Deburring equipment fault diagnosis system and method based on artificial intelligence

Abstract

The invention discloses a deburring equipment fault diagnosis system and method based on artificial intelligence, and relates to the technical field of equipment fault diagnosis, and the diagnosis method comprises the following steps: obtaining a historical fault record of deburring equipment, and extracting an abnormal feature when the deburring equipment sends an abnormal prompt; classifying historical fault records of the deburring equipment to obtain a plurality of fault types; analyzing operation parameters of the deburring equipment in each fault record, and calculating a total distribution value of abnormal characteristics; carrying out weight distribution on the extracted abnormal features, and calculating to obtain a risk value of the fault record; the real-time abnormal value of the deburring equipment is calculated, and the probability that the deburring equipment is abnormal is obtained; if the probability that the deburring equipment is abnormal exceeds a set risk threshold value, whether the deburring equipment is abnormal or not is analyzed; and adjusting the weight for distributing each abnormal feature in combination with the change trend of each abnormal feature.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment fault diagnosis, and specifically to a deburring equipment fault diagnosis system and method based on artificial intelligence. Background Art

[0002] A deburring device is a device used in the manufacturing industry to remove burrs generated on workpieces; burrs are unwanted, rough or sharp edges that appear when cutting or machining workpieces; the deburring device removes the burrs on the workpiece surface through various methods, such as mechanical deburring, chemical deburring or electronic deburring, etc., thereby improving the quality and safety of the final product.

[0003] Usually, the deburring device may not work properly due to mechanical component wear, sensor or controller failures, working environment and improper operation, etc.; when the deburring device malfunctions, the staff cannot detect the abnormality in the first place; and when the abnormality is detected, they cannot quickly know the specific fault that caused the abnormality, and the staff needs to check one by one, which seriously affects the efficiency. At the same time, the deburring device may also be further damaged because the fault cannot be detected in time. Summary of the Invention

[0004] The purpose of the present invention is to provide a deburring equipment fault diagnosis system and method based on artificial intelligence to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A deburring equipment fault diagnosis method based on artificial intelligence, the diagnosis method includes the following steps:

[0006] Step S100: Set the process of the deburring device sending an abnormal reminder once as a fault record; obtain the historical fault records of the deburring device, and extract the abnormal characteristics when the deburring device sends an abnormal reminder; classify the historical fault records of the deburring device according to the differences between the abnormal characteristics to obtain several fault types.

[0007] Step S200: Analyze the operating parameters of the deburring device in each fault record, and calculate the total assigned value of the abnormal characteristics of each fault record; according to the total assigned value of the abnormal characteristics and the fault type corresponding to the fault record, assign weights to the extracted abnormal characteristics, and calculate the risk value of the fault record.

[0008] Step S300: Obtain the operating parameters of the deburring device during operation every unit time, calculate the real-time abnormal value of the deburring device, and obtain the probability of the deburring device having an abnormality; if the probability of the deburring device having an abnormality exceeds the set risk threshold, analyze whether the deburring device has an abnormality.

[0009] Step S400: According to the actual operation time of the deburring device and the change trend of each abnormal feature, the weight assigned to each abnormal feature is adjusted.

[0010] Furthermore, step S100 includes the following steps:

[0011] Step S101: Set the i-th fault record to A i , get fault record A i The deburring equipment is configured to generate a distribution diagram of each operating parameter changing with time; the data of each distribution diagram is screened by using a box plot, and the upper and lower boundaries of the box plot are set. If the values of some data in the distribution diagram do not fall within the range of the lower and upper boundaries, the total time length t of the data is obtained. y , we get the data anomaly ratio α=t in the distribution diagram y / t, where t is the operating time of the deburring device; setting an abnormal threshold β, if α>β, extracting abnormal features from the operating parameters corresponding to the distribution diagram;

[0012] Step S102: Obtain fault record A i All abnormal features in the fault record A are obtained i The abnormal feature set B i ; Compare the abnormal feature sets of historical fault records with each other. If there are several fault records with the same number of abnormal features and the same abnormal features in the abnormal feature sets, the several fault records are classified into the same fault type; classify the historical fault records, considering that the same fault types are relatively close in the subsequent weight allocation and abnormal value calculation, which is conducive to more reasonable and accurate subsequent calculations.

