CN-121981672-A - Intelligent auditing method based on gas security inspection system

CN121981672ACN 121981672 ACN121981672 ACN 121981672ACN-121981672-A

Abstract

The invention relates to the technical field of data processing and utility safety management, and discloses an intelligent auditing method based on a gas safety inspection system. The safety inspection duration monitoring module dynamically calculates standard duration by utilizing a linear regression model based on factors such as house area, equipment number and the like, solves the problem of fixed threshold judgment rigidification, calculates theoretical minimum passing time by utilizing a space-time collision monitoring module based on spherical distance, checks physical movement limit to identify position fake, and utilizes an isolated forest algorithm to mine unknown outlier characteristics. And the system calculates the comprehensive risk value according to the weighting of the results of each module, and accordingly performs grading treatment of freezing, marking review or automatic archiving on the work order. The invention overcomes the defects of hysteresis of traditional manual auditing and incomplete coverage of a single rule model.

Inventors

WANG YUE
LI JING
HAN CHAOHUI
WU BIN
XIA YINGJIE

Assignees

博康智能信息技术有限公司

Dates

Publication Date: 20260505
Application Date: 20260106

Claims (10)

1. An intelligent auditing method based on a gas security inspection system is characterized by comprising the following steps: Acquiring gas security check worksheet data to be checked, wherein the gas security check worksheet data comprise timestamp information, geographical position coordinates, security check task attributes and on-site image credentials; Inputting the gas security check worksheet data into a preset parallel anomaly detection engine, and performing anomaly analysis of multiple dimensions in parallel to obtain a corresponding multi-dimensional detection result; The parallel anomaly detection engine comprises at least one of a security check duration monitoring module, a space-time collision monitoring module, a geographic position checking module and a workload load monitoring module; Based on the multi-dimensional detection result, calculating a comprehensive risk value of the gas security check work order data, and judging a risk level according to the comprehensive risk value; Triggering corresponding treatment operations according to the risk level, wherein the treatment operations comprise real-time freezing, marking manual review or automatic passing.
2. The intelligent auditing method based on a gas security inspection system according to claim 1, characterized in that the parallel execution of anomaly analysis of multiple dimensions comprises the following steps performed by a security inspection duration monitoring module: calculating the actual operation duration of the current task based on the check-in time and the submitting time in the timestamp information; acquiring the house area, the number of gas equipment and the traffic jam coefficient of the area where the gas equipment is located in the security inspection task attribute; respectively giving preset weights to the house area, the number of the gas equipment and the traffic jam coefficient, summing, and generating a dynamic standard duration for the current task after the superposition foundation is fixed and time-consuming; And comparing the actual operation duration with the dynamic standard duration, and if the actual operation duration is lower than the preset lower limit proportion of the dynamic standard duration or higher than the preset upper limit proportion of the dynamic standard duration, generating a detection result of duration abnormality.
3. The intelligent auditing method based on a gas security inspection system according to claim 1, characterized in that the parallel execution of anomaly analysis in multiple dimensions comprises the execution of the following steps by a space-time collision monitoring module: Obtaining the geographical position coordinates and the submitting time of the current gas security check work order data, and the geographical position coordinates and the submitting time of the previous gas security check work order data of the same security inspector; calculating the spherical physical distance or road network passing distance between two points according to the geographic position coordinates of the current and last gas security check worksheet data; calculating the time interval of two work orders according to the submitting time, and calculating the theoretical minimum passing time by combining the preset average passing speed; And if the time interval is smaller than the theoretical minimum passing time, judging that the non-physical movement behavior exists, and generating a detection result of the space-time collision abnormality.
4. The intelligent auditing method based on a gas security inspection system according to claim 1, characterized in that the parallel execution of anomaly analysis of multiple dimensions comprises the execution of the following steps by a geographic location verification module: acquiring a registration address coordinate of a target user, and acquiring a site report coordinate generated by fusing GPS (global positioning system), base station and WiFi (wireless fidelity) fingerprints in the gas security check worksheet data; calculating Euclidean distance deviation between the site report coordinates and the registered address coordinates; Judging whether the Euclidean distance deviation exceeds a preset error radius threshold value or judging whether the administrative city to which the on-site reporting coordinate belongs is consistent with the administrative city to which the registration address coordinate belongs; and if the error radius threshold value is exceeded or administrative cities are inconsistent, generating a detection result of abnormal positioning.
5. The intelligent auditing method based on a gas security system of claim 1, in which the performing of anomaly analysis in multiple dimensions in parallel includes performing the following steps by a workload monitoring module: Counting the total amount of work orders submitted by the current security inspector in the current day; Acquiring a historical average daily completion single-quantity average value of a current area, and adjusting the historical average daily completion single-quantity average value by utilizing a preset multiplying factor and a seasonal floating factor to obtain a dynamic workload upper limit threshold of the current day; and if the total quantity of the work orders exceeds the dynamic workload upper limit threshold, generating a detection result of workload load abnormality.
6. The intelligent auditing method based on a gas security inspection system according to claim 1, characterized in that the performing of anomaly analysis in multiple dimensions in parallel further comprises performing the following steps by an unsupervised anomaly detection module: extracting feature vectors in the gas security check worksheet data, wherein the feature vectors at least comprise operation duration, positioning deviation distance, number of uploaded photos and submitting time period; Processing the feature vector by using the constructed isolation forest model, and calculating the average path length of the current work order in a plurality of isolation trees; mapping the average path length into an anomaly score, and if the anomaly score indicates that the sample is an outlier, generating a detection result of the high-risk outlier.
7. The intelligent auditing method based on a gas security inspection system according to claim 1, characterized in that the performing exception analysis of multiple dimensions in parallel further comprises performing the following steps by a credential verification module: Detecting whether the on-site image certificate contains a preset key bill photo or not; Identifying the forced watermark information on the site image evidence, and checking whether the time and positioning data in the forced watermark information are consistent with the time stamp information and the geographic position coordinates in the gas security check work order data; if the key bill photo is missing or watermark information is inconsistent, a detection result of abnormal certificates is generated.
8. The intelligent auditing method based on a gas security inspection system according to claim 1, wherein the calculating the comprehensive risk value of the gas security inspection work order data based on the multi-dimensional detection result comprises: Respectively presetting corresponding risk weight coefficients for the security check duration monitoring module, the space-time collision monitoring module, the geographic position checking module and the workload load monitoring module; carrying out quantization processing on the detection results generated by each module to obtain risk scores of each dimension; And carrying out weighted summation on the risk scores of the dimensions and the corresponding risk weight coefficients to obtain the comprehensive risk value.
9. The intelligent auditing method based on a gas security system according to claim 1, wherein the performing the corresponding treatment operation according to the risk level comprises: If the comprehensive risk value is higher than a preset high risk threshold, judging that the comprehensive risk value is high in risk level, executing the operation of freezing the work order in real time and sending an alarm, and blocking the follow-up settlement flow; if the comprehensive risk value is between a preset low risk threshold value and a preset high risk threshold value, judging that the comprehensive risk value is a medium risk level, executing the operation of marking to be manually rechecked, and highlighting an abnormal dimension on an auditing interface; And if the comprehensive risk value is lower than the low risk threshold, judging that the comprehensive risk value is of a low risk level, and executing operations of logging and automatically passing through.
10. The intelligent auditing method based on a gas security inspection system according to claim 2, characterized in that the respectively giving preset weights to the floor area, the number of gas devices and the traffic jam coefficient and summing the weights comprises: multiplying the house area by a first weight coefficient to obtain an area influence duration; Multiplying the number of the gas equipment by a second weight coefficient to obtain equipment influence duration; multiplying the traffic jam coefficient by a third weight coefficient to obtain traffic influence duration; And adding the area influence time length, the equipment influence time length, the traffic influence time length and the basic fixed time consumption to be used as the dynamic standard time length.

