship is connected. The number of connections between each sensor module and other sensor modules in the relationship network is counted, and the sensor modules are sorted in descending order according to the number of connections.
[0043] Secondly, obtain the sensor modules MZ in order of arrangement one The vibration acceleration in the three-axis direction collected in real time is substituted into the MZ one Associated sensor module MZ two In the final expression between, the expected vibration acceleration in the three-axis direction is calculated, and the sensor module MZ is calculated. twoThe difference between the vibration acceleration in the three-axis direction collected in real time and the expected vibration acceleration is set to the deviation threshold g. When the difference in the three-axis direction is greater than g, the sensor module MZ is two Mark the exception.
[0044] Finally, when with MZ one After all associated sensor modules are calculated, MZ is obtained in order of arrangement. one Subsequent sensor modules calculate the difference in vibration acceleration and mark the anomaly until all sensor modules in the arrangement order have completed the calculation and marked the anomaly. Then, the number of times each sensor module in the relationship network is marked as anomaly is counted, and the sensor module with the most number of marked anomalies and a number greater than M is selected as the suspicious sensor module.
[0045] The damage detection module includes an index calculation unit and a data grooming unit.
[0046] The index calculation unit is used to calculate the abnormal index of the suspicious sensor module. First, mark the sensor modules that have an impact relationship with the suspicious sensor module in the relationship network, count the number k of all marked sensor modules, obtain the vibration acceleration in the three-axis direction collected by each marked sensor module in real time, and bring them into the relationship formula with the suspicious sensor module to calculate the expected vibration acceleration. Then collect the vibration acceleration in the three-axis direction of the suspicious sensor module, according to the formula: Calculate the abnormal index of the suspicious sensor module. i is the vibration acceleration of the suspicious sensor module in the i-th axis direction, The estimated vibration acceleration of the suspicious sensor module in the i-th axis direction is calculated for the h-th marked sensor module.
[0047] The data grooming unit conducts rapid grooming and abnormal alarms for the data in the sensor module. First, set the sampling frequency SF and the interval time, and calculate the abnormal index of the suspicious sensor module according to the sampling frequency SF. Second, set the index threshold RT th , when the abnormal index of the suspicious sensor module is greater than RT th And when the duration is greater than time, it enters the alarm state. Finally, the sensor module data collection frequency of the alarm state is increased and transmitted to the smart terminal first. The smart terminal uploads the sensor module data of the alarm state to the cloud platform and issues an abnormal alarm.
[0048] The intelligent early warning module displays the mechanical equipment and sensor modules in alarm status through a large visual screen, obtains the data uploaded by the alarm status sensor module, and pushes the abnormal alarm information to the management personnel for decision-making.
[0049] Compared with the prior art, the beneficial effects achieved by the present invention are:
[0050] This application first analyzes the parameter records of each sensor module in the operation log, associates the parameter records in pairs, and then identifies the sensor modules that affect each other and builds a relationship network by analyzing the relationship between the vibration accelerations at the same time node and in the coaxial direction under different parameter records. Secondly, the vibration accelerations collected by each sensor module in the relationship network are marked as abnormal, and suspicious sensor modules are screened out according to the number of marked abnormalities. Finally, the sensor module with real problems is found by calculating the abnormality index and analyzing the changes in the abnormality index, and the data collected is quickly guided and abnormal alarms are issued. Compared with the abnormality detection method in the prior art that detects whether the business data is abnormal or whether the vibration change breaks through the interval, it is more intelligent and efficient, and can quickly detect damage and issue timely warnings.
[0051] In summary, compared with traditional technologies, the present invention has the advantages of rapid anomaly screening and intelligent data management, and can improve the efficiency of damage detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:
[0053] Figure 1 It is a schematic diagram of a flow chart of a sensor damage detection method of the present invention;
[0054] Figure 2 It is a structural schematic diagram of a sensor damage detection system of the present invention. DETAILED DESCRIPTION
[0055] 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.
[0056] See also Figure 1 The present invention provides a sensor damage detection method, comprising the following steps:
[0057] S100, collecting the operation log of the mechanical equipment, and collecting the vibration acceleration in three-axis directions in real time through the sensor module installed on the mechanical equipment.
[0058] S200, analyzing the parameter records of each sensor module in the operation log, calculating the influence coefficients between the sensor modules and analyzing whether there is an influence relationship, establishing a relationship network based on the influence relationship and screening out suspicious sensor modules.