[0013] Further, step S200 includes the following steps:

[0014] Step S201: Set the i-th fault record A i The abnormal feature set B i The number of abnormal features in is m, and the jth abnormal feature is selected, and the number of occurrences of the jth abnormal feature in the historical abnormal feature set is q j , the occurrence frequency of the jth abnormal feature is f j =q j / x, where x is the number of historical abnormal feature sets. A high frequency of occurrence of an abnormal feature indicates that the abnormal feature has a great impact on the abnormality of the device. Calculating the frequency of occurrence is conducive to more accurate subsequent weight allocation.

[0015] Step S202: Obtain fault record A iThe number of records for the same fault type is p, and the fault record A is obtained i The occurrence frequency is d i = p / x; According to the formula:

[0016]

[0017] The total assigned value Y of the abnormal characteristics of the fault record A is calculated i ; Combining with the occurrence frequency of the same fault type can illustrate the occurrence frequency of this fault, and then adjust the total assigned value Y of the abnormal characteristics i , making the result more accurate; i

[0018] Step S203: Obtain all historical different abnormal characteristics as the operating characteristic set of the deburring device. Set the number of operating characteristics in the operating characteristic set of the deburring device as n. According to the formula:

[0019]

[0020] Among them, k is the kth characteristic in the operating characteristic set, and Ju(k) is a judgment function to judge whether the kth characteristic in the operating characteristic set is an abnormal characteristic of the fault record A i If the kth characteristic is an abnormal characteristic, then Ju(k) = 1; otherwise, Ju(k) = 0; f k is the occurrence frequency of the kth characteristic when the kth characteristic is an abnormal characteristic; the assignment weight Q of the fault record A i assigned to the kth characteristic is calculated k ;

[0021] The assignment of weights includes all characteristics, so it is necessary to assign a part of the weights to the normal characteristics as well; since it is the assignment of the total assigned value of the abnormal characteristics, and the total assigned value of the abnormal characteristics is calculated based on the abnormal characteristics without involving other normal characteristics, it is necessary to add one to every other characteristic to ensure that the weight assigned to the normal characteristics is not 0, which is beneficial to a more reasonable weight assignment;

[0022] Step S204: Obtain the operating time T of the deburring device in the fault record A i According to the formula: i

[0023]

[0024] The abnormal value Z of the fault record A is calculated i ; i

[0025] ​​​Step S205: Obtain the outliers of each historical fault record, and select the outlier with the smallest value as the anomaly detection threshold Z of the deburring device max 。

[0026] Further, step S300 includes the following steps:

[0027] Step S301: Detect the operating parameters corresponding to each operating characteristic in the deburring device every unit time Δt, and select the operating parameters of the k-th operating characteristic; Use a box plot to view the distribution of the operating parameters, and obtain the abnormal time length of the values that are not between the upper and lower boundaries in one unit time. Among them, let the abnormal time length in the s-th unit time be t ’ s ,Calculate the abnormal time proportion of the operating parameters of the k-th operating characteristic as where r is the number of unit times elapsed;

[0028] Step S302: Obtain the actual operating time T of the deburring device s ,According to the formula:

[0029]

[0030] Calculate the real-time outlier Z of the deburring device s ;If Z s >Z max ,where Z max is the anomaly detection threshold, then count the number of records with outliers less than the real-time outlier Z s in the historical x fault records as x ’ ;According to the comparison of the size of the real-time outlier in all historical fault records, the possibility of abnormality in the deburring device can be preliminarily confirmed, which affects the subsequent anomaly probability calculation;

[0031] Step S303: Set the anomaly threshold as β, and obtain the average abnormal time proportion of the operating parameters of the k-th operating characteristic as (ε k ) ave =ε k / r. If (ε k ) ave >β, then regard the k-th operating characteristic as an abnormal characteristic; Obtain all abnormal characteristics, generate the actual abnormal characteristic set of the deburring device, compare the actual abnormal characteristic set with the abnormal characteristic set of the historical fault records, and determine the fault type corresponding to the actual abnormal characteristic set; Similarly, compare the values of the same fault type to further confirm the possibility of device abnormality, making the subsequent anomaly probability more accurate;

