A method and device for early warning of highway electromechanical equipment abnormalities

Abstract

The application is suitable for the field of intelligent traffic technology, and provides a highway electromechanical equipment abnormal early warning method and device.In the embodiment, for the electromechanical equipment on the highway, the balance detection rate and the false alarm rate are first determined through the historical operation data of the electromechanical equipment or the same type electromechanical equipment, then the false alarm rate threshold value determined based on the road section characteristic parameter of the electromechanical equipment is combined to determine the abnormal prediction threshold value of the electromechanical equipment, and then real-time monitoring is carried out during the operation of the electromechanical equipment, the real-time abnormal prediction value of the electromechanical equipment is determined according to the real-time data, and the device is warned when it is determined that the device meets the abnormal condition through comparison.The application can set different false alarm rate standards according to different road section characteristics of the equipment, thereby solving the problems of missing key faults and generating a large number of invalid alarms, resulting in insufficient accuracy of the alarm and waste of human resources.In addition, the threshold value of the maximum detection rate is set as the abnormal prediction threshold value, and the discovery ability of the real fault is maximized, thereby realizing the minimization of the safety risk on the premise of ensuring the controllable operation and maintenance cost.

Description

Technical Field

[0001] This application relates to the field of intelligent transportation technology, and in particular to a method and device for early warning of abnormalities in highway electromechanical equipment. Background Technology

[0002] With the continuous expansion of my country's expressway network and the ongoing improvement of its intelligent level, electromechanical equipment, as the core support system for expressway operation and management, is directly related to road network safety and traffic efficiency through its stable operation. However, current anomaly warnings for key electromechanical equipment such as ETC gantries, surveillance cameras, information boards, and weather stations mainly rely on manual inspections or simple rule judgments based on fixed thresholds. This is difficult to adapt to the differentiated characteristics of different road sections and equipment types, and generally suffers from high false alarm rates and high risks of missed alarms. This leads to a waste of operation and maintenance resources or untimely fault response, seriously hindering the improvement of the level of intelligent operation and maintenance management of expressways.

[0003] While some existing technologies attempt to use machine learning methods to detect anomalies in equipment operation data, most neglect the impact of road segment characteristics on early warning strategies. They often employ a uniform false alarm rate standard, leading to missed critical faults on important hub sections due to overly strict thresholds, and numerous invalid alarms on less important sections due to overly lenient thresholds. This results in insufficient alarm accuracy and wasted human resources. Furthermore, existing methods lack a systematic optimization strategy to maximize fault detection rate under acceptable false alarm rate constraints when optimizing thresholds, making it difficult to achieve the optimal balance between maintenance costs and safety assurance, and failing to meet the urgent needs of modern intelligent transportation platforms for accurate early warning of electromechanical equipment. Summary of the Invention

[0004] In view of this, this application provides a method and apparatus for early warning of abnormalities in highway electromechanical equipment, so as to improve the accuracy of equipment abnormality warning.

[0005] The first aspect of this application provides a method for early warning of abnormalities in highway electromechanical equipment, applied to an intelligent transportation platform, the method comprising: Historical operating data containing several operating parameters of the target electromechanical equipment is obtained, and then sample data of the target electromechanical equipment is determined through the historical operating data. The sample data includes the abnormal prediction values ​​and the true labels of the historical operating data. The road segment characteristic parameters of the target electromechanical equipment are determined, and the false alarm rate threshold of the target electromechanical equipment is determined through the road segment characteristic parameters. The balanced detection rate and false alarm rate of the historical operation data are determined based on the sample data. The road segment characteristic parameters include the basic information of the target electromechanical equipment, its layout area, and traffic flow. Then, by using the balanced detection rate, false alarm rate, and false alarm rate threshold, an anomaly prediction threshold that maximizes the detection rate of the historical operating data is determined, and the early warning conditions for the target electromechanical equipment are determined based on the anomaly prediction threshold. The system monitors the operating data of the target electromechanical equipment in real time, determines the real-time anomaly prediction value based on the operating data, and issues an anomaly warning when the real-time anomaly prediction value meets the warning conditions.

[0006] Optionally, determining the sample data of the target electromechanical equipment through the historical operating data includes: Feature extraction is performed on the historical operation data, and the extracted data is preprocessed to obtain the historical feature values ​​of the historical operation data; The historical feature values ​​are input into the trained fault prediction model to obtain the anomaly prediction values ​​of the historical operating data. The true label of the historical operating data is determined by analyzing the fault logs of the target electromechanical equipment. The abnormal predicted value and the true label are determined as the sample data of the target electromechanical equipment.

