An intelligent supervision system and method for electromechanical equipment based on artificial intelligence

Abstract

The application discloses an intelligent supervision system and method for electromechanical equipment based on artificial intelligence, relates to the technical field of electromechanical equipment supervision, and comprises an attribute combination set generation module, a key matching rule output module, a necessary optimization index analysis module, a comprehensive service influence degree analysis module, a rule updating module and a secondary verification response module; the attribute combination set generation module is used for generating an equipment attribute combination set; the key matching rule output module is used for extracting a key matching rule of an electromechanical equipment monitoring event record; the necessary optimization index analysis module is used for matching a necessary optimization index of a rule; the comprehensive service influence degree analysis module is used for analyzing a comprehensive service influence degree of a gray scale rule in which equipment attributes of a to-be-analyzed equipment group are located; the rule updating module is used for updating a key matching rule of the to-be-analyzed equipment group under a corresponding attribute combination; and the secondary verification response module is used for performing secondary verification on a target electromechanical equipment group in which the key matching rule is replaced.

Description

Technical Field

[0001] This invention relates to the field of electromechanical equipment monitoring technology, specifically an intelligent monitoring system and method for electromechanical equipment based on artificial intelligence. Background Technology

[0002] In the field of electromechanical equipment supervision, especially in scenarios such as highways, electromechanical equipment is diverse, widely distributed, and significantly heterogeneous, encompassing various devices such as UPS, optical cables, information boards, and industrial control computers. The physical attributes, operating status, and business relevance of different devices vary significantly. Existing gray-scale release rule configurations largely rely on fixed dimensions or manual experience, lacking precise adaptation to equipment attributes and monitoring scenarios. On the one hand, attribute combination extraction fails to consider the core requirements of different monitoring scenarios, leading to weakened core attribute correlations, excessive redundant combinations, and impacting rule matching efficiency. On the other hand, the construction of candidate rule lists lacks systematicity, and the selection of key matching rules lacks unified quantitative standards, making it difficult to address multi-rule conflicts. Furthermore, existing technologies do not fully integrate historical data for comprehensive rule adaptability assessment, lacking multi-dimensional considerations of business continuity, segmentation accuracy, gray-scale target achievement, and system performance stability, resulting in highly subjective rule selection and insufficient effectiveness. In addition, the lack of a closed-loop verification mechanism after rule application prevents timely detection and dynamic optimization of adaptation issues, thus affecting the efficiency and security of gray-scale releases and failing to meet the refined and intelligent requirements of electromechanical equipment supervision. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent monitoring system and method for electromechanical equipment based on artificial intelligence, so as to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring method for electromechanical equipment based on artificial intelligence, the method comprising:

[0005] Step S100: Traverse historical data, extract attribute combinations of electromechanical equipment under different monitoring scenarios, and generate equipment attribute combination set; associate and mark the target electromechanical equipment groups monitored by each attribute combination in the equipment attribute combination set;

[0006] Step S200: Construct a list of candidate rules from the grayscale release rules of the independent attributes of each attribute combination in the device attribute combination set, and extract the key matching rules of the electromechanical equipment monitoring event records; the electromechanical equipment monitoring event refers to the grayscale release application event corresponding to the same target electromechanical equipment group containing the same grayscale release rule; calculate the necessary optimization index of the matching release rules for all electromechanical equipment monitoring events corresponding to the same target electromechanical equipment group;

[0007] Step S300: Set the necessary optimization index threshold, and mark the target electromechanical equipment group whose necessary optimization index is greater than or equal to the necessary optimization index threshold as the equipment group to be analyzed; analyze the comprehensive business impact of the gray scale rules of the equipment attributes of the equipment group to be analyzed;

[0008] Step S400: Based on the comprehensive business impact verification, update the key matching rules of the equipment group to be analyzed under the corresponding attribute combination, and provide early warning recommendations when responding to the candidate rule column in real time; perform secondary verification on the target electromechanical equipment group with the changed key matching rules.

[0009] Furthermore, step S100 includes the following processes:

[0010] Step S110: Construct an attribute dimension system for electromechanical equipment. The attribute dimension system includes physical attribute dimensions, operational attribute dimensions, business association dimensions, and environmental attribute dimensions.

[0011] Step S120: Traverse the historical monitoring data of electromechanical equipment and extract the attribute combinations under different monitoring scenarios. The monitoring scenarios include UPS fault prediction scenario, optical cable interruption monitoring scenario, information board display abnormality scenario and industrial control computer operation status monitoring scenario.

[0012] Step S130: During the attribute combination extraction process, perform differentiated calculations on the attribute association strength under different scenarios; the differentiated calculations include the following:

[0013] The basic correlation degree R(A,B) of any two device attributes in the attribute combination co-occurring in historical data is calculated as follows: R(A,B)=P(A∩B) / P(A)*P(B), where P(A) represents the probability of device attribute A occurring, P(B) represents the probability of device attribute B occurring, and P(A∩B) represents the probability of device attribute A and device attribute B occurring in the same attribute combination.

[0014] Assign scene weight S to all monitoring scenarios for historical data storage; and distinguish between core attributes and non-core attributes for the device attributes contained in each monitoring scenario, assigning weight W1 to core attributes and weight W2 to non-core attributes, where W1>W2;

[0015] Then calculate the attribute differential correlation strength S for each monitoring scenario. AB S AB =R(A,B)*S*W, W={W1,W2};

[0016] When the number of independent device attributes contained in an attribute combination is greater than 2, the differential association strength of the corresponding attribute combination is calculated by the geometric mean of the pairwise attribute association strengths.