[0032] Step S304: Count the outliers less than the real-time outlier Z in the same fault types The number of records is x ” , let the number of fault records in the same fault type be y, according to the formula:

[0033]

[0034] Calculate the probability G of the deburring device having an abnormality; set a risk threshold τ. When G > τ, monitor the set of abnormal characteristics of the deburring device. If the values of several abnormal characteristics are not always between the upper and lower boundaries, send an abnormal reminder to the deburring device; predict the abnormal probability through the number of abnormal values greater than the historical abnormal values and the number of abnormal values greater than those in the same fault type, which is conducive to more accurate results; judge whether an abnormality occurs through the predicted probability, which is conducive to reducing the workload of the staff.

[0035] Furthermore, step S400 includes the following steps:

[0036] Step S401: Let the actual running time of the deburring device when an abnormal reminder is sent be T s , obtain the distribution diagram of the operating parameters corresponding to each abnormal characteristic in the set of abnormal characteristics changing with time, and extract the device running time T when the operating parameters corresponding to the j-th abnormal characteristic start to show abnormalities ’ j , obtain the abnormal running time of the operating parameters corresponding to the j-th abnormal characteristic as T ” j = T s - T ’ j ;

[0037] Step S402: Obtain the allocation weight Q j of the j-th abnormal characteristic and the abnormal running time of each abnormal characteristic, according to the formula:

[0038]

[0039] where m is the number of abnormal characteristics; calculate the new allocation weight Q ’ j ; There are two situations for the deburring device to have an abnormality. One is that there is an abnormality but the device can still run, and the other is that the device cannot run; calculating the abnormal value of the device can avoid the situation where the device cannot run, but it cannot ensure that the abnormality is detected when the device just starts to have an abnormality. Therefore, it is necessary to continuously adjust the allocation weight. When the abnormal detection threshold remains unchanged, increasing the weight can reach the abnormal detection threshold with the reduction of the device running time, and the device abnormality can be detected earlier.

[0040] To better implement the above method, a deburring equipment fault diagnosis system is also proposed. The diagnosis system includes a fault classification module, a weight assignment module, a fault detection module, and a judgment and adjustment module;

[0041] The fault classification module is used to set the process of the deburring equipment sending an abnormal reminder once as a fault record; obtain the historical fault records of the deburring equipment, and extract the abnormal characteristics when the deburring equipment sends an abnormal reminder; classify the historical fault records of the deburring equipment according to the differences between the abnormal characteristics to obtain several fault types;

[0042] The weight assignment module is used to analyze the operating parameters of the deburring equipment in each fault record, calculate the total assigned value of the abnormal characteristics of each fault record; according to the total assigned value of the abnormal characteristics and the fault type corresponding to the fault record, assign weights to the extracted abnormal characteristics, and calculate the risk value of the fault record;

[0043] The fault detection module is used to obtain the operating parameters of the deburring equipment during operation every other unit time, calculate the real-time abnormal value of the deburring equipment, and obtain the probability of the deburring equipment having an abnormality; if the probability of the deburring equipment having an abnormality exceeds the set risk threshold, analyze whether the deburring equipment has an abnormality;

[0044] The judgment and adjustment module is used to adjust the weights assigned to each abnormal characteristic according to the actual operating time of the deburring equipment and in combination with the change trends of each abnormal characteristic.

[0045] Furthermore, the fault classification module includes a feature extraction unit and a type division unit;

[0046] The feature extraction unit is used to set the process of the deburring equipment sending an abnormal reminder once as a fault record; obtain the historical fault records of the deburring equipment, and extract the abnormal characteristics when the deburring equipment sends an abnormal reminder; the type division unit is used to classify the historical fault records of the deburring equipment according to the differences between the abnormal characteristics to obtain several fault types.

[0047] Furthermore, the weight assignment module includes a total assigned value calculation unit and a risk value calculation unit;

[0048] The total assigned value calculation unit is used to analyze the operating parameters of the deburring equipment in each fault record, calculate the total assigned value of the abnormal characteristics of each fault record; the risk value calculation unit is used to assign weights to the extracted abnormal characteristics according to the total assigned value of the abnormal characteristics and the fault type corresponding to the fault record, and calculate the risk value of the fault record.