[0007] Optionally, determining the false alarm rate threshold of the target electromechanical equipment through the road segment feature parameters includes: The risk value of the target electromechanical equipment is determined by quantifying each parameter in the road segment feature parameters and then combining the preset weight values ​​of each parameter. The mapping relationship between the risk value and the false alarm rate threshold is preset based on the business rules, and then the corresponding false alarm rate threshold is determined through the risk value.

[0008] Optionally, determining the anomaly prediction threshold that maximizes the detection rate of the historical operational data by balancing the detection rate, the false alarm rate, and the false alarm rate threshold includes: Determine the maximum and minimum values ​​of the anomaly prediction threshold, Tmax and Tmin, and then use the formula... Determine the anomaly prediction threshold T that maximizes the detection rate, where, The maximum value of the balanced detection rate. The false alarm rate is M, and the false alarm rate threshold is M.

[0009] Optionally, the road segment characteristic parameters may also include natural environmental parameters that characterize meteorological conditions.

[0010] A second aspect of this application provides a highway electromechanical equipment anomaly early warning device, applied to an intelligent transportation platform, the device comprising: The sample data determination unit is used to acquire historical operating data containing several operating parameters of the target electromechanical equipment, and then determine the sample data of the target electromechanical equipment through the historical operating data, wherein the sample data includes the abnormal prediction value and the true label of the historical operating data; The data processing unit is used to determine the road segment characteristic parameters of the target electromechanical equipment, determine the false alarm rate threshold of the target electromechanical equipment through the road segment characteristic parameters, and determine the balanced detection rate and false alarm rate of the historical operation data based on the sample data. The road segment characteristic parameters include the basic information of the target electromechanical equipment, its layout area, and traffic flow. The early warning condition determination unit is used to determine the anomaly prediction threshold that maximizes the detection rate of the historical operating data by further using the balanced detection rate, false alarm rate and the false alarm rate threshold, and to determine the early warning condition of the target electromechanical equipment based on the anomaly prediction threshold. An anomaly warning unit is used to monitor the operating data of the target electromechanical equipment in real time, determine the real-time anomaly prediction value through the operating data, and issue an anomaly warning when the real-time anomaly prediction value meets the warning conditions.

[0011] Optionally, the sample data for determining the target electromechanical equipment through the historical operating data in the sample data determination unit includes: Feature extraction is performed on the historical operation data, and the extracted data is preprocessed to obtain the historical feature values ​​of the historical operation data; The historical feature values ​​are input into the trained fault prediction model to obtain the anomaly prediction values ​​of the historical operating data. The true label of the historical operating data is determined by analyzing the fault logs of the target electromechanical equipment. The abnormal predicted value and the true label are determined as the sample data of the target electromechanical equipment.

[0012] Optionally, the data processing unit's determination of the false alarm rate threshold for the target electromechanical equipment based on the road segment feature parameters includes: The risk value of the target electromechanical equipment is determined by quantifying each parameter in the road segment feature parameters and then combining the preset weight values ​​of each parameter. The mapping relationship between the risk value and the false alarm rate threshold is preset based on the business rules, and then the corresponding false alarm rate threshold is determined through the risk value.

[0013] Optionally, the anomaly prediction threshold for maximizing the detection rate of the historical operating data, determined by the early warning condition determination unit through the balanced detection rate, false alarm rate, and false alarm rate threshold, includes: Determine the maximum and minimum values ​​of the anomaly prediction threshold, Tmax and Tmin, and then use the formula... Determine the anomaly prediction threshold T that maximizes the detection rate, where, The maximum value of the balanced detection rate. The false alarm rate is M, and the false alarm rate threshold is M.

[0014] Optionally, the road segment characteristic parameters in the data processing unit may also include natural environmental parameters characterizing meteorological conditions.