[0059] S300, calculating an abnormality index through the vibration acceleration of the suspicious sensor module, judging whether to enter an alarm state according to changes in the abnormality index, and quickly guiding and giving abnormality alarms for the data in the sensor module in the alarm state.
[0060] S400, the cloud platform displays the data information uploaded by the alarm device in real time through a large visual screen, and notifies the management personnel of the occurrence of abnormal events.
[0061] In S100, mechanical equipment refers to construction machinery used for building construction. The operation log refers to the parameter record of each sensor module during the operation of the mechanical equipment. Each sensor module corresponds to a parameter record, including a code and a vibration acceleration relationship diagram. The code is used to distinguish different sensor modules, and the vibration acceleration relationship diagram represents the vibration acceleration in the three-axis direction collected by the sensor module at different times. The three-axis direction refers to the direction defined along three mutually perpendicular coordinate axes in three-dimensional space; the sensor module refers to a functional module composed of different types of sensors.
[0062] Sensor modules are installed at different locations of mechanical equipment, and various types of sensors are used to collect data at pre-set frequencies. Each sensor module includes at least one triaxial acceleration sensor. These sensor modules are connected to a smart terminal, and all sensor modules transmit the collected data to the smart terminal, which then uploads it to the cloud platform. Each mechanical equipment has at least one smart terminal.
[0063] In S200, the specific steps are as follows:
[0064] S201, obtain all parameter records in the operation log and set u time nodes, mark the u time nodes at the same time in the vibration acceleration relationship diagram in each parameter record, and analyze the vibration acceleration in the three-axis direction at each time node. Count the number of all parameter records n, and associate these parameter records two by two to obtain (n 2 -n)÷2 pairs of parameter records. Analyze the vibration acceleration of each pair of parameter records in the three axes at each time node, and associate the vibration acceleration of different parameter records at the same time node and in the same axis direction.
[0065] The vibration acceleration relationship diagram is a line chart of the relationship between vibration acceleration and time. Each vibration acceleration relationship diagram can simultaneously display the vibration acceleration of the sensor module in the three-axis direction over time. By analyzing the vibration acceleration relationship diagram, the vibration acceleration in the three-axis direction at different times can be obtained. The vibration acceleration relationship diagram is drawn in real time based on the data collected by the three-axis acceleration sensor.
[0066] S202, group the associated vibration accelerations according to the axial direction and establish a training set for each group. The associated vibration accelerations in the same group are respectively used as independent variables and dependent variables and packaged as samples, and these u samples are placed in the training set corresponding to the group to which they belong. Establish a linear regression model, substitute the independent variables of the samples in each training set as input values into the linear regression model expression, calculate the difference between the output result of each sample and the dependent variable, sum the differences of all samples in the training set as the gap coefficient, and obtain the minimum gap coefficient PL by adjusting the intercept F and the regression coefficient I in the expression. min , thus obtaining the final expression:
[0067] E=F+IL
[0068] In the formula, E is the dependent variable and L is the independent variable.
[0069] S203. Each pair of parameter records corresponds to three expressions. Calculate the standard deviation F of the intercepts in these three expressions. sd , standard deviation of regression coefficient I sd and the standard deviation of the minimum gap coefficient PL sd , substitute into the formula to calculate the influence coefficient PX; set the coefficient threshold PX th , the influence coefficient is greater than PX th The encoding of each pair of parameter records is associated with the corresponding sensor module, and the associated sensor modules are marked with the influence relationship. The influence coefficient calculation formula is as follows:
[0070]
[0071] The influence coefficient indicates the similarity of the three expressions corresponding to each pair of parameter records. When the standard deviation of the intercept, regression coefficient and minimum gap coefficient in these expressions is smaller, the similarity of the corresponding expressions is higher. The two sensor modules corresponding to this pair of parameter records have a mutual influence relationship in terms of vibration acceleration.
