Power transformation unplanned operation multi-dimensional monitoring system based on artificial intelligence

Abstract

The invention discloses a power transformation unplanned operation multi-dimensional monitoring system based on artificial intelligence, which relates to the technical field of intelligent substation monitoring and comprises a pre-acquisition module, a plan judgment module, a heat source area screening module, a risk prediction module and an early warning module. The method comprises the following steps: constructing a plurality of functional areas according to the monitoring requirements of substation field operation, and pre-collecting the activity time of operators in the substation field area and the planned operation authorized entry and exit time; unplanned operators in the current substation field area are judged, and a heat source analysis instruction is sent out; screening out a heat source gathering function area in the substation field area, and sending a risk prediction signal to the outside; carrying out prediction analysis on the unplanned operation risk degree of the heat source gathering functional area; and judging whether the unplanned operation performed in the current heat source gathering functional area has a risk or not, and obtaining and executing a risk early warning instruction of a corresponding level.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent substation monitoring, and specifically to a multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence. Background Technique

[0002] With the continuous and in-depth application of artificial intelligence technology in industrial systems, the intelligent operation and maintenance of power systems has become one of the important development directions in the construction of modern power grids; especially in the field of intelligent substation monitoring, the use of artificial intelligence technology to achieve job behavior perception, risk warning and dispatching decision-making assistance gradually replaces the traditional human-dominated safety monitoring mode; especially in the key scenario of on-site substation operation monitoring, higher intelligent upgrade requirements are put forward for the substation operation monitoring system.

[0003] However, in the actual operation of substations, due to the frequent occurrence of sudden tasks, temporary repairs and missing reports of operation records, when unplanned operation behaviors occur, the current on-site substation operation monitoring system lacks the dynamic recognition of unauthorized unplanned operations; at the same time, the existing on-site substation operation monitoring systems generally lack the screening of unplanned operation aggregation areas, and it is difficult to comprehensively judge risks based on multi-source information such as heat source status, personnel trajectories and electrical disturbances; resulting in problems such as lagging recognition of unplanned operation behaviors, high misjudgment rates, poor operation risk assessment capabilities and untimely risk warning responses, seriously affecting the operation and maintenance safety and dispatching efficiency of substations. Therefore, there is an urgent need to establish a new type of intelligent monitoring system that integrates perception, analysis and warning response to achieve high-precision recognition and risk control of unplanned operations. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence, which solves the problems in the above background technology.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence, including a pre-acquisition module, a plan determination module, a heat source area screening module, a risk prediction module and a warning module; The pre-acquisition module is used to construct several functional areas according to the monitoring requirements of on-site substation operations, and pre-acquire the activity time of operating personnel in the on-site substation area and the planned operation authorized entry and exit time. The plan determination module is used to determine unplanned operating personnel in the current on-site substation area and send out a heat source analysis instruction. The heat source area screening module, based on the received heat source analysis instruction, screens out the heat source aggregation functional areas in the on-site substation area and sends out a risk prediction signal. The risk prediction module is used to predict and analyze the risk level of unplanned operations in the heat source aggregation functional area after receiving a risk prediction signal; The early warning module is used to determine whether there is a risk in the unplanned operations carried out in the current heat source aggregation functional area, so as to obtain a risk early warning instruction of the corresponding level and execute it.

[0006] Preferably, the pre-acquisition module includes a deployment unit and a time period acquisition unit; The deployment unit is used to divide the substation site area into several functional areas according to the monitoring requirements of substation on-site operations, combined with the substation GIS structure diagram and the functional uses of the substation site area. Use the inspection robot integrated with multiple groups of sensors to conduct all-day inspections on the constructed several functional areas. At the same time, voltage sensors are deployed on the bus nodes of the substation site area, and an access control system is deployed at the substation entrance and exit. Combined with wireless communication technology, relevant data information is uploaded to the central data platform; The functional uses of the substation site area include high-voltage main equipment operation area, secondary control area, access control supervision area, tool storage area and computer room area; multiple groups of sensors include high-definition optical cameras, infrared thermal imagers, timers, audible and visual alarms, edge computing modules and wireless communication modules; the access control system is used to store the authorized entry and exit times of substation planned operations and the identity verification of entry and exit personnel; the central data platform is used to preprocess, perform AI edge computing and store the uploaded relevant data information.