[0015] In the embodiments provided in this application, for electromechanical equipment that needs to be monitored on highways, the balanced detection rate and false alarm rate are first determined using historical operating data of the electromechanical equipment or similar models. Then, an anomaly prediction threshold for the electromechanical equipment is determined by combining the false alarm rate threshold based on the road segment characteristic parameters of the electromechanical equipment. Real-time monitoring is then performed during operation, and the real-time anomaly prediction value of the electromechanical equipment is determined based on real-time data. An early warning is issued when the equipment meets the anomaly conditions by comparison. This application can set different false alarm rate standards according to different road segment characteristics of the equipment, thereby solving the problems of missed critical faults and the generation of a large number of invalid alarms, leading to insufficient alarm accuracy and wasted human resources, and improving the accuracy of early warnings. Furthermore, setting the threshold that maximizes the detection rate as the anomaly prediction threshold maximizes the ability to detect real faults, thereby minimizing safety risks while ensuring controllable operation and maintenance costs. Attached Figure Description

[0016] Figure 1 A flowchart illustrating the method provided in this application embodiment; Figure 2 This is a structural diagram of the device provided in the embodiments of this application; Figure 3 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0017] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0018] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0019] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0020] This application provides a method and apparatus for early warning of abnormalities in highway electromechanical equipment, so as to improve the accuracy of early warning.

[0021] The technical solutions of this application will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0022] like Figure 1 The diagram shown is a flowchart of a method for early warning of abnormalities in highway electromechanical equipment provided in this application. The process may include the following steps: Step S101: Obtain historical operating data containing several operating parameters of the target electromechanical equipment, and then determine the sample data of the target electromechanical equipment through the historical operating data.

[0023] In this embodiment, the historical operating data may include basic performance data such as CPU and memory usage, network latency, packet loss rate, etc. It may also include electrical data such as voltage, current, temperature at various locations, fan speed, etc. This historical operating data can be obtained from the previous operating data of the target electromechanical equipment. For historical operating data, such as for newly used electromechanical equipment, the historical operating data can also be obtained from other electromechanical equipment of the same model that has already been used; this application does not impose any limitations on this.

[0024] After determining the historical operating data, it is processed to obtain sample data for the target electromechanical equipment. In this embodiment, the sample data includes the predicted anomalies and true labels of the historical operating data. The specific processing procedure is as follows: 1. Feature extraction of historical operating data can be performed using a multi-dimensional feature extraction strategy. Statistical features such as sliding window mean, variance, trend, frequency domain features such as the main frequency energy after Fourier transform, and correlation features such as the correlation coefficient between equipment parameters can be extracted from the historical operating data. Then, the extracted data is preprocessed, i.e., standardized or normalized to eliminate dimensional differences, missing value imputation and outlier handling are used to ensure data quality, and feature selection techniques such as importance ranking based on variance or tree models are used to remove redundant features. Finally, the historical feature values ​​of the historical operating data are obtained.

[0025] 2. Input historical feature values ​​into the trained fault prediction model to obtain anomaly prediction values ​​for the historical operational data. This model can be trained based on the historical operational data or a comprehensive model trained using operational data from various other devices. By learning the complex mapping relationship between features and fault states, the model comprehensively evaluates the device state at each historical time point and outputs a continuous anomaly prediction value between 0 and 1. This value represents the confidence level or risk probability of a device failure at the corresponding time, thus transforming multi-dimensional feature information into a unified, quantified anomaly degree index, providing a direct basis for subsequent threshold optimization and early warning decisions. The specific training process of the model does not constitute a limitation of this application and will not be elaborated here.

[0026] 3. The fault status of each historical operational data item can be determined from fault reports or logs, and then its true label can be determined. For example, for operational data collected at time t, check whether a fault actually occurred within a future period, such as (t, t+24 hours). If a fault occurred, the true label of the sample is marked as a positive example; otherwise, it is marked as a negative example.

[0027] Using the methods described above, the abnormal predicted values ​​and true labels for each historical data point can be determined.

[0028] Step S102: Determine the road segment characteristic parameters of the target electromechanical equipment, determine the false alarm rate threshold of the target electromechanical equipment through the road segment characteristic parameters, and determine the balanced detection rate and false alarm rate of the historical operation data based on the sample data.

[0029] In this embodiment, the road segment characteristic parameters include the basic information of the target electromechanical equipment, the deployment area, and the traffic flow. The basic information may include factory parameters, performance indicators, etc., which can be determined from the equipment's product manual. The deployment area can be registered with the intelligent transportation platform when the target electromechanical equipment is deployed. The traffic flow can be determined by the intelligent transportation platform through monitoring and statistics of the road segment containing the target electromechanical equipment.

[0030] After determining the characteristic parameters of each road segment, the basic information of the equipment, the layout area, and the traffic flow are quantified, encoded, and weighted. Then, a mapping model is constructed based on expert experience rules or historical operation and maintenance cost-benefit analysis. For example, a strict false alarm rate threshold is set for traffic hubs and high traffic flow areas, such as M=0.03, while a lenient threshold is set for minor road segments, such as M=0.10. Finally, the automatic configuration of differentiated early warning strategies is realized.