[0072] S204, establish a relationship network, connect each pair of associated sensor modules with an impact relationship. Count the number of connections between each sensor module and other sensor modules in the relationship network, sort the sensor modules in descending order according to the number of connections; obtain the sensor module MZ in order according to the sorting order oneThe vibration acceleration in the three-axis direction collected in real time is substituted into the MZ one Associated sensor module MZ two In the final expression between, the expected vibration acceleration in the three-axis direction is calculated, and the sensor module MZ is calculated. two The difference between the vibration acceleration in the three-axis direction collected in real time and the expected vibration acceleration is set to the deviation threshold g. When the difference in the three-axis direction is greater than g, the sensor module MZ is two Marks an exception.
[0073] S205, when with MZ one After all associated sensor modules are calculated, MZ is obtained in order of arrangement. one Subsequent sensor modules calculate the difference in vibration acceleration and mark the anomaly until all sensor modules in the arrangement order have completed the calculation and marked the anomaly. Then, the number of times each sensor module in the relationship network is marked as anomaly is counted, and the sensor module with the most number of marked anomalies and a number greater than M is selected as the suspicious sensor module.
[0074] Each sensor module may have an influence relationship with multiple other sensor modules, and the established relationship network may also have multiple influence relationship chains. In the influence chain, when the sensor module with problems is used as calculation data, all other sensor modules that have influence relationships but have no problems will be marked as abnormal. When each sensor module without problems is used as calculation data, the sensor module with problems will be marked as abnormal multiple times. By counting the number of times each sensor module is marked as abnormal, the suspicious sensor module with real problems can be found.
[0075] In S300, the specific steps are as follows:
[0076] S301. Mark the sensor modules that have an influencing relationship with the suspicious sensor modules in the relationship network, count the number k of all marked sensor modules, obtain the vibration acceleration in the three-axis direction collected by each marked sensor module in real time, and bring them into the relationship formula between the suspicious sensor module and the suspicious sensor module to calculate the expected vibration acceleration. Collect the vibration acceleration in the three-axis direction of the suspicious sensor module and bring it into the formula to calculate the abnormal index of the suspicious sensor module:
[0077]
[0078] In the formula, RT is the abnormality index, va i is the vibration acceleration of the suspicious sensor module in the i-th axis direction, The estimated vibration acceleration of the suspicious sensor module in the i-th axis direction is calculated for the h-th marked sensor module.
[0079] S302, set the sampling frequency SF and the interval time, and calculate the abnormal index of the suspicious sensor module according to the sampling frequency SF. Set the index threshold RT th , when the abnormal index of the suspicious sensor module is greater than RT th When the duration is longer than time, the alarm state is entered. The frequency of data collection of the sensor module in the alarm state is increased and the data is transmitted to the smart terminal first. The smart terminal uploads the sensor module data in the alarm state to the cloud platform and issues an abnormal alarm.
[0080] The sampling frequency and interval duration are set in advance by the management personnel. The vibration acceleration of the suspicious sensor module is collected multiple times in succession and the abnormal index is calculated. The dual factors of the sampling interval duration and the size of the abnormal index are used to comprehensively determine whether to issue an abnormal alarm and prevent false alarms.
[0081] The cloud platform uses a large visual screen to display the mechanical equipment and specific sensor modules in alarm status in real time, collects and verifies the data uploaded by the alarm status sensor module, and pushes abnormal alarm information to management personnel for decision-making.
[0082] See also Figure 2 The present invention provides a sensor damage detection system, including a data acquisition module, an abnormality analysis module, a damage detection module and an intelligent early warning module.
[0083] The data acquisition module collects the operation log of the mechanical equipment, and collects the vibration acceleration in the three-axis direction in real time through the sensor module installed on the mechanical equipment. The abnormal analysis module analyzes the parameter records of each sensor module in the operation log, calculates the influence coefficient between the sensor modules and analyzes whether there is an influence relationship, establishes a relationship network and screens out suspicious sensor modules. The damage detection module calculates the abnormal index through the vibration acceleration of the suspicious sensor module, determines whether to enter the alarm state according to the change of the abnormal index, and quickly guides and alarms the data in the sensor module in the alarm state. The intelligent early warning module displays the data information uploaded by the alarm device in real time through a large visual screen, and notifies the management personnel of the occurrence of abnormal events.
[0084] The data collection module includes a parameter information collection unit and an operation log collection unit.
[0085] The parameter information acquisition unit collects vibration acceleration in three-axis directions in real time through the sensor module installed on the mechanical equipment.