[0007] Preferably, the time period acquisition unit is used to monitor the activity time status of the operating personnel in the substation site area according to the high-definition optical camera and timer integrated in the inspection robot, combined with the human body recognition algorithm, to obtain the set of personnel activity time periods Tj. And according to the access control system deployed at the substation entrance and exit, collect the authorized entry and exit times of the planned operations stored in the access control system, obtain the set of authorized entry and exit time periods Tc for operations, and upload the obtained set of personnel activity time periods Tj and the set of authorized entry and exit time periods Tc for operations to the central data platform for storage.

[0008] Preferably, the plan determination module includes a coupling degree analysis unit and a determination unit; The coupling degree analysis unit is used to perform feature recognition on the relevant data information of the central data platform, correlate the extracted set of personnel activity time periods Tj and the set of authorized entry and exit time periods Tc for operations, and analyze the coupling degree between the activity time period of the operating personnel in the substation site area and the entry time period of the authorized operating personnel, to obtain the time coupling degree coefficient Xoh, which is specifically calculated through the following formula; ; In the formula, It is represented as the intersection period of the set Tj of personnel activity time periods and the set Tc of operation authorization entry and exit time periods. It is represented as the union period of the set Tj of personnel activity time periods and the set Tc of operation authorization entry and exit time periods, and A represents the first correction constant. The determination unit is used to determine whether the operating personnel in the substation site area are unauthorized unplanned operating personnel after comparative analysis according to the obtained value of the time coupling coefficient Xoh. The specific analysis process is as follows: If 0 ≤ time coupling coefficient Xoh < 1, it means that the operating personnel in the current substation site area are unauthorized unplanned operating personnel, and at this time, a heat source analysis instruction is sent outwards. If the time coupling coefficient Xoh = 1, it means that the operating personnel in the current substation site area are not unauthorized unplanned operating personnel, and at this time, no additional heat source analysis instruction is sent.

[0009] Preferably, the heat source area screening module includes a heat source analysis unit and a screening unit. After receiving the heat source analysis instruction, the heat source analysis unit is used to, based on a number of constructed functional areas, and according to the infrared thermal imager integrated in the inspection robot, take infrared images of the shooting areas of each functional area, extract the heat source target area by using the temperature threshold segmentation algorithm, map the extracted heat source target area into the pseudo-color image space, construct the color scale map of each functional area, analyze the pixel traversal of the red channel in the color scale map, and use the local extreme value extraction algorithm to identify the maximum value of the color scale of each functional area to obtain the heat source area color scale peak Ssj in each functional area; based on the number of pixels in the color scale map of each functional area, and use the physical area conversion model to convert it into the actual heat source area to obtain the total heat source area Smj in each functional area, and through the connected domain analysis method, number and identify different heat source areas in each functional area to obtain the number Nss of independent heat source areas in each functional area; construct the heat source area color scale peak Ssj, the total heat source area Smj and the number Nss of independent heat source areas in each functional area into relevant heat source state data information. According to the dimensions of each functional area in the substation CAD design drawing, obtain the area Sqy of each functional area, and associate it with the relevant heat source state data information. After dimensionless processing, analyze the heat source distribution of each functional area to obtain the heat source density coefficient Xmj of each functional area, which is specifically calculated by the following formula: ; In the formula, represents the heat source density coefficient of the i-th functional area, represents the area of the i-th functional area, The peak value of the heat source area color scale of the i-th functional area The total area of the heat source area of the i-th functional area The number of independent heat source areas of the i-th functional area, where 、 and are all represented as weight values

[0010] Preferably, the screening unit calculates the average value of the heat source density coefficients Xmj of each functional area in the substation site area based on the obtained heat source density coefficients Xmj of each functional area and according to the statistical mean algorithm, and obtains the average heat source density coefficient of the substation site area ; After comparing and analyzing the heat source density coefficients Xmj of each functional area in the substation site area with the average heat source density coefficient of the substation site area respectively, the heat source aggregation functional areas are screened out and a risk warning is sent outwards. The specific content of the comparative analysis is as follows After comparing and analyzing the heat source density coefficients Xmj of each functional area in the substation site area with the average heat source density coefficient of the substation site area respectively, if the heat source density coefficient Xmj of the current functional area exceeds the average heat source density coefficient , the corresponding functional area is classified as a heat source aggregation functional area and marked at the same time. At this time, a risk prediction signal is sent outwards. If the heat source density coefficient Xmj of the current functional area does not exceed the average heat source density coefficient , the corresponding functional area is classified as a normal heat source functional area, and no additional risk prediction signal is sent out at this time