[0031] In this embodiment, the process of determining the balanced detection rate and false alarm rate based on sample data is as follows: 1. A preset range for anomaly prediction thresholds is generally (0, 1). Based on the anomaly prediction values ​​and corresponding ground truth labels in the sample data, by iterating through all possible anomaly prediction thresholds within this range, the detection rate Recall(T) = TP / (TP+FN) and the false positive rate FPR(T) = FP / (FP+TN) are calculated for each candidate threshold T. Here, TP represents data where the anomaly prediction value is predicted as anomaly and is also detected as anomaly by the ground truth label; FN represents data where the anomaly prediction value is predicted as normal and is actually anomaly by the ground truth label; FP represents data where the anomaly prediction value is predicted as anomaly and is actually normal by the ground truth label; and FP represents data where the anomaly prediction value is predicted as anomaly and is also detected as normal by the ground truth label.

[0032] In another embodiment, the aforementioned road segment characteristic parameters also include natural environmental parameters characterizing meteorological conditions, such as temperature, humidity, wind speed, rainfall, visibility, and dust concentration. This embodiment can quantify, encode, and weight these parameters. Then, by combining the aforementioned basic information, layout area, and traffic flow parameters, a false alarm rate threshold is determined. This further improves the accuracy of the false alarm rate threshold.

[0033] Step S103: Determine the anomaly prediction threshold that maximizes the detection rate of the historical operating data by balancing the detection rate, the false alarm rate, and the false alarm rate threshold, and determine the early warning conditions of the target electromechanical equipment based on the anomaly prediction threshold.

[0034] In this embodiment, it can be achieved through the formula Construct a Recall-FPR curve, where M represents the false alarm rate threshold mentioned above. Then, locate the feasible region on the curve that satisfies the constraint "false alarm rate ≤ preset threshold". Within this feasible region, search for the maximum value point along the Recall axis; the threshold corresponding to this point is the desired anomaly prediction threshold T that maximizes the detection rate. In practice, all candidate thresholds can be iterated and filtered, retaining only those with an FPR value that does not exceed the limit. The threshold with the highest Recall value is then selected as the final solution. If multiple thresholds have the same Recall value, the solution with the smaller FPR or higher threshold is selected.

[0035] Once the anomaly prediction threshold T is determined, it can be directly used as the decision boundary. The warning condition is defined as "the real-time anomaly prediction value of the equipment is greater than or equal to the anomaly prediction threshold T". When this condition is met, the warning process is triggered immediately.

[0036] Step S104: Monitor the operating data of the target electromechanical equipment in real time, determine the real-time anomaly prediction value through the operating data, and issue an anomaly warning when the real-time anomaly prediction value meets the warning conditions.

[0037] In this embodiment, the intelligent transportation platform continuously collects real-time operating parameter streams of the target electromechanical equipment, performs feature engineering processing on each real-time data point consistent with historical data, and inputs the obtained real-time feature values ​​into the deployed fault prediction model to obtain real-time anomaly prediction values. Subsequently, the predicted value is compared with the aforementioned anomaly prediction threshold T. When the predicted value is greater than or equal to the anomaly prediction threshold T, an early warning process is immediately triggered.

[0038] This concludes the process. Figure 1 The process is shown below.

[0039] In this embodiment, for electromechanical equipment on highways, the balanced detection rate and false alarm rate are first determined using historical operating data of the equipment or similar models. Then, an anomaly prediction threshold for the equipment is determined by combining this with a false alarm rate threshold based on road segment characteristic parameters. Real-time monitoring is then performed during operation, and the real-time anomaly prediction value is determined based on real-time data. An early warning is issued when the equipment meets an anomaly condition through comparison. This application can set different false alarm rate standards based on different road segment characteristics of the equipment, thereby solving the problems of missed critical faults and the generation of numerous invalid alarms, leading to insufficient alarm accuracy and wasted human resources. Furthermore, the threshold maximizing the detection rate is set as the anomaly prediction threshold, maximizing the ability to detect real faults, thereby minimizing safety risks while ensuring controllable operation and maintenance costs.