[0086] The operation log collection unit is used to collect parameter records of each sensor module. The parameter record includes a code and a vibration acceleration relationship diagram. The code is used to distinguish different sensor modules, and the vibration acceleration relationship diagram represents the vibration acceleration in the three-axis direction collected by the sensor module at different times. The three-axis direction refers to the direction defined along three mutually perpendicular coordinate axes in three-dimensional space.
[0087] The anomaly analysis module includes a relationship analysis unit and an anomaly screening unit.
[0088] The relationship analysis unit is used to analyze whether there is an influence relationship between sensor modules.
[0089] First, set u time nodes and mark these u time nodes at the same time in the vibration acceleration relationship diagram in each parameter record. Count the number of all parameter records n, and associate these parameter records two by two to obtain (n 2 -n)÷2 pairs of parameter records. Analyze the vibration acceleration of each pair of parameter records in the three axes at each time node, and associate the vibration acceleration of different parameter records at the same time node and in the same axis direction.
[0090] Secondly, group the associated vibration accelerations according to the axial direction and establish a training set for each group. The associated vibration accelerations in the same group are used as independent variables and dependent variables and packaged as samples. These u samples are placed in the training set corresponding to the group to which they belong. Establish a linear regression model, substitute the independent variables of the samples as input values into the linear regression model expression, calculate the difference between the output result of each sample and the dependent variable, and sum the differences of all samples in the training set as the gap coefficient. The minimum gap coefficient PL is obtained by adjusting the intercept F and the regression coefficient I in the expression. min , thus obtaining the final expression: E=F+IL, where E is the dependent variable and L is the independent variable.
[0091] Finally, calculate the standard deviation F of the intercepts in the three expressions corresponding to each pair of parameter records sd , standard deviation of regression coefficient I sd and the standard deviation of the minimum gap coefficient PL sd , according to the formula Calculate the influence coefficient; set the coefficient threshold PX th , the influence coefficient is greater than PX th The encoding of each pair of parameter records is associated with the corresponding sensor module, and the associated sensor modules are marked with the influence relationship.
[0092] The abnormal screening unit is used to screen suspicious sensor modules.
[0093] First, a relationship network is established, and each pair of associated sensor modules with an impact relationship is connected. The number of connections between each sensor module and other sensor modules in the relationship network is counted, and the sensor modules are sorted in descending order according to the number of connections.
[0094] Secondly, obtain the sensor modules MZ in order of arrangement one The vibration acceleration in the three-axis direction collected in real time is substituted into the MZ one Associated sensor module MZ two In the final expression between, the expected vibration acceleration in the three-axis direction is calculated, and the sensor module MZ is calculated. two The difference between the vibration acceleration in the three-axis direction collected in real time and the expected vibration acceleration is set to the deviation threshold g. When the difference in the three-axis direction is greater than g, the sensor module MZ is two Mark the exception.
[0095] Finally, when with MZ one After all associated sensor modules are calculated, MZ is obtained in order of arrangement. one Subsequent sensor modules calculate the difference in vibration acceleration and mark the anomaly until all sensor modules in the arrangement order have completed the calculation and marked the anomaly. Then, the number of times each sensor module in the relationship network is marked as anomaly is counted, and the sensor module with the most number of marked anomalies and a number greater than M is selected as the suspicious sensor module.
[0096] The damage detection module includes an index calculation unit and a data grooming unit.
[0097] The index calculation unit is used to calculate the abnormal index of the suspicious sensor module. First, mark the sensor modules that have an impact relationship with the suspicious sensor module in the relationship network, count the number k of all marked sensor modules, obtain the vibration acceleration in the three-axis direction collected by each marked sensor module in real time, and bring them into the relationship formula with the suspicious sensor module to calculate the expected vibration acceleration. Then collect the vibration acceleration in the three-axis direction of the suspicious sensor module, according to the formula: Calculate the abnormal index of the suspicious sensor module. i is the vibration acceleration of the suspicious sensor module in the i-th axis direction, The estimated vibration acceleration of the suspicious sensor module in the i-th axis direction is calculated for the h-th marked sensor module.
[0098] The data grooming unit conducts rapid grooming and abnormal alarms for the data in the sensor module. First, set the sampling frequency SF and the interval time, and calculate the abnormal index of the suspicious sensor module according to the sampling frequency SF. Second, set the index threshold RTth , when the abnormal index of the suspicious sensor module is greater than RT th And when the duration is greater than time, it enters the alarm state. Finally, the sensor module data collection frequency of the alarm state is increased and transmitted to the smart terminal first. The smart terminal uploads the sensor module data of the alarm state to the cloud platform and issues an abnormal alarm.