[0011] Preferably, the risk prediction module includes a behavioral risk analysis unit, a disturbance analysis unit and a prediction unit The behavioral risk analysis unit is used to monitor the behavioral states of unplanned workers in the heat source aggregation functional area during the personnel activity period according to the high-definition optical camera and RFID reader integrated in the inspection robot after receiving the risk prediction signal, and obtain relevant personnel behavioral state data information. Among them, the relevant personnel behavioral state data information includes the number of times Ccs of personnel climbing heights, the climbing duration Tsc and the proportion Bcd of safety-wearing personnel in the heat source aggregation functional area By performing feature recognition on the data information of the behavior status of relevant personnel, the number of times of personnel climbing heights Ccs, the duration of climbing heights Tsc, and the proportion of personnel wearing safety equipment Bcd in the heat source aggregation functional area are correlated. After dimensionless processing, the behavior risk degree of unplanned operation personnel in the heat source aggregation functional area is analyzed to obtain the behavior risk coefficient Xxw, which is specifically calculated through the following formula; ; In the formula, , and respectively represent the weight values of the number of times of personnel climbing heights Ccs, the duration of climbing heights Tsc, and the proportion of personnel wearing safety equipment Bcd.

[0012] Preferably, the perturbation analysis unit is used to monitor the bus voltage status in the heat source aggregation functional area in real time according to the voltage sensors deployed on the bus nodes, and obtain relevant voltage status data information. Among them, the relevant voltage status data information includes the voltage value Vss at each monitoring time point in the heat source aggregation functional area. By performing feature extraction on the historical data stored in the central data platform and combining the statistical mean algorithm, the historical voltage mean value Vjz of the heat source aggregation functional area is obtained. By correlating the voltage value Vss at each monitoring time point in the heat source aggregation functional area with the historical voltage mean value Vjz, after dimensionless processing, the bus voltage perturbation situation in the heat source aggregation functional area during the personnel activity period is analyzed to obtain the bus voltage perturbation coefficient Xrd, which is specifically calculated through the following formula; ; In the formula, represents the voltage value at the jth monitoring time point, j = 1, 2, 3,..., n, and n represents the number of monitoring time points during the personnel activity period, represents the degree of fluctuation of the voltage value Vss at the jth monitoring time point relative to the historical voltage mean value Vjz.

[0013] Preferably, the prediction unit is used to correlate the heat source density coefficient Xmj of each functional area in the heat source aggregation functional area with the behavior risk coefficient Xxw and the bus voltage perturbation coefficient Xrd of the heat source aggregation functional area. After dimensionless processing, the unplanned operation risk degree in the heat source aggregation functional area is predicted and analyzed to obtain the unplanned operation risk prediction index Zyc of the heat source aggregation functional area, which is specifically obtained through the following formula: ; In the formula, represents the heat source density coefficient of the zth functional area in the heat source aggregation functional area, z = 1, 2, 3,..., m, and m represents the number of functional areas in the heat source aggregation functional area, The behavior risk coefficient represented as the heat source aggregation function area The bus voltage disturbance coefficient represented as the heat source aggregation function area, where , and are all represented as weight values, and R is represented as the second correction constant.