[0017] Step S140: When SAB When S is greater than or equal to the minimum association strength threshold of the scene, the corresponding attribute combination is valid. AB When the minimum association strength threshold of the scene is reached, the corresponding attribute combination is deleted; all valid attribute combinations are retained, and attribute combinations consisting of different independent attributes are classified into one category to generate a set of device attribute combinations; the set of device attribute combinations is stored in the form of a triple of "scene ID-attribute combination code-associated device feature"; and a mapping relationship with the target electromechanical equipment group is established to complete the association marking.

[0018] Furthermore, step S200 includes the following specific steps:

[0019] Step S210: Establish a basic template library for grayscale release rules. The template library contains basic rule templates corresponding to different attribute dimensions. For each attribute combination in the device attribute combination set, extract the independent attributes contained in each attribute combination, and match the basic rule templates corresponding to each independent attribute from the basic rule template library. Arrange all independent basic rule templates of the same attribute combination to generate a list of candidate rules.

[0020] Step S220: Filter events from the electromechanical equipment monitoring event records that meet the same target electromechanical equipment group, the same gray-scale release rules, and a gray-scale execution success rate greater than or equal to the success rate threshold. Define these as the same type of valid monitoring events. Gray-scale execution success rate = (number of fault-free operating devices during the gray-scale period + number of fault-repaired devices) / total number of gray-scale devices recorded for the target electromechanical equipment group * 100%; extract the basic rule templates applied in the same type of valid monitoring events as key matching rules.

[0021] Step S230: Taking the effective monitoring events as the analysis object, extract the business continuity assessment value C1, the segmentation accuracy assessment value C2, the gray target achievement assessment value C3, and the system performance stability assessment value C4 of the effective monitoring event records;

[0022] The business continuity assessment value C1 is calculated as follows: C1 = 1 - (core business interruption duration / total grayscale duration) * 0.5 - (number of false alarms / total number of alarms) * 0.3 - (business recovery delay duration / standard recovery duration) * 0.2; the value ranges from [0,1], and the larger the value, the better the business continuity.

[0023] The segmentation accuracy evaluation value C2 = rule matching accuracy * 0.4 + target device coverage * 0.3 + reciprocal of non-target device missegmentation rate * 0.3; where rule matching accuracy = number of target devices that meet the rule / total number of devices that match the rule * 100%; target device coverage = number of target devices that match the rule / total number of target devices * 100%; non-target device missegmentation rate = number of non-target devices that are mismatched by the rule / total number of devices that match the rule * 100%; the value of C2 ranges from [0,1], and the larger the value, the better the segmentation accuracy.

[0024] The gray-scale target achievement evaluation value C3 = Adaptability verification compliance rate × 0.5 + Risk control compliance rate × 0.25 + Gray-scale cycle compliance rate × 0.25; where, Adaptability verification compliance rate = Number of new version compatible devices in gray-scale devices / Total number of gray-scale devices × 100%, Risk control compliance rate = Number of times the gray-scale ratio of core devices is less than or equal to the preset threshold / Total number of executions × 100%, Gray-scale cycle compliance rate = Actual gray-scale verification time / Planned gray-scale cycle × 100%; C3 ranges from [0,1], and the larger the value, the better the target achievement.

[0025] The system performance stability evaluation value C4 is calculated as follows: C4 = 1 - (duration of exceeding resource utilization limit / total gray-scale duration) × 0.4 - (number of times routing response delay exceeds limit / total number of requests) × 0.3 - (number of system errors / total number of executions) × 0.3; the value range is [0,1]. The larger the value, the more stable the system performance.

[0026] Step S240: Calculate the rule comprehensive fit index A, A=C1*0.38+C2*0.32+C3*0.18+C4*0.12; the larger A is, the better the rule fit; define the necessary optimization index Q for the effective monitoring events corresponding to the same target electromechanical equipment group, Q=1-A0, where A0 represents the average comprehensive fit index of all effective monitoring events recorded for the same target electromechanical equipment group.

[0027] Furthermore, analyzing the overall business impact of the group of devices under analysis, including the grayscale rules of the device attributes, includes the following steps:

[0028] Step S310: Extract the number m of gray-scale rules contained in the group of devices to be analyzed. Each gray-scale rule is measured by n evaluation indicators, and an initial indicator matrix X is constructed.

[0029]

[0030] Where x ij This represents the original data of the i-th grayscale rule on the j-th metric; x ij ={x 11 ,x 12 ,...,xmn};for x ij Standardization yields the standardized matrix Z, z ij This represents the data of the i-th grayscale rule after standardization on the j-th metric;

[0031]

[0032] Step S320: Based on the data collection time of the indicator, for each z ij Assign a time decay factor *at*: assign *at*=k1 to data within 1 year, *at*=k2 to data within 1-3 years, and *at*=k3 to data older than 3 years, where *k1*>*k2*>*k3*; calculate the time weighting matrix *Zt*, where *Zt(i,j)*=z ij *at(i,j); where at(i,j) represents the time decay factor of the j-th indicator data corresponding to the i-th grayscale rule; using the formula: calculate the final weighting matrix Zts, Zts(i,j)=Zt(i,j)*p s p s This indicates the scenario adaptation factor based on the monitoring scenario settings; all rules under the same monitoring scenario share the same p. s Ensure consistent evaluation standards within the scenario;

[0033] Step S330: Calculate the weight w of the j-th indicator. j w j =(1-e j ) / ∑(1-e j ), and satisfy ∑w j =1; e j The information entropy of the j-th indicator is represented; the final weighted matrix Zts is then compared with the indicator weights w. j Multiplying them together yields the weighted standardized matrix V; V(i,j) = Zts(i,j) * w j ;

[0034] Step S340: Determine the ideal state with the greatest impact on integrated business as the positive ideal solution V. + V + =(max(V(1,j),V(2,j),...,V(m,j))|j=1,2,...,n); The ideal state with the minimum impact on integrated business is determined as the negative ideal solution V. - V - =(min(V(1,j),V(2,j),...,V(m,j))|j=1,2,...,n); Calculate the distance Di between the i-th rule and the positive ideal solution. + ={∑[V(1,j)-V + (j)] 2} 1 / 2The distance Di between the i-th rule and the negative ideal solution - ={∑[V(1,j)-V - (j)] 2} 1 / 2 Then calculate the comprehensive business impact Yi of the i-th gray-scale rule, Yi = Di - / (Di + +Di - The larger the value, the greater the business impact of the application of the corresponding group of electromechanical equipment, and the more important the rule is.