[0040] This application also provides a highway electromechanical equipment abnormality early warning device, such as... Figure 2 As shown, the device includes: The sample data determination unit 201 is used to acquire historical operating data containing several operating parameters of the target electromechanical equipment, and then determine the sample data of the target electromechanical equipment through the historical operating data, wherein the sample data includes the abnormal prediction value and the true label of the historical operating data. Data processing unit 202 is used to determine the road segment characteristic parameters of the target electromechanical equipment, determine the false alarm rate threshold of the target electromechanical equipment through the road segment characteristic parameters, and determine the balanced detection rate and false alarm rate of the historical operation data based on the sample data. The road segment characteristic parameters include the basic information of the target electromechanical equipment, the layout area, and the traffic flow. The early warning condition determination unit 203 is used to determine the anomaly prediction threshold that maximizes the detection rate of the historical operating data by further using the balanced detection rate, false alarm rate and the false alarm rate threshold, and to determine the early warning condition of the target electromechanical equipment based on the anomaly prediction threshold. The anomaly warning unit 204 is used to monitor the operating data of the target electromechanical equipment in real time, determine the real-time anomaly prediction value through the operating data, and issue an anomaly warning when the real-time anomaly prediction value meets the warning conditions.

[0041] In another embodiment, the sample data for determining the target electromechanical equipment through the historical operating data in the sample data determining unit includes: Feature extraction is performed on the historical operation data, and the extracted data is preprocessed to obtain the historical feature values ​​of the historical operation data; The historical feature values ​​are input into the trained fault prediction model to obtain the anomaly prediction values ​​of the historical operating data. The true label of the historical operating data is determined by the fault log of the target electromechanical equipment; The abnormal predicted value and the true label are determined as the sample data of the target electromechanical equipment.

[0042] In another embodiment, the data processing unit's determination of the false alarm rate threshold for the target electromechanical equipment using the road segment feature parameters includes: The risk value of the target electromechanical equipment is determined by quantifying each parameter in the road segment feature parameters and then combining the preset weight values ​​of each parameter. The mapping relationship between the risk value and the false alarm rate threshold is preset based on the business rules, and then the corresponding false alarm rate threshold is determined through the risk value.

[0043] In another embodiment, the anomaly prediction threshold for maximizing the detection rate of the historical operational data, determined by the early warning condition determination unit through the balanced detection rate, false alarm rate, and false alarm rate threshold, includes: Determine the maximum and minimum values ​​of the anomaly prediction threshold, Tmax and Tmin, and then use the formula... Determine the anomaly prediction threshold T that maximizes the detection rate, where, The maximum value of the balanced detection rate. The false alarm rate is M, and the false alarm rate threshold is M.

[0044] In another embodiment, the road segment characteristic parameters in the data processing unit also include natural environmental parameters characterizing meteorological conditions.

[0045] The present invention provides a method for early warning of abnormalities in highway electromechanical equipment in the above embodiments, and provides a device for early warning of abnormalities in highway electromechanical equipment based on the method. The above method and device can improve the accuracy of early warning of electromechanical equipment, and can also maximize the ability to detect real faults within a limited budget.

[0046] This embodiment also discloses a computer device, such as... Figure 3 As shown, the computer device includes a processor and a memory, the memory storing at least one instruction, which is loaded and executed by the processor to implement any of the above-described methods for early warning of abnormalities in highway electromechanical equipment.

[0047] Furthermore, in the above-described implementation of the highway electromechanical equipment abnormality early warning device, the logical division of each program module is merely illustrative. In actual applications, the above functions can be assigned to different program modules as needed, for example, for the sake of corresponding hardware configuration requirements or the convenience of software implementation. That is, the internal structure of the alarm analysis system based on multimodal machine learning can be divided into different program modules to complete all or part of the functions described above.

[0048] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for early warning of abnormalities in highway electromechanical equipment, characterized in that, Applied to intelligent transportation platforms, the method includes: Historical operating data containing several operating parameters of the target electromechanical equipment is obtained, and then sample data of the target electromechanical equipment is determined through the historical operating data. The sample data includes the abnormal prediction values ​​and the true labels of the historical operating data. The road segment characteristic parameters of the target electromechanical equipment are determined, and the false alarm rate threshold of the target electromechanical equipment is determined through the road segment characteristic parameters. The balanced detection rate and false alarm rate of the historical operation data are determined based on the sample data. The road segment characteristic parameters include the basic information of the target electromechanical equipment, its layout area, and traffic flow. Then, by using the balanced detection rate, false alarm rate, and false alarm rate threshold, an anomaly prediction threshold that maximizes the detection rate of the historical operating data is determined, and the early warning conditions for the target electromechanical equipment are determined based on the anomaly prediction threshold. The system monitors the operating data of the target electromechanical equipment in real time, determines the real-time anomaly prediction value based on the operating data, and issues an anomaly warning when the real-time anomaly prediction value meets the warning conditions.