[0099] The intelligent early warning module displays the mechanical equipment and sensor modules in alarm status through a large visual screen, obtains the data uploaded by the alarm status sensor module, and pushes the abnormal alarm information to the management personnel for decision-making.
[0100] Embodiment 1:
[0101] Assuming that the standard deviation of the intercepts in the three expressions corresponding to a pair of parameter records is 1.2, the standard deviation of the regression coefficient is 2.1, and the standard deviation of the minimum gap coefficient is 1.5, when the number of time nodes is 8, the influence coefficient of this pair of parameter records is calculated by the formula:
[0102] Influence coefficient:
[0103] When the coefficient threshold is 0.5, the encoding corresponding to the sensor module of the pair of parameter records is associated.
[0104] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0105] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A sensor damage detection method, characterized in that: The method comprises the following steps: S100, collecting the operation log of the mechanical equipment, and collecting the vibration acceleration in three-axis directions in real time through the sensor module installed on the mechanical equipment; S200, analyzing the parameter records of each sensor module in the operation log, calculating the influence coefficients between the sensor modules and analyzing whether there is an influence relationship, establishing a relationship network based on the influence relationship and screening out suspicious sensor modules; S300, calculating an abnormality index through the vibration acceleration of the suspicious sensor module, judging whether to enter an alarm state according to the change of the abnormality index, and quickly guiding and giving an abnormality alarm for the data in the sensor module in the alarm state; S400, the cloud platform displays the data information uploaded by the alarm device in real time through a large visual screen, and notifies the management personnel of the occurrence of abnormal events.
2. A sensor damage detection method according to claim 1, characterized in that: In S100, mechanical equipment refers to construction machinery used in construction; operation log refers to parameter records of each sensor module during the operation of mechanical equipment, and each sensor module corresponds to a parameter record, including a code and a vibration acceleration relationship diagram; the code is used to distinguish different sensor modules, and the vibration acceleration relationship diagram represents the vibration acceleration in the three-axis directions collected by the sensor module at different times; the three-axis directions refer to the directions defined along three mutually perpendicular coordinate axes in three-dimensional space; the sensor module refers to a functional module composed of different types of sensors.
3. A sensor damage detection method according to claim 2, characterized in that: In S200, the specific steps are as follows: S201, obtain all parameter records in the operation log and set u time nodes, mark the u time nodes at the same time in the vibration acceleration relationship diagram in each parameter record, and analyze the vibration acceleration in the three-axis direction at each time node; count the number of all parameter records n, and associate these parameter records two by two to obtain (n 2 -n)÷2 pairs of parameter records; analyze the vibration acceleration of each pair of parameter records in the three-axis directions at each time node, and associate the vibration acceleration of different parameter records at the same time node and in the same axis direction; S202, grouping the associated vibration accelerations according to the axial direction and establishing a training set for each group, the associated vibration accelerations in the same group are respectively used as independent variables and dependent variables and packaged as samples, and these u samples are placed in the training set corresponding to the group to which they belong; establishing a linear regression model, substituting the independent variables of the samples in each training set as input values into the linear regression model expression, calculating the difference between the output result of each sample and the dependent variable, summing the differences of all samples in the training set as the gap coefficient, and obtaining the minimum gap coefficient PL by adjusting the intercept F and the regression coefficient I in the expression min , thus obtaining the final expression: E=F+IL In the formula, E is the dependent variable and L is the independent variable; S203. Each pair of parameter records corresponds to three expressions. Calculate the standard deviation F of the intercepts in these three expressions. sd , standard deviation of regression coefficient I sd and the standard deviation of the minimum gap coefficient PL sd , substitute into the formula to calculate the influence coefficient PX; Set the coefficient threshold PX th , the influence coefficient is greater than PX th The encoding of each pair of parameter records is associated with the corresponding sensor module, and the associated sensor modules are marked with the influence relationship; The influence coefficient calculation formula is as follows: S204, establishing a relationship network, connecting each pair of associated sensor modules with annotated influence relationships; counting the number of connections between each sensor module and other sensor modules in the relationship network, and sorting the sensor modules in descending order according to the number of connections; Get the sensor modules MZ in order one The vibration acceleration in the three-axis direction collected in real time is substituted into the MZ one Associated sensor module MZ two In the final expression between, the expected vibration acceleration in the three-axis direction is calculated, and the sensor module MZ is calculated. two The difference between the vibration acceleration in the three-axis direction collected in real time and the expected vibration acceleration is set to the deviation threshold g. When the difference in the three-axis direction is greater than g, the sensor module MZ is two Mark an exception. S205, when with MZ one After all associated sensor modules are calculated, MZ is obtained in order of arrangement. one Subsequent sensor modules calculate the difference in vibration acceleration and mark the anomaly until all sensor modules in the arrangement order have completed the calculation and marked the anomaly. Then, the number of times each sensor module in the relationship network is marked as anomaly is counted, and the sensor module with the most number of marked anomalies and a number greater than M is selected as the suspicious sensor module.