[0049] Furthermore, the fault detection module includes a probability calculation unit and an abnormality judgment unit;

[0050] A probability calculation unit is configured to obtain the operating parameters of the deburring device every other unit time, calculate the real-time anomaly value of the deburring device, and obtain the probability of the deburring device having an anomaly; an anomaly judgment unit is configured to analyze whether the deburring device has an anomaly if the probability of the deburring device having an anomaly exceeds a set risk threshold.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0052] 1. The present invention judges whether the device has an anomaly by calculating the anomaly value of the deburring device, can perform fault diagnosis automatically, reduces the workload of manual troubleshooting, and can understand the operating conditions of the device faster at the same time;

[0053] 2. The present invention classifies the fault types based on the historical fault records, which is beneficial for subsequent weight assignment and calculation of anomaly values;

[0054] 3. The present invention adjusts the weight assignment according to the time difference between the time when the device starts to have an anomaly and the time when the anomaly reminder is sent, which can prompt the discovery of the abnormal behavior of the device earlier, and can detect the anomaly as early as possible when the device just starts to have an anomaly. Description of the Drawings

[0055] Figure 1 It is a schematic diagram of the steps of a deburring device fault diagnosis method based on artificial intelligence;

[0056] Figure 2 It is a schematic diagram of the structure of a deburring device fault diagnosis system based on artificial intelligence. Detailed Embodiments

[0057] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0058] Embodiment: As Figures 1 to 2 shown, the present invention provides a deburring device fault diagnosis method based on artificial intelligence, and the diagnosis method includes the following steps:

[0059] Step S100: Set the process of the deburring device sending an anomaly reminder once as a fault record; obtain the historical fault records of the deburring device, and extract the anomaly features when the deburring device sends an anomaly reminder; classify the historical fault records of the deburring device according to the differences between the anomaly features to obtain several fault types;

[0060] Wherein, step S100 includes the following steps:

[0061] Step S101: Set the i-th fault record to A i , get fault record A i The deburring equipment is configured to generate a distribution diagram of each operating parameter changing with time; the data of each distribution diagram is screened by using a box plot, and the upper and lower boundaries of the box plot are set. If the values of some data in the distribution diagram do not fall within the range of the lower and upper boundaries, the total time length t of the data is obtained. y , we get the data anomaly ratio α=t in the distribution diagram y / t, where t is the operating time of the deburring device; setting an abnormal threshold β, if α>β, extracting abnormal features from the operating parameters corresponding to the distribution diagram;

[0062] Step S102: Obtain fault record A i All abnormal features in the fault record A are obtained i The abnormal feature set B i ; The abnormal feature sets of historical fault records are compared with each other. If there are several fault records with the same number of abnormal features and the same abnormal features in the abnormal feature sets, the several fault records are classified as the same fault type.

[0063] Step S200: analyzing the operating parameters of the deburring equipment in each fault record, and calculating the total distribution value of the abnormal features of each fault record; performing weight distribution on the extracted abnormal features according to the total distribution value of the abnormal features and the fault type corresponding to the fault record, and calculating the risk value of the fault record;

[0064] Wherein, step S200 includes the following steps:

[0065] Step S201: Set the i-th fault record A i The abnormal feature set B i The number of abnormal features in is m, and the jth abnormal feature is selected, and the number of occurrences of the jth abnormal feature in the historical abnormal feature set is q j , the occurrence frequency of the jth abnormal feature is f j =q j / x, where x is the number of historical anomaly feature sets;

[0066] Step S202: Obtain fault record A i The number of records of the same fault type is p, and the fault record A is obtained i The frequency of occurrence is d i =p / x; According to the formula:

[0067]

[0068] Calculate the fault record A i The total assigned value Y of abnormal features i ;

[0069] Step S203: Obtain all historical different abnormal features as the operating feature set of the deburring device. Set the number of operating features in the operating feature set of the deburring device to n. According to the formula:

[0070]