2. The method according to claim 1, characterized in that, The sample data for determining the target electromechanical equipment through the historical operating data includes: Feature extraction is performed on the historical operation data, and the extracted data is preprocessed to obtain the historical feature values ​​of the historical operation data; The historical feature values ​​are input into the trained fault prediction model to obtain the anomaly prediction values ​​of the historical operating data. The true label of the historical operating data is determined by analyzing the fault logs of the target electromechanical equipment. The abnormal predicted value and the true label are determined as the sample data of the target electromechanical equipment.

3. The method according to claim 1, characterized in that, The step of determining the false alarm rate threshold of the target electromechanical equipment through the road segment feature parameters includes: The risk value of the target electromechanical equipment is determined by quantifying each parameter in the road segment feature parameters and then combining the preset weight values ​​of each parameter. The mapping relationship between the risk value and the false alarm rate threshold is preset based on the business rules, and then the corresponding false alarm rate threshold is determined through the risk value.

4. The method according to claim 1, characterized in that, The method for determining the anomaly prediction threshold that maximizes the detection rate of the historical operational data by balancing the detection rate, the false alarm rate, and the false alarm rate threshold includes: Determine the maximum and minimum values ​​of the anomaly prediction threshold, Tmax and Tmin, and then use the formula... Determine the anomaly prediction threshold T that maximizes the detection rate, where, The maximum value of the balanced detection rate. The false alarm rate is M, and the false alarm rate threshold is M.

5. The method according to claim 1 or 3, characterized in that, The road segment characteristic parameters also include natural environmental parameters that characterize meteorological conditions.

6. A pre-warning device for abnormal electromechanical equipment on highways, characterized in that, The device, applied to an intelligent transportation platform, includes: The sample data determination unit is used to acquire historical operating data containing several operating parameters of the target electromechanical equipment, and then determine the sample data of the target electromechanical equipment through the historical operating data, wherein the sample data includes the abnormal prediction value and the true label of the historical operating data; The data processing unit is used to determine the road segment characteristic parameters of the target electromechanical equipment, determine the false alarm rate threshold of the target electromechanical equipment through the road segment characteristic parameters, and determine the balanced detection rate and false alarm rate of the historical operation data based on the sample data. The road segment characteristic parameters include the basic information of the target electromechanical equipment, its layout area, and traffic flow. The early warning condition determination unit is used to determine the anomaly prediction threshold that maximizes the detection rate of the historical operating data by further using the balanced detection rate, false alarm rate and the false alarm rate threshold, and to determine the early warning condition of the target electromechanical equipment based on the anomaly prediction threshold. An anomaly warning unit is used to monitor the operating data of the target electromechanical equipment in real time, determine the real-time anomaly prediction value through the operating data, and issue an anomaly warning when the real-time anomaly prediction value meets the warning conditions.

7. The apparatus according to claim 6, characterized in that, The sample data for determining the target electromechanical equipment using the historical operating data in the sample data determination unit includes: Feature extraction is performed on the historical operation data, and the extracted data is preprocessed to obtain the historical feature values ​​of the historical operation data; The historical feature values ​​are input into the trained fault prediction model to obtain the anomaly prediction values ​​of the historical operating data. The true label of the historical operating data is determined by analyzing the fault logs of the target electromechanical equipment. The abnormal predicted value and the true label are determined as the sample data of the target electromechanical equipment.

8. The apparatus according to claim 6, characterized in that, The data processing unit determines the false alarm rate threshold of the target electromechanical equipment based on the road segment feature parameters, including: The risk value of the target electromechanical equipment is determined by quantifying each parameter in the road segment feature parameters and then combining the preset weight values ​​of each parameter. The mapping relationship between the risk value and the false alarm rate threshold is preset based on the business rules, and then the corresponding false alarm rate threshold is determined through the risk value.

9. The apparatus according to claim 6, characterized in that, The anomaly prediction threshold in the early warning condition determination unit, which maximizes the detection rate of the historical operating data by balancing the detection rate, the false alarm rate, and the false alarm rate threshold, includes: Determine the maximum and minimum values ​​of the anomaly prediction threshold, Tmax and Tmin, and then use the formula... Determine the anomaly prediction threshold T that maximizes the detection rate, where, The maximum value of the balanced detection rate. The false alarm rate is M, and the false alarm rate threshold is M.

10. The apparatus according to claim 6 or 8, characterized in that, The road segment characteristic parameters in the data processing unit also include natural environmental parameters that characterize meteorological conditions.

Images