4. A sensor damage detection method according to claim 3, characterized in that: In S300, the specific steps are as follows: S301. Mark the sensor modules that have an influencing relationship with the suspicious sensor modules in the relationship network, count the number k of all marked sensor modules, obtain the vibration acceleration in the three-axis direction collected by each marked sensor module in real time, and bring them into the relationship formula between the suspicious sensor module and the suspicious sensor module to calculate the expected vibration acceleration; collect the vibration acceleration in the three-axis direction of the suspicious sensor module, and bring it into the formula to calculate the abnormal index of the suspicious sensor module: In the formula, RT is the abnormality index, va i is the vibration acceleration of the suspicious sensor module in the i-th axis direction, Calculate the estimated vibration acceleration of the suspicious sensor module in the i-th axis direction for the h-th marked sensor module; S302, setting a sampling frequency SF and an interval time time, and calculating an abnormality index of a suspicious sensor module according to the sampling frequency SF; Set the exponential threshold RT th , when the abnormal index of the suspicious sensor module is greater than RT th When the duration is greater than time, the system enters the alarm state. The data collection frequency of the sensor module in the alarm state is increased and transmitted to the smart terminal first. The smart terminal uploads the sensor module data in the alarm state to the cloud platform and issues an abnormal alarm.
5. A sensor damage detection method according to claim 4, characterized in that: The cloud platform uses a large visual screen to display the mechanical equipment and specific sensor modules in alarm status in real time, collects and verifies the data uploaded by the alarm status sensor module, and pushes abnormal alarm information to management personnel for decision-making.
6. A sensor damage detection system, characterized in that: The system includes a data acquisition module, an abnormality analysis module, a damage detection module and an intelligent early warning module; The data acquisition module collects the operation log of the mechanical equipment and collects the vibration acceleration in three axes in real time through the sensor modules installed on the mechanical equipment; the abnormal analysis module analyzes the parameter records of each sensor module in the operation log, calculates the influence coefficient between the sensor modules and analyzes whether there is an influence relationship, establishes a relationship network and screens out suspicious sensor modules; The damage detection module calculates the abnormality index through the vibration acceleration of the suspicious sensor module, and determines whether to enter the alarm state based on the change of the abnormal index. In the alarm state, it quickly guides and issues abnormal alarms for the data in the sensor module; the intelligent early warning module displays the data information uploaded by the alarm device in real time through a large visual screen, and notifies the management personnel of the occurrence of abnormal events.
7. A sensor damage detection system according to claim 6, characterized in that: The data acquisition module includes a parameter information acquisition unit and an operation log acquisition unit; The parameter information acquisition unit collects the vibration acceleration in three-axis directions in real time through the sensor module installed on the mechanical equipment; The operation log collection unit is used to collect parameter records of each sensor module; the parameter records include codes and vibration acceleration relationship diagrams; the codes are used to distinguish different sensor modules, and the vibration acceleration relationship diagrams represent the vibration accelerations in the three-axis directions collected by the sensor modules at different times; the three-axis directions refer to the directions defined along three mutually perpendicular coordinate axes in three-dimensional space.