[0071] where k is the k-th feature in the operating feature set, and Ju(k) is a judgment function to judge whether the k-th feature in the operating feature set is an abnormal feature of the fault record A i in. If the k-th feature is an abnormal feature, then Ju(k)=1; otherwise, Ju(k)=0; f k is the occurrence frequency of the k-th feature when the k-th feature is an abnormal feature; calculate the fault record A i The assigned weight Q assigned to the k-th feature k ;

[0072] Step S204: Obtain the operating time T of the deburring device in the fault record A i and calculate the abnormal value Z of the fault record A according to the formula: i

[0073]

[0074] Calculate the fault record A i The abnormal value Z i ;

[0075] Step S205: Obtain the abnormal values of each historical fault record, and select the abnormal detection threshold Z of the deburring device with the smallest value max .

[0076] Step S300: Obtain the operating parameters of the deburring device every unit time, calculate the real-time abnormal value of the deburring device, and obtain the probability of the deburring device having an abnormality; if the probability of the deburring device having an abnormality exceeds the set risk threshold, analyze whether the deburring device has an abnormality;

[0077] Among them, step S300 includes the following steps:

[0078] ​Step S301: Detect the operating parameters corresponding to each operating feature in the deburring device every unit time Δt, and select the operating parameters of the k-th operating feature; use a box plot to view the distribution of the operating parameters, and obtain the abnormal time length of the values that are not between the upper and lower boundaries in one unit time. Among them, let the abnormal time length in the s-th unit time be t ’ s , and calculate that the abnormal time proportion of the operating parameters of the k-th operating feature is where r is the number of unit times elapsed;

[0079] Step S302: Obtain the actual operating time T of the deburring device s , according to the formula:

[0080]

[0081] calculate the real-time abnormal value Z of the deburring device s ; if Z s > Z max , where Z max is the abnormal detection threshold, then count the number of records with abnormal values less than the real-time abnormal value Z s in the historical x fault records as x ’ ;

[0082] Step S303: Set the abnormal threshold as β, and obtain the average abnormal time proportion of the operating parameters of the k-th operating feature as (ε k ) ave = ε k / r. If (ε k ) ave > β, then regard the k-th operating feature as an abnormal feature; obtain all abnormal features, generate the actual abnormal feature set of the deburring device, and compare the actual abnormal feature set with the abnormal feature set of the historical fault records to determine the fault type corresponding to the actual abnormal feature set;

[0083] Step S304: Count the number of records with abnormal values less than the real-time abnormal value Z s in the same fault type as x ” , let the number of fault records in the same fault type be y, according to the formula:

[0084]

[0085] calculate the probability G of the deburring device having an abnormality; set a risk threshold τ. When G > τ, monitor the abnormal feature set of the deburring device. If there are several abnormal feature values that are always not between the upper and lower boundaries, send an abnormal reminder to the deburring device.

[0086] Step S400: Adjust the weight assigned to each abnormal feature according to the actual running time of the deburring device and the changing trend of each abnormal feature.

[0087] Among them, step S400 includes the following steps:

[0088] Step S401: Let the actual running time when the deburring device sends an abnormal reminder be T s , obtain the distribution diagram of the operating parameters corresponding to each abnormal feature in the abnormal feature set changing with time, and extract the device running time T when the operating parameters corresponding to the j-th abnormal feature start to show abnormalities ’ j , and obtain the abnormal running time of the operating parameters corresponding to the j-th abnormal feature as T ” j = T s - T ’ j ;

[0089] Step S402: Obtain the assigned weight Q j of the j-th abnormal feature and the abnormal running time of each abnormal feature. According to the formula:

[0090]

[0091] Among them, m is the number of abnormal features; calculate the new assigned weight Q ’ j .