[0035] Furthermore, step S400 includes the following:

[0036] All rules for the same group of devices to be analyzed are sorted in descending order of Yi, and the rule corresponding to the maximum value is taken as the key matching rule under the corresponding attribute combination.

[0037] When the intelligent monitoring system receives a new gray-scale release request, it extracts the combination of device attributes and target device group information from the request.

[0038] Search the list of candidate rules for key matching rules and send alerts accordingly;

[0039] After the new key matching rule is applied, the gray-scale release execution process of the target electromechanical equipment group is continuously monitored, and the running data of all application events is recorded. Based on the running data of application events, the necessary optimization index corresponding to each application event record is recalculated. If the necessary optimization index of all application event records is greater than or equal to the corresponding threshold, the secondary verification is deemed normal, and the new rule can continue to be executed. If the necessary optimization index of any application event record is less than the threshold, the secondary verification is deemed abnormal. The complete data of the abnormal event is used as the data basis, and the process returns to step S300 to re-analyze the comprehensive business impact of all release rules corresponding to the target electromechanical equipment group and select the optimal key matching rule again.

[0040] An intelligent monitoring system for electromechanical equipment based on artificial intelligence, comprising an attribute combination set generation module, a key matching rule output module, a necessary optimization index analysis module, a comprehensive business impact analysis module, a rule update module, and a secondary verification response module;

[0041] The attribute combination set generation module is used to traverse historical data, extract attribute combinations formed by electromechanical equipment attributes under different monitoring scenarios, and generate equipment attribute combination sets.

[0042] The key matching rule output module is used to extract key matching rules from the monitoring event records of electromechanical equipment;

[0043] The Necessary Optimization Index Analysis module is used to calculate the necessary optimization index for the matching and release rules of monitoring events for all electromechanical equipment groups with the same target.

[0044] The comprehensive business impact analysis module is used to analyze the comprehensive business impact of the group of devices under analysis, including the grayscale rules of the device attributes.

[0045] The rule update module is used to update the key matching rules for the group of devices to be analyzed under the corresponding attribute combinations;

[0046] The secondary verification response module is used to provide early warning recommendations when responding to the candidate rule column in real time; and to perform secondary verification on the target group of electromechanical equipment for changing the key matching rules.

[0047] Furthermore, the necessary optimization index analysis module includes a parameter calculation unit and a necessary optimization index calculation unit;

[0048] The parameter calculation unit is used to calculate the business continuity assessment value, segmentation accuracy assessment value, gray target achievement assessment value, and system performance stability assessment value of the effective monitoring event records, taking the effective monitoring events as the analysis object.

[0049] The necessary optimization index calculation unit is used to calculate the comprehensive adaptation index based on the output parameters of the parameter calculation unit, and to define the necessary optimization index for effective monitoring events corresponding to the same target electromechanical equipment group based on the comprehensive adaptation index.

[0050] Furthermore, the comprehensive business impact analysis module includes a standardized matrix construction unit, a weighted matrix calculation unit, a weighted standardized matrix calculation unit, and a comprehensive business impact calculation unit;

[0051] The standardized matrix construction unit is used to extract the number of gray-scale rules contained in the group of devices to be analyzed, as well as the evaluation index of each gray-scale rule record, and to standardize the data to generate a standardized matrix.

[0052] The weighted matrix calculation unit is used to assign a time decay factor and a scene adaptation factor to each data point in the matrix, ultimately obtaining the weighted matrix.

[0053] The weighted standardization matrix calculation unit is used to generate a weighted standardization matrix based on the weights of the indicators;

[0054] The comprehensive business impact calculation unit is used to calculate the comprehensive business impact based on the distance to the positive ideal solution and the negative ideal solution.

[0055] Furthermore, the secondary verification response module includes a data extraction unit and a necessary optimization index secondary verification unit;

[0056] The data extraction unit is used to continuously monitor the gray-scale release execution process of the target electromechanical equipment group after the application of new key matching rules, and record the running data of all application events;

[0057] The necessary optimization index secondary verification unit is used to recalculate the necessary optimization index corresponding to each application event record; and to verify whether the update was successful based on the index result.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] 1. This invention constructs a multi-dimensional device attribute system, combines it with differentiated association strength calculation, strengthens the association of core attributes in the monitoring scenario, filters out invalid and redundant combinations, generates a highly targeted set of device attribute combinations, provides accurate data support for rule configuration, and improves the accuracy of rule matching.

[0060] 2. Based on historical and effective monitoring events, this invention quantifies the rule comprehensive adaptation index and necessary optimization index through a four-dimensional evaluation system, and combines it with comprehensive business impact analysis to achieve objective judgment of multiple rule conflicts and optimal rule selection, avoiding reliance on human experience and improving the scientificity and rationality of rule selection.

[0061] 3. This invention introduces a time decay factor and a scenario adaptation factor to optimize the comprehensive business impact assessment, making the rule adaptation more in line with the actual business situation; through secondary verification, a closed-loop optimization mechanism is formed to correct rule adaptation deviations in a timely manner, effectively ensuring business continuity, system performance stability and the achievement of gray-scale targets, and reducing operation and maintenance risks.