Description

Intelligent auditing method based on gas security inspection system Technical Field The invention relates to the technical field of data processing and utility safety management, in particular to an intelligent auditing method based on a gas safety inspection system. Background With the explosive growth of urban gas users, regular household safety inspection becomes a key defense line for guaranteeing safe operation of the tail end of a gas pipe network and preventing gas accidents. In the face of massive security check worksheets and scattered and flowing security check operators, how to ensure the authenticity and effectiveness of each security check service becomes a pain point for operation management of gas enterprises. The current security inspection quality control technology mainly relies on manual post-hoc inspection or rule matching systems based on simple logic. However, this conventional management mode has significant limitations in practical applications. The existing automatic auditing system generally adopts a fixed threshold value to carry out one-cut type judgment, for example, only a unified shortest operation duration standard is simply set, but the obvious influence of objective factors such as house area, gas equipment quantity, traffic environment and the like on the actual operation efficiency is ignored. Such rigidified decision logic not only can easily misjudge the efficient operation of a skilled worker as a violation, but also makes it difficult to identify hidden fake behaviors in which an operator avoids rules by deliberately hanging up. In addition, with the popularization of intelligent terminal technology, the technical cheating means such as virtual positioning software, photo tampering tools and the like are increasingly renovated. The existing supervision technology often lacks depth consistency verification of worksheet space-time logic, is difficult to capture abnormal behaviors crossing geographic distances in a short time from a physical motion rule layer, and also lacks depth fake identification capability of unstructured image credentials. More importantly, the traditional monitoring mechanism based on the fixed rule base can only identify the known violation pattern, and the lack of autonomous mining and sensing capability of unknown abnormal features in data distribution leads to the fact that the supervision means is always lagged behind the updating of the cheating means, and cannot realize real-time accurate blocking of the security inspection whole flow risk, so that the intelligent auditing method based on the gas security inspection system is provided for overcoming the defects. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an intelligent auditing method based on a gas security inspection system, which solves the technical problems that in the existing gas security inspection supervision mode, the artificial auditing hysteresis is strong, a single fixed rule model cannot adapt to complex operation scenes, so that misjudgment and report omission is caused, and hidden fake behaviors such as space-time logic conflict and the like are difficult to recognize in real time. The intelligent auditing method based on the gas safety inspection system is realized by constructing a parallel anomaly detection engine to automatically analyze the safety inspection worksheet data in full dimension. The method comprises the steps of obtaining gas security check worksheet data to be audited, wherein the gas security check worksheet data comprise timestamp information, geographical position coordinates, security check task attributes and site image credentials, inputting the gas security check worksheet data into a preset parallel anomaly detection engine, performing anomaly analysis of multiple dimensions in parallel to obtain corresponding multi-dimensional detection results, wherein the parallel anomaly detection engine comprises at least one of a security check duration monitoring module, a space-time collision monitoring module, a geographical position checking module, a workload monitoring module, an unsupervised anomaly detection module and a credential checking module, calculating comprehensive risk values of the gas security check worksheet data based on the multi-dimensional detection results, judging risk levels according to the comprehensive risk values, and triggering corresponding treatment operations according to the risk levels, wherein the treatment operations comprise real-time freezing, manual review of marks or automatic passing. In the technical scheme of security check duration monitoring, the invention innovatively introduces a dynamic duration judging mechanism based on task characteristics, and abandons the traditional fixed threshold judging mode. In specific implementation, calculating the actual job duration of the current task based on the check-in time and the submission time in the timestamp informationAcq