[0092] A deburring device fault diagnosis system, the diagnosis system includes a fault classification module, a weight assignment module, a fault detection module and a judgment and adjustment module;

[0093] The fault classification module is used to set the process of the deburring device sending an abnormal reminder as a fault record; obtain the historical fault records of the deburring device, and extract the abnormal features when the deburring device sends an abnormal reminder; classify the historical fault records of the deburring device according to the differences between the abnormal features to obtain several fault types;

[0094] The weight assignment module is used to analyze the operating parameters of the deburring device in each fault record, calculate the total assigned value of the abnormal features of each fault record; according to the total assigned value of the abnormal features and the fault type corresponding to the fault record, assign weights to the extracted abnormal features, and calculate the risk value of the fault record;

[0095] A fault detection module, which is used to obtain the operating parameters of the deburring device every other unit time, calculate the real-time anomaly value of the deburring device, and obtain the probability of the deburring device having an anomaly; if the probability of the deburring device having an anomaly exceeds the set risk threshold, analyze whether the deburring device has an anomaly.

[0096] A judgment and adjustment module, which is used to adjust the weight assigned to each anomaly feature according to the actual operating time of the deburring device and in combination with the change trend of each anomaly feature.

[0097] Among them, the fault classification module includes a feature extraction unit and a type division unit.

[0098] The feature extraction unit is used to set the process of the deburring device sending an anomaly reminder once as a fault record; obtain the historical fault records of the deburring device, and extract the anomaly features when the deburring device sends an anomaly reminder; the type division unit is used to classify the historical fault records of the deburring device according to the differences between the anomaly features, and obtain several fault types.

[0099] Among them, the weight assignment module includes a total assignment value calculation unit and a risk value calculation unit.

[0100] The total assignment value calculation unit is used to analyze the operating parameters of the deburring device in each fault record and calculate the total assignment value of the anomaly features of each fault record; the risk value calculation unit is used to assign weights to the extracted anomaly features according to the total assignment value of the anomaly features and the fault type corresponding to the fault record, and calculate the risk value of the fault record.

[0101] Among them, the fault detection module includes a probability calculation unit and an anomaly judgment unit.

[0102] The probability calculation unit is used to obtain the operating parameters of the deburring device every other unit time, calculate the real-time anomaly value of the deburring device, and obtain the probability of the deburring device having an anomaly; the anomaly judgment unit is used to analyze whether the deburring device has an anomaly if the probability of the deburring device having an anomaly exceeds the set risk threshold.

[0103] Obtain the historical 3 fault records. The first fault is temperature anomaly, and the anomaly feature is temperature; the second fault is signal strength anomaly, and the anomaly feature is signal strength; the third fault is data transmission rate anomaly, and the anomaly feature is transmission rate; therefore, the historical fault records are divided into 3 categories, and the occurrence frequency of each fault is 1 / 3, and the occurrence frequency in its respective fault type is 1. Therefore, calculate the total anomaly feature assignment value of the first fault as Y i =1×2×1 = 2, the total anomaly feature assignment value of the second fault is 2, and the total anomaly feature assignment value of the third fault is 2.

[0104] Obtain all different historical abnormal features as the operation feature set of the deburring device. Set the number of operation features in the operation feature set of the deburring device to n = 3, and calculate the distribution weight Q of the temperature feature of the first failure k =(1 + 1 / 3) / 4 = 0.67, the distribution weight Q of the signal strength feature of the first failure k = 1 / 4×2 = 0.5, and the distribution weight of the transmission rate feature of the first failure is 0.5; calculate that the outlier of the first failure is Zi = 100×1.67 = 167, where the actual operation time is 100 time points, and so on, to obtain the outliers of the second and third failures, and set the outlier of the first failure as the anomaly detection threshold;

[0105] If the outlier of the deburring device obtained in real time is 170, then analyze the abnormal conditions of each feature of the deburring device, and it is detected that it is a temperature anomaly, so the abnormal feature is temperature; calculate the probability G of the deburring device having an anomaly = 2 / 3, and set the risk threshold to 0.5. Therefore, monitor the temperature change of the deburring device and send an anomaly reminder;

[0106] Set the actual operation time of the deburring device when sending the anomaly reminder to 100 time points, and the operation time when the device starts to have an anomaly to 80 time points, and obtain the abnormal operation time of the deburring device as 20 time points. According to the formula: Q ’ j = 0.67×(1 + 20 / 20) = 1.34, to obtain the new distribution weight of the temperature feature of the deburring device.

[0107] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, in any aspect, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights involved.