[0062] 4. This invention automates the entire process of attribute combination extraction, rule filtering, optimization and updating, and verification response, reducing manual intervention and improving the efficiency of electromechanical equipment supervision. The system has good scalability and adaptability, which can meet the supervision needs of different monitoring scenarios and different types of equipment, and provide an intelligent traffic segmentation solution for the gray release of electromechanical equipment. Attached Figure Description

[0063] Figure 1 This is a flowchart illustrating an intelligent monitoring method for electromechanical equipment based on artificial intelligence, according to the present invention. Detailed Implementation

[0064] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Example: Figure 1 As shown, this invention provides an intelligent monitoring method for electromechanical equipment based on artificial intelligence, the method comprising:

[0066] Step S100: Traverse historical data, extract attribute combinations of electromechanical equipment under different monitoring scenarios, and generate equipment attribute combination set; associate and mark the target electromechanical equipment groups monitored by each attribute combination in the equipment attribute combination set;

[0067] Step S200: Construct a list of candidate rules from the grayscale release rules of the independent attributes of each attribute combination in the device attribute combination set, and extract the key matching rules of the electromechanical equipment monitoring event records; the electromechanical equipment monitoring event refers to the grayscale release application event corresponding to the same target electromechanical equipment group containing the same grayscale release rule; calculate the necessary optimization index of the matching release rules for all electromechanical equipment monitoring events corresponding to the same target electromechanical equipment group;

[0068] Step S300: Set the necessary optimization index threshold, and mark the target electromechanical equipment group whose necessary optimization index is greater than or equal to the necessary optimization index threshold as the equipment group to be analyzed; analyze the comprehensive business impact of the gray scale rules of the equipment attributes of the equipment group to be analyzed;

[0069] Step S400: Based on the comprehensive business impact verification, update the key matching rules of the equipment group to be analyzed under the corresponding attribute combination, and provide early warning recommendations when responding to the candidate rule column in real time; perform secondary verification on the target electromechanical equipment group with the changed key matching rules.

[0070] Step S100 includes the following processes:

[0071] Step S110: Construct an attribute dimension system for electromechanical equipment. The attribute dimension system includes physical attribute dimensions, operational attribute dimensions, business association dimensions, and environmental attribute dimensions. The physical attribute dimension includes equipment manufacturer, model, deployment time, and hardware configuration parameters. The operational attribute dimension includes real-time load rate, number of failures in the past 30 days, average mean time between failures (MTBF), and firmware version. The business association dimension includes the road segment, associated business type, and importance level. The associated business type includes core toll collection and non-core service areas, and the importance level includes Level 1, Level 2, and Level 3. The environmental attribute dimension includes installation location, average ambient temperature, and vibration frequency.

[0072] Step S120: Traverse the historical monitoring data of electromechanical equipment and extract the attribute combinations under different monitoring scenarios. The monitoring scenarios include UPS fault prediction scenario, optical cable interruption monitoring scenario, information board display abnormality scenario and industrial control computer operation status monitoring scenario.

[0073] Step S130: During the attribute combination extraction process, perform differentiated calculations on the attribute association strength under different scenarios; the differentiated calculations include the following:

[0074] The basic correlation degree R(A,B) of any two device attributes in the attribute combination co-occurring in historical data is calculated as follows: R(A,B)=P(A∩B) / P(A)*P(B), where P(A) represents the probability of device attribute A occurring, P(B) represents the probability of device attribute B occurring, and P(A∩B) represents the probability of device attribute A and device attribute B occurring in the same attribute combination.

[0075] Assign scene weight S to all monitoring scenarios for historical data storage; and distinguish between core attributes and non-core attributes for the device attributes contained in each monitoring scenario, assigning weight W1 to core attributes and weight W2 to non-core attributes, where W1>W2;

[0076] Then calculate the attribute differential correlation strength S for each monitoring scenario. AB S AB =R(A,B)*S*W, W={W1,W2};

[0077] When the number of independent device attributes contained in an attribute combination is greater than 2, the differential association strength of the corresponding attribute combination is calculated by the geometric mean of the pairwise attribute association strengths.

[0078] Step S140: When S AB When S is greater than or equal to the minimum association strength threshold of the scene, the corresponding attribute combination is valid. AB When the minimum association strength threshold of the scene is reached, the corresponding attribute combination is deleted; all valid attribute combinations are retained, and attribute combinations consisting of different independent attributes are classified into one category to generate a set of device attribute combinations; the set of device attribute combinations is stored in the form of a triple of "scene ID-attribute combination code-associated device feature"; and a mapping relationship with the target electromechanical equipment group is established to complete the association marking.

[0079] Step S200 includes the following specific steps:

[0080] Step S210: Establish a basic template library for grayscale release rules. The template library contains basic rule templates corresponding to different attribute dimensions. For each attribute combination in the device attribute combination set, extract the independent attributes contained in each attribute combination, and match the basic rule templates corresponding to each independent attribute from the basic rule template library. Arrange all independent basic rule templates of the same attribute combination to generate a list of candidate rules.

[0081] Step S220: Filter events from the electromechanical equipment monitoring event records that meet the same target electromechanical equipment group, the same gray-scale release rules, and a gray-scale execution success rate greater than or equal to the success rate threshold. Define these as the same type of valid monitoring events. Gray-scale execution success rate = (number of fault-free operating devices during the gray-scale period + number of fault-repaired devices) / total number of gray-scale devices recorded for the target electromechanical equipment group * 100%; extract the basic rule templates applied in the same type of valid monitoring events as key matching rules.