8. A sensor damage detection system according to claim 7, characterized in that: The anomaly analysis module includes a relationship analysis unit and an anomaly screening unit; The relationship analysis unit is used to analyze whether there is an influence relationship between sensor modules; First, set u time nodes and mark the u time nodes at the same time in the vibration acceleration relationship diagram in each parameter record; count the number of all parameter records n, and associate these parameter records two by two to obtain (n 2 -n)÷2 pairs of parameter records; analyze the vibration acceleration of each pair of parameter records in the three-axis directions at each time node, and associate the vibration acceleration of different parameter records at the same time node and in the same axis direction; Secondly, the associated vibration accelerations are grouped according to the axial direction and a training set is established for each group. The associated vibration accelerations in the same group are respectively used as independent variables and dependent variables and packaged as samples, and these u samples are placed in the training set corresponding to the group to which they belong; a linear regression model is established, and the independent variables of the samples are substituted into the linear regression model expression as input values, and the difference between the output result of each sample and the dependent variable is calculated. The difference of all samples in the training set is summed as the gap coefficient, and the minimum gap coefficient PL is obtained by adjusting the intercept F and the regression coefficient I in the expression. min , thus obtaining the final expression: E = F + IL, where E is the dependent variable and L is the independent variable; Finally, calculate the standard deviation F of the intercepts in the three expressions corresponding to each pair of parameter records sd , standard deviation of regression coefficient I sd and the standard deviation of the minimum gap coefficient PL sd , according to the formula Calculate the influence coefficient; Set the coefficient threshold PX th , the influence coefficient is greater than PX th The encoding of each pair of parameter records is associated with the corresponding sensor module, and the associated sensor modules are marked with the influence relationship; The abnormal screening unit is used to screen suspicious sensor modules; First, a relationship network is established, and each pair of associated sensor modules with an impact relationship is connected; the number of connections between each sensor module and other sensor modules in the relationship network is counted, and the sensor modules are sorted in descending order according to the number of connections; Secondly, obtain the sensor modules MZ in order of arrangement one The vibration acceleration in the three-axis direction collected in real time is substituted into the MZ one Associated sensor module MZ two In the final expression between, the expected vibration acceleration in the three-axis direction is calculated, and the sensor module MZ is calculated. two The difference between the vibration acceleration in the three-axis direction collected in real time and the expected vibration acceleration is set to the deviation threshold g. When the difference in the three-axis direction is greater than g, the sensor module MZ is two Marking anomalies; Finally, when with MZ one After all associated sensor modules are calculated, MZ is obtained in order of arrangement. one Subsequent sensor modules calculate the difference in vibration acceleration and mark the anomaly until all sensor modules in the arrangement order have completed the calculation and marked the anomaly. Then, the number of times each sensor module in the relationship network is marked as anomaly is counted, and the sensor module with the most number of marked anomalies and a number greater than M is selected as the suspicious sensor module.
9. A sensor damage detection system according to claim 8, characterized in that: The damage detection module includes an index calculation unit and a data grooming unit; The index calculation unit is used to calculate the abnormal index of the suspicious sensor module; first mark the sensor modules that have an influence relationship with the suspicious sensor module in the relationship network, count the number k of all marked sensor modules, obtain the vibration acceleration in the three-axis direction collected in real time by each marked sensor module, and bring them into the relationship formula between the suspicious sensor module and the module to calculate the expected vibration acceleration; Then collect the vibration acceleration of the suspicious sensor module in the three-axis direction according to the formula: Calculate the abnormal index of the suspicious sensor module; where va i is the vibration acceleration of the suspicious sensor module in the i-th axis direction, Calculate the estimated vibration acceleration of the suspicious sensor module in the i-th axis direction for the h-th marked sensor module; The data grooming unit conducts rapid grooming and abnormal alarm for the data in the sensor module. First, the sampling frequency SF and the interval time are set, and the abnormal index of the suspicious sensor module is calculated according to the sampling frequency SF. Next, set the index threshold RT th , when the abnormal index of the suspicious sensor module is greater than RT th And when the duration is greater than time, it enters the alarm state; finally, the data collection frequency of the sensor module in the alarm state is increased and transmitted to the smart terminal first. The smart terminal uploads the sensor module data in the alarm state to the cloud platform and issues an abnormal alarm.
10. A sensor damage detection system according to claim 9, characterized in that: The intelligent early warning module displays the mechanical equipment and sensor modules in alarm status through a large visual screen, obtains the data uploaded by the alarm status sensor module, and pushes the abnormal alarm information to the management personnel for decision-making.
Citation Information
Cited By
Active damping method and system
CN120521855A