Claims

1. A method for fault diagnosis of a deburring device based on artificial intelligence, characterized in that: The described diagnostic method includes the following steps: Step S100: Set the process of the deburring device sending an abnormal reminder once as a fault record; obtain the historical fault records of the deburring device, and extract the abnormal features when the deburring device sends an abnormal reminder; classify the historical fault records of the deburring device according to the differences between the abnormal features to obtain several fault types; Step S200: Analyze the operating parameters of the deburring device in each fault record, and calculate the total allocation value of the abnormal features of each fault record; according to the total allocation value of the abnormal features and the fault type corresponding to the fault record, allocate weights to the extracted abnormal features, and calculate the risk value of the fault record; Step S300: Obtain the operating parameters of the deburring device during operation every other unit time, calculate the real-time abnormal value of the deburring device, and obtain the probability of the deburring device having an abnormality; if the probability of the deburring device having an abnormality exceeds the set risk threshold, analyze whether the deburring device has an abnormality; Step S400: Adjust the weights assigned to each abnormal feature according to the actual operating time of the deburring device and the change trend of each abnormal feature.

2. The fault diagnosis method of a deburring device based on artificial intelligence according to claim 1, characterized in that: The described Step S100 includes the following steps: Step S101: Set the i-th fault record as A i , and obtain the fault record A i For several operating parameters of the deburring device in it, generate a distribution map of each operating parameter changing with time; use a box plot to screen the data of each distribution map, set the upper and lower boundaries of the box plot. If there are several data values in the distribution map that do not fall within the range of the lower and upper boundaries, obtain the total duration t y , and obtain the abnormal data ratio α = t y / t, where t is the operating duration of the deburring device; set an abnormality threshold β. If α > β, extract abnormal features from the operating parameters corresponding to the distribution map; Step S102: Obtain the fault record A i All the abnormal features in it to obtain the fault record A i The abnormal feature set B i ; Compare the abnormal feature sets of the historical fault records with each other. If there are several fault records with the same number of abnormal features and the same abnormal features in the abnormal feature sets, then divide the several fault records into the same fault type.

3. A method for fault diagnosis of a deburring device based on artificial intelligence according to claim 2, characterized in that: The described Step S200 includes the following steps: Step S201: Set the i-th fault record A i The abnormal feature set B i in it has m abnormal features. Select the j-th abnormal feature among them, and obtain the number q of occurrences of the j-th abnormal feature in the historical abnormal feature set j , and obtain the occurrence frequency f of the j-th abnormal feature j = q j / x, where x is the number of the historical abnormal feature set; Step S202: Obtain the number p of records of the same fault type as the fault record A i to obtain the occurrence frequency d of the fault record A i = p / x, and according to the formula: i ​ Calculate the fault record A i The total allocated value Y of abnormal features i ; Step S203: Obtain all different historical abnormal features as the operating feature set of the deburring device, set the number of operating features in the operating feature set of the deburring device to n, and calculate to obtain the fault record A i The allocation weight Q assigned to the k-th feature k ; Step S204: Obtain the fault record A i The running time of the deburring device in i is T i , according to the formula: Calculated fault record A i of the outlier Z i ; Step S205: Obtain the outliers of each historical failure record, and select the outlier with the smallest value as the anomaly detection threshold Z of the deburring device max .

4. The method for diagnosing faults of a deburring device based on artificial intelligence according to claim 3, wherein: The described Step S300 includes the following steps: Step S301: Detect the operating parameters corresponding to each operating feature in the deburring device every unit time Δt, and select the operating parameters of the k-th operating feature; use a box plot to view the distribution of the operating parameters, and obtain the abnormal time length of the values that are not between the upper and lower boundaries in one unit time. Among them, let the abnormal time length in the s-th unit time be t ’ s , and calculate that the abnormal time proportion of the operating parameters of the k-th operating feature is where r is the number of unit times elapsed; Step S302: Obtain the actual running time T of the deburring device s , according to the formula: Calculate the real-time outlier Z of the deburring device s ; If Z s > Z max , where Z max is the anomaly detection threshold, then count the number of records with outliers less than the real-time outlier Z s in the historical x fault records as x ’ ; Step S303: Set the anomaly threshold as β, and obtain that the average anomaly time proportion of the operating parameters of the k-th operating feature is (ε k ) ave = ε k / r. If (ε k ) ave > β, then regard the k-th operating feature as an abnormal feature; obtain all abnormal features, generate the actual abnormal feature set of the deburring device, compare the actual abnormal feature set with the abnormal feature set of the historical fault record, and determine the fault type corresponding to the actual abnormal feature set; Step S304: Count the number of records where the outlier in the same fault type is less than the real-time outlier Z s as x ” , let the number of fault records in the same fault type be y, according to the formula: Calculate the probability G of the deburring device having an abnormality; set a risk threshold τ, when G > τ, monitor the abnormal feature set of the deburring device, and if there are several abnormal feature values that are not always between the upper and lower boundaries, send an abnormal reminder to the deburring device.