[0082] Step S230: Taking the effective monitoring events as the analysis object, extract the business continuity assessment value C1, the segmentation accuracy assessment value C2, the gray target achievement assessment value C3, and the system performance stability assessment value C4 of the effective monitoring event records;

[0083] The business continuity assessment value C1 is calculated as follows: C1 = 1 - (core business interruption duration / total grayscale duration) * 0.5 - (number of false alarms / total number of alarms) * 0.3 - (business recovery delay duration / standard recovery duration) * 0.2; the value ranges from [0,1], and the larger the value, the better the business continuity.

[0084] The segmentation accuracy evaluation value C2 = rule matching accuracy * 0.4 + target device coverage * 0.3 + reciprocal of non-target device missegmentation rate * 0.3; where rule matching accuracy = number of target devices that meet the rule / total number of devices that match the rule * 100%; target device coverage = number of target devices that match the rule / total number of target devices * 100%; non-target device missegmentation rate = number of non-target devices that are mismatched by the rule / total number of devices that match the rule * 100%; the value of C2 ranges from [0,1], and the larger the value, the better the segmentation accuracy.

[0085] The gray-scale target achievement evaluation value C3 = Adaptability verification compliance rate × 0.5 + Risk control compliance rate × 0.25 + Gray-scale cycle compliance rate × 0.25; where, Adaptability verification compliance rate = Number of new version compatible devices in gray-scale devices / Total number of gray-scale devices × 100%, Risk control compliance rate = Number of times the gray-scale ratio of core devices is less than or equal to the preset threshold / Total number of executions × 100%, Gray-scale cycle compliance rate = Actual gray-scale verification time / Planned gray-scale cycle × 100%; C3 ranges from [0,1], and the larger the value, the better the target achievement.

[0086] The system performance stability evaluation value C4 is calculated as follows: C4 = 1 - (duration of exceeding resource utilization limit / total gray-scale duration) × 0.4 - (number of times routing response delay exceeds limit / total number of requests) × 0.3 - (number of system errors / total number of executions) × 0.3; the value range is [0,1]. The larger the value, the more stable the system performance.

[0087] Step S240: Calculate the rule comprehensive fit index A, A=C1*0.38+C2*0.32+C3*0.18+C4*0.12; the larger A is, the better the rule fit; define the necessary optimization index Q for the effective monitoring events corresponding to the same target electromechanical equipment group, Q=1-A0, where A0 represents the average comprehensive fit index of all effective monitoring events recorded for the same target electromechanical equipment group.

[0088] Analyzing the overall business impact of the group of devices to be analyzed, including the grayscale rules of the device attributes, includes the following steps:

[0089] Step S310: Extract the number m of gray-scale rules contained in the group of devices to be analyzed. Each gray-scale rule records n evaluation indicators, such as specific indicators under dimensions like business value impact and fault propagation impact; construct the initial indicator matrix X.

[0090]

[0091] Where x ij This represents the original data of the i-th grayscale rule on the j-th metric; x ij ={x 11 ,x 12 ,...,x mn};for x ij Standardization yields the standardized matrix Z, z ij This represents the data of the i-th grayscale rule after standardization on the j-th metric;

[0092]

[0093] The standardization process begins with analyzing the indicators, categorizing them into positive, negative, and Boolean indicators. A larger positive indicator value indicates a greater impact. Standardization is represented by z. ij =[x ij -min(x j )] / [max(x j )-min(x j The smaller the value of the negative indicator, the greater its impact; it is standardized as: z ij =[max(x j )-x ij ] / [max(x j )-min(x j Boolean indexes retain the original value as z based on the principle of 1 for yes and 0 for no. ij ;

[0094] Step S320: Based on the data collection time of the indicator, for each z ijAssign a time decay factor *at*: assign *at*=k1 to data within 1 year, *at*=k2 to data within 1-3 years, and *at*=k3 to data older than 3 years, where *k1*>*k2*>*k3*; calculate the time weighting matrix *Zt*, where *Zt(i,j)*=z ij *at(i,j); where at(i,j) represents the time decay factor of the j-th indicator data corresponding to the i-th grayscale rule; using the formula: calculate the final weighting matrix Zts, Zts(i,j)=Zt(i,j)*p s p s This indicates the scenario adaptation factor based on the monitoring scenario settings; all rules under the same monitoring scenario share the same p. s Ensure consistent evaluation standards across different scenarios; such as the billing equipment and UPS power supply settings in core business scenarios. s =1.2; Non-core business scenarios such as service area lighting settings p s =0.8;

[0095] Step S330: Calculate the weight w of the j-th indicator. j w j =(1-e j ) / ∑(1-e j ), and satisfy ∑w j =1; e j The information entropy of the j-th indicator is represented; the final weighted matrix Zts is then compared with the indicator weights w. j Multiplying them together yields the weighted standardized matrix V; V(i,j) = Zts(i,j) * w j ;

[0096]

[0097] If Zts(i,j)=0, then lnZts(i,j)=0;

[0098] Step S340: Determine the ideal state with the greatest impact on integrated business as the positive ideal solution V. + V + =(max(V(1,j),V(2,j),...,V(m,j))|j=1,2,...,n); The ideal state with the minimum impact on integrated business is determined as the negative ideal solution V. - V - =(min(V(1,j),V(2,j),...,V(m,j))|j=1,2,...,n); Calculate the distance Di between the i-th rule and the positive ideal solution. + ={∑[V(1,j)-V + (j)] 2} 1 / 2 The distance Di between the i-th rule and the negative ideal solution -={∑[V(1,j)-V - (j)] 2} 1 / 2 Then calculate the comprehensive business impact Yi of the i-th gray-scale rule, Yi = Di - / (Di + +Di - The larger the value, the greater the business impact of the application of the corresponding group of electromechanical equipment, and the more important the rule is.

[0099] Step S400 includes the following:

[0100] All rules for the same group of devices to be analyzed are sorted in descending order of Yi, and the rule corresponding to the maximum value is taken as the key matching rule under the corresponding attribute combination.

[0101] When the intelligent monitoring system receives a new gray-scale release request, it extracts the combination of device attributes and target device group information from the request.

[0102] Search the list of candidate rules for key matching rules and send alerts accordingly;

[0103] After the new key matching rule is applied, the gray-scale release execution process of the target electromechanical equipment group is continuously monitored, and the running data of all application events is recorded. Based on the running data of application events, the necessary optimization index corresponding to each application event record is recalculated. If the necessary optimization index of all application event records is greater than or equal to the corresponding threshold, the secondary verification is deemed normal, and the new rule can continue to be executed. If the necessary optimization index of any application event record is less than the threshold, the secondary verification is deemed abnormal. The complete data of the abnormal event is used as the data basis, and the process returns to step S300 to re-analyze the comprehensive business impact of all release rules corresponding to the target electromechanical equipment group and select the optimal key matching rule again.

[0104] An intelligent monitoring system for electromechanical equipment based on artificial intelligence, comprising an attribute combination set generation module, a key matching rule output module, a necessary optimization index analysis module, a comprehensive business impact analysis module, a rule update module, and a secondary verification response module;

[0105] The attribute combination set generation module is used to traverse historical data, extract attribute combinations formed by electromechanical equipment attributes under different monitoring scenarios, and generate equipment attribute combination sets.

[0106] The key matching rule output module is used to extract key matching rules from the monitoring event records of electromechanical equipment;

[0107] The Necessary Optimization Index Analysis module is used to calculate the necessary optimization index for the matching and release rules of monitoring events for all electromechanical equipment groups with the same target.

[0108] The comprehensive business impact analysis module is used to analyze the comprehensive business impact of the group of devices under analysis, including the grayscale rules of the device attributes.

[0109] The rule update module is used to update the key matching rules for the group of devices to be analyzed under the corresponding attribute combinations;

[0110] The secondary verification response module is used to provide early warning recommendations when responding to the candidate rule column in real time; and to perform secondary verification on the target group of electromechanical equipment for changing the key matching rules.

[0111] The necessary optimization index analysis module includes a parameter calculation unit and a necessary optimization index calculation unit;

[0112] The parameter calculation unit is used to calculate the business continuity assessment value, segmentation accuracy assessment value, gray target achievement assessment value, and system performance stability assessment value of the effective monitoring event records, taking the effective monitoring events as the analysis object.

[0113] The necessary optimization index calculation unit is used to calculate the comprehensive adaptation index based on the output parameters of the parameter calculation unit, and to define the necessary optimization index for effective monitoring events corresponding to the same target electromechanical equipment group based on the comprehensive adaptation index.

[0114] The comprehensive business impact analysis module includes a standardized matrix construction unit, a weighted matrix calculation unit, a weighted standardized matrix calculation unit, and a comprehensive business impact calculation unit.

[0115] The standardized matrix construction unit is used to extract the number of gray-scale rules contained in the group of devices to be analyzed, as well as the evaluation index of each gray-scale rule record, and to standardize the data to generate a standardized matrix.

[0116] The weighted matrix calculation unit is used to assign a time decay factor and a scene adaptation factor to each data point in the matrix, ultimately obtaining the weighted matrix.

[0117] The weighted standardization matrix calculation unit is used to generate a weighted standardization matrix based on the weights of the indicators;

[0118] The comprehensive business impact calculation unit is used to calculate the comprehensive business impact based on the distance to the positive ideal solution and the negative ideal solution.

[0119] The secondary verification response module includes a data extraction unit and a secondary verification unit for necessary optimization indices;

[0120] The data extraction unit is used to continuously monitor the gray-scale release execution process of the target electromechanical equipment group after the application of new key matching rules, and record the running data of all application events;

[0121] The necessary optimization index secondary verification unit is used to recalculate the necessary optimization index corresponding to each application event record; and to verify whether the update was successful based on the index result.