5. A fault diagnosis method for a deburring device based on artificial intelligence according to claim 4, characterized in that: The described Step S400 includes the following steps: Step S401: Set the actual running time when the deburring device sends an exception reminder as T s , obtain the distribution diagram of the operating parameters corresponding to each exception feature in the exception feature set changing with time, and extract the device running time T when the operating parameters corresponding to the j-th exception feature start to show abnormalities ’ j , and obtain the abnormal running time of the operating parameters corresponding to the j-th exception feature as T ” j = T s - T ’ j ; Step S402: Obtain the allocation weight Q of the j-th abnormal feature j and the abnormal running time of each abnormal feature, and calculate the new allocation weight Q of the j-th abnormal feature ’ j .

6. A deburring equipment fault diagnosis system for executing a deburring equipment fault diagnosis method based on artificial intelligence according to any one of claims 1-5, characterized in that: The described diagnostic system includes a fault classification module, a weight allocation module, a fault detection module, and a judgment and adjustment module; The fault classification module is used to set the process of the deburring device sending an abnormal reminder once as a fault record; obtain the historical fault records of the deburring device, and extract the abnormal features when the deburring device sends an abnormal reminder; Classify the historical fault records of the deburring device according to the differences between the abnormal features to obtain several fault types; The weight allocation module is used to analyze the operating parameters of the deburring device in each fault record, and calculate the total allocation value of the abnormal features of each fault record; According to the total allocation value of the abnormal features and the fault type corresponding to the fault record, allocate weights to the extracted abnormal features, and calculate the risk value of the fault record; The fault detection module is used to obtain the operating parameters of the deburring device during operation every other unit time, calculate the real-time abnormal value of the deburring device, and obtain the probability of the deburring device having an abnormality; If the probability of the deburring device having an abnormality exceeds the set risk threshold, analyze whether the deburring device has an abnormality; The judgment and adjustment module is used to adjust the weights assigned to each abnormal feature according to the actual operating time of the deburring device and the change trend of each abnormal feature.

7. A deburring equipment fault diagnosis system according to claim 6, characterized in that: The fault classification module includes a feature extraction unit and a type division unit; The feature extraction unit is used to set the process of the deburring device sending an abnormal reminder as a fault record once; obtain the historical fault records of the deburring device, and extract the abnormal features when the deburring device sends an abnormal reminder. The type classification unit is used to classify the historical fault records of the deburring device according to the differences between the abnormal features, and obtain several fault types.

8. A deburring equipment fault diagnosis system according to claim 6, characterized in that: The weight assignment module includes a total assignment value calculation unit and a risk value calculation unit. The total assignment value calculation unit is used to analyze the operating parameters of the deburring device in each fault record and calculate the total assignment value of the abnormal features of each fault record. The risk value calculation unit is used to assign weights to the extracted abnormal features according to the total assignment value of the abnormal features and the fault type corresponding to the fault record, and calculate the risk value of the fault record.

9. The deburring equipment fault diagnosis system according to claim 6, characterized in that: The fault detection module includes a probability calculation unit and an abnormality judgment unit. The probability calculation unit is used to obtain the operating parameters of the deburring device during operation every other unit time, calculate the real-time abnormal value of the deburring device, and obtain the probability of the deburring device having an abnormality; the abnormality judgment unit is used to analyze whether the deburring device has an abnormality if the probability of the deburring device having an abnormality exceeds the set risk threshold.

Images