[0122] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent monitoring of electromechanical equipment based on artificial intelligence, characterized in that: The method includes: Step S100: Traverse historical data, extract attribute combinations of electromechanical equipment under different monitoring scenarios, and generate equipment attribute combination set; associate and mark the target electromechanical equipment groups monitored by each attribute combination in the equipment attribute combination set; Step S110: Construct an attribute dimension system for electromechanical equipment, which includes physical attribute dimensions, operational attribute dimensions, business association dimensions, and environmental attribute dimensions; Step S120: Traverse the historical monitoring data of electromechanical equipment and extract the attribute combinations under different monitoring scenarios. The monitoring scenarios include UPS fault prediction scenario, optical cable interruption monitoring scenario, information board display abnormality scenario and industrial control computer operation status monitoring scenario. Step S130: In the process of attribute combination extraction, perform differentiated calculations on the attribute association strength under different scenarios; Step S140: When the attribute differentiation association strength is greater than or equal to the minimum association strength threshold of the scene, the corresponding attribute combination is output as valid; when the attribute differentiation association strength is less than the minimum association strength threshold of the scene, the corresponding attribute combination is deleted. All valid attribute combinations are retained, and attribute combinations consisting of different independent attributes are classified into one category to generate a device attribute combination set. The device attribute combination set is stored in the form of a triple of "scene ID-attribute combination code-associated device feature". A mapping relationship with the target electromechanical equipment group is established to complete the association marking. Step S200: Construct a list of candidate rules from the grayscale release rules of the independent attributes of each attribute combination in the device attribute combination set, and extract the key matching rules of the electromechanical equipment monitoring event records; the electromechanical equipment monitoring event refers to the grayscale release application event corresponding to the same target electromechanical equipment group containing the same grayscale release rule; calculate the necessary optimization index of the matching release rules for all electromechanical equipment monitoring events corresponding to the same target electromechanical equipment group; Step S200 includes the following specific steps: Step S210: Establish a basic template library for grayscale release rules, which contains basic rule templates corresponding to different attribute dimensions; for each attribute combination in the device attribute combination set, extract the independent attributes contained in each attribute combination, and match the basic rule templates corresponding to each independent attribute from the basic rule template library; arrange all independent basic rule templates of the same attribute combination to generate a list of candidate rules; Step S220: Filter events from the electromechanical equipment monitoring event records that meet the same target electromechanical equipment group, the same gray-scale release rules, and a gray-scale execution success rate greater than or equal to the success rate threshold. Define these as the same type of valid monitoring events. Gray-scale execution success rate = (number of fault-free operating devices during the gray-scale period + number of fault-repaired devices) / total number of gray-scale devices recorded for the target electromechanical equipment group * 100%; extract the basic rule templates applied in the same type of valid monitoring events as key matching rules. Step S230: Taking the effective monitoring events as the analysis object, extract the business continuity assessment value C1, the segmentation accuracy assessment value C2, the gray target achievement assessment value C3, and the system performance stability assessment value C4 of the effective monitoring event records; The business continuity assessment value C1 is calculated as follows: C1 = 1 - (core business interruption duration / total grayscale duration) * 0.5 - (number of false alarms / total number of alarms) * 0.3 - (business recovery delay duration / standard recovery duration) * 0.2; The segmentation accuracy evaluation value C2 = rule matching accuracy * 0.4 + target device coverage * 0.3 + reciprocal of non-target device missegmentation rate * 0.3; where rule matching accuracy = number of target devices that match the rule / total number of devices matching the rule * 100%; target device coverage = number of target devices that match the rule / total number of target devices * 100%; non-target device missegmentation rate = number of non-target devices that mismatch the rule / total number of devices matching the rule * 100%. The gray-scale target achievement evaluation value C3 = Adaptability verification compliance rate × 0.5 + Risk control compliance rate × 0.25 + Gray-scale cycle compliance rate × 0.25; where, Adaptability verification compliance rate = Number of gray-scale devices that are compatible with the new version / Total number of gray-scale devices × 100%, Risk control compliance rate = Number of times the gray-scale ratio of core devices is less than or equal to the preset threshold / Total number of executions × 100%, Gray-scale cycle compliance rate = Actual time to complete gray-scale verification / Planned gray-scale cycle × 100%; The system performance stability assessment value C4 = 1 - (duration of exceeding resource utilization limit / total gray-scale duration) × 0.4 - (number of times routing response latency exceeds limit / total number of requests) × 0.3 - (number of system errors / total number of executions) × 0.3; Step S240: Calculate the rule comprehensive adaptation index A, A=C1*0.38+C2*0.32+C3*0.18+C4*0.12; Define the necessary optimization index Q for the effective monitoring events corresponding to the same target electromechanical equipment group, Q=1-A0, where A0 represents the average comprehensive adaptation index of all effective monitoring events recorded for the same target electromechanical equipment group; Step S300: Set the necessary optimization index threshold, and mark the target electromechanical equipment group whose necessary optimization index is greater than or equal to the necessary optimization index threshold as the equipment group to be analyzed; analyze the comprehensive business impact of the gray scale rules of the equipment attributes of the equipment group to be analyzed; Step S400: Based on the comprehensive business impact verification, update the key matching rules of the equipment group to be analyzed under the corresponding attribute combination, and provide early warning recommendations when responding to the candidate rule column in real time; perform secondary verification on the target electromechanical equipment group with the changed key matching rules.

2. The intelligent monitoring method for electromechanical equipment based on artificial intelligence according to claim 1, characterized in that: Step S130 includes the following process: The differential calculation includes the following: The basic correlation degree R(A,B) of any two device attributes in the attribute combination co-occurring in historical data is calculated as follows: R(A,B)=P(A∩B) / P(A)*P(B), where P(A) represents the probability of device attribute A occurring, P(B) represents the probability of device attribute B occurring, and P(A∩B) represents the probability of device attribute A and device attribute B occurring in the same attribute combination. Assign a scene weight S to all monitoring scenarios for storing historical data; Furthermore, for each monitoring scenario, the device attributes are distinguished into core attributes and non-core attributes. Core attributes are assigned a weight W1, and non-core attributes are assigned a weight W2, where W1 > W2. S = R(A, B) * S * W, W = {W1, W2} AB AB S = R(A, B) * S * W, W = {W1, W2}​ When the number of independent device attributes contained in an attribute combination is greater than 2, the differential association strength of the corresponding attribute combination is calculated by the geometric mean of the pairwise attribute association strength.

3. The intelligent monitoring method for electromechanical equipment based on artificial intelligence according to claim 1, characterized in that: The analysis of the overall business impact of the group of devices to be analyzed, including the grayscale rules of the device attributes, includes the following steps: Step S310: Extract the number m of gray-scale rules contained in the group of devices to be analyzed. Each gray-scale rule is measured by n evaluation indicators, and an initial indicator matrix X is constructed. wherein x ij represents the original data of the i-th gray scale rule on the j-th index; x ij 11 12 mn ; the standardized matrix Z is obtained by standardizing x ij ; z ij represents the data of the i-th gray scale rule standardized on the j-th index;​​​ Step S320: Based on the data collection time of the indicator, for each z ij Assign a time decay factor *at*: assign *at*=k1 to data within 1 year, *at*=k2 to data within 1-3 years, and *at*=k3 to data older than 3 years, where *k1*>*k2*>*k3*; calculate the time weighting matrix *Zt*, where *Zt(i,j)*=z ij *at(i,j); where at(i,j) represents the time decay factor of the j-th indicator data corresponding to the i-th grayscale rule; using the formula: calculate the final weighting matrix Zts, Zts(i,j)=Zt(i,j)*p s p s This indicates the scene adaptation factor based on the monitoring scene settings; Step S330: Calculate the weight w of the j-th indicator. j w j =(1-e j ) / ∑(1-e j ), and satisfy ∑w j =1; e j The information entropy of the j-th indicator is represented; the final weighted matrix Zts is then compared with the indicator weights w. j Multiplying them together yields the weighted standardized matrix V; V(i,j) = Zts(i,j) * w j ; Step S340: Determine the ideal state with the greatest impact on integrated business as the positive ideal solution V. + V + =(max(V(1,j),V(2,j),...,V(m,j))|j=1,2,...,n); The ideal state with the minimum impact on integrated business is determined as the negative ideal solution V. - V - =(min(V(1,j),V(2,j),...,V(m,j))|j=1,2,...,n); Calculate the distance Di between the i-th rule and the positive ideal solution. + ={∑[V(1,j)-V + (j)] 2 } 1 / 2 The distance Di between the i-th rule and the negative ideal solution - ={∑[V(1,j)-V - (j)] 2 } 1 / 2 Then calculate the comprehensive business impact Yi of the i-th gray-scale rule, Yi = Di - / (Di + +Di - ).

4. The intelligent monitoring method for electromechanical equipment based on artificial intelligence according to claim 3, characterized in that: Step S400 includes the following: All rules for the same group of devices to be analyzed are sorted in descending order of Yi, and the rule corresponding to the maximum value is taken as the key matching rule under the corresponding attribute combination. When the intelligent monitoring system receives a new gray-scale release request, it extracts the combination of device attributes and target device group information from the request. Search the list of candidate rules for key matching rules and send alerts accordingly; After the new key matching rules are applied, the canary release execution process of the target electromechanical equipment group is continuously monitored, and the running data of all application events is recorded; based on the running data of application events, the necessary optimization index corresponding to each application event record is recalculated. If the necessary optimization index of all application event records is greater than or equal to the corresponding threshold, the secondary verification is deemed normal and the new rule can continue to be executed; if the necessary optimization index of any application event record is less than the threshold, the secondary verification is deemed abnormal. The complete data of the abnormal event is used as the data basis, and the process returns to step S300 to re-analyze the comprehensive business impact of all published rules corresponding to the target electromechanical equipment group and to select the optimal key matching rule again.

5. An intelligent monitoring system for electromechanical equipment based on artificial intelligence, using the intelligent monitoring method for electromechanical equipment based on artificial intelligence according to any one of claims 1-4, characterized in that: The system includes an attribute combination set generation module, a key matching rule output module, a necessary optimization index analysis module, a comprehensive business impact analysis module, a rule update module, and a secondary verification response module. The attribute combination set generation module is used to traverse historical data, extract attribute combinations formed by electromechanical equipment attributes under different monitoring scenarios, and generate equipment attribute combination sets. The key matching rule output module is used to extract key matching rules from the monitoring event records of electromechanical equipment; The necessary optimization index analysis module is used to calculate the necessary optimization index of the matching and release rules for all electromechanical equipment monitoring events corresponding to the same target electromechanical equipment group; The comprehensive business impact analysis module is used to analyze the comprehensive business impact of the group of devices to be analyzed, including the grayscale rules of the device attributes. The rule update module is used to update the key matching rules for the group of devices to be analyzed under the corresponding attribute combinations; The secondary verification response module is used to provide early warning recommendations when responding to the candidate rule column in real time; A secondary verification is performed on the target group of electromechanical equipment for which the key matching rules have been changed.

6. The intelligent monitoring system for electromechanical equipment based on artificial intelligence according to claim 5, characterized in that: The necessary optimization index analysis module includes a parameter calculation unit and a necessary optimization index calculation unit. The parameter calculation unit is used to calculate the business continuity assessment value, segmentation accuracy assessment value, grayscale target achievement assessment value, and system performance stability assessment value of the effective monitoring event records, taking the effective monitoring events as the analysis object. The necessary optimization index calculation unit is used to calculate the comprehensive adaptation index based on the output parameters of the parameter calculation unit, and to define the necessary optimization index for effective monitoring events corresponding to the same target electromechanical equipment group using the comprehensive adaptation index.

7. The intelligent monitoring system for electromechanical equipment based on artificial intelligence according to claim 6, characterized in that: The comprehensive business impact analysis module includes a standardized matrix construction unit, a weighted matrix calculation unit, a weighted standardized matrix calculation unit, and a comprehensive business impact calculation unit. The standardized matrix construction unit is used to extract the number of gray-scale rules contained in the group of devices to be analyzed, as well as the evaluation index of each gray-scale rule record, and to standardize the data to generate a standardized matrix. The weighted matrix calculation unit is used to assign a time decay factor and a scene adaptation factor to each data in the matrix to finally obtain the weighted matrix. The weighted standardization matrix calculation unit is used to generate a weighted standardization matrix based on the weights of the indicators; The comprehensive business impact calculation unit is used to calculate the comprehensive business impact based on the distance to the positive ideal solution and the negative ideal solution.

8. The intelligent monitoring system for electromechanical equipment based on artificial intelligence according to claim 7, characterized in that: The secondary verification response module includes a data extraction unit and a necessary optimization index secondary verification unit. The data extraction unit is used to continuously monitor the gray-scale release execution process of the target electromechanical equipment group after the application of new key matching rules, and record the running data of all application events; The necessary optimization index secondary verification unit is used to recalculate the necessary optimization index corresponding to each application event record. Verify whether the update was successful based on the index results.

Images