[0014] Preferably, the early warning module is used to compare and analyze the unplanned operation risk prediction index Zyc of the heat source aggregation function area with the preset risk prediction threshold Y to determine whether there is a risk in the unplanned operation currently carried out in the heat source aggregation function area, so as to obtain the risk early warning instruction of the corresponding level and execute it. The specific content is as follows: If the unplanned operation risk prediction index Zyc of the heat source aggregation function area ≥ the risk prediction threshold Y, it means that there is a risk in the unplanned operation currently carried out in the heat source aggregation function area. Generate a first-level risk early warning instruction and execute it. The execution content is: guide the inspection robot to move to the heat source aggregation function area, and emit a sound and light alarm through the integrated sound and light alarm, and at the same time play the voice prompt of "There is a risk in the current unplanned operation, please stop the operation immediately"; link the access control system to prohibit unplanned operation personnel from entering again, and notify the power supervisor to check the on-site situation; If the unplanned operation risk prediction index Zyc of the heat source aggregation function area < the risk prediction threshold Y, it means that there is no risk in the unplanned operation currently carried out in the heat source aggregation function area. Generate a second-level risk early warning instruction and execute it. The execution content is: guide the inspection robot to move to the heat source aggregation function area to play the voice prompt of "The current operation behavior is unauthorized, please pay attention to operation safety", and continuously monitor the unplanned operation behavior of the heat source aggregation function area. At the same time, mark the current operation event as an unplanned operation behavior without risk and record it in the central data platform.

[0015] The present invention provides a multi-dimensional monitoring system for substation unplanned operations based on artificial intelligence, which has the following beneficial effects: (1) By constructing a multi-dimensional monitoring system composed of pre-acquisition, plan determination, heat source screening, risk prediction and early warning response, it breaks through the technical bottleneck of the traditional substation monitoring system that fails to identify unplanned operation behaviors in a timely manner and respond inaccurately; the system integrates artificial intelligence behavior recognition algorithms, infrared heat source image analysis technology and voltage disturbance monitoring models, and can realize the whole-process dynamic perception, classification recognition, focusing and positioning, and risk level prediction of unplanned operations at the substation site. Based on hierarchical early warning control, it effectively improves the accuracy of operation behavior recognition, the stability of system operation and the intelligent level of dispatching and handling, and constructs a more predictive and proactive safety guarantee mechanism for the substation.

[0016] (2) By deploying inspection robots with the capabilities of human body recognition and time synchronization perception, collect the activity time periods of operators in each functional area of the substation, and combine with the authorized entry and exit times recorded in the access control system. Use the time coupling coefficient Xoh to analyze the consistency between the operation behavior and the planned authorization, so as to achieve rapid identification and classification judgment of unauthorized operation behaviors; greatly reduce the misjudgment rate of unplanned operations, can quickly identify potential violations at the initial stage of the behavior, effectively shorten the system response time and enhance the initiative of on-site monitoring.

[0017] (3) Based on the infrared thermal imager integrated in the inspection robot, collect the infrared thermal images in the functional area and construct a pseudo-color map of the heat source. Through algorithms such as temperature threshold segmentation, color scale peak extraction, and connected component analysis, obtain the color scale peak Ssj of the heat source area, the total heat source area Smj, and the number of independent heat sources Nss in each area, and combine with the area of the area to calculate the heat source density coefficient Xmj, so as to realize the intelligent screening of the functional areas with heat source aggregation in the substation site area, effectively solving the technical problem that it is difficult for the substation site operation monitoring system to identify abnormal heat source areas.

[0018] (4) In the risk prediction link, comprehensively consider the heat source density coefficient Xmj of the functional area with heat source aggregation, the behavior risk coefficient Xxw of the functional area with heat source aggregation, and the bus voltage disturbance coefficient Xrd, construct a multi-dimensional risk prediction model, obtain the unplanned operation risk prediction index Zyc of the functional area with heat source aggregation, and realize a multi-dimensional fusion evaluation mechanism for unplanned operation risks; not only can grade the risk levels of the identified unplanned behaviors, but also can predict their potential impacts on system stability and trigger differential level warning response strategies, effectively avoiding the situation in the traditional system where the risk assessment lags behind, seriously affecting the operation and maintenance safety and dispatching efficiency of the substation. Brief Description of the Drawings

[0019] Figure 1 It is a block diagram of a multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence of the present invention; Figure 2 It is a logical thinking diagram of a multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence of the present invention; Figure 3 It is a device connection relationship diagram of a multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence of the present invention. Detailed Embodiment

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Embodiment 1 Please refer to Figure 1 and Figure 2 , the present invention provides a multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence, including a pre-acquisition module, a plan determination module, a heat source area screening module, a risk prediction module, and an early warning module; The pre-acquisition module is used to construct several functional areas according to the monitoring requirements of substation on-site operations, and pre-acquire the activity time of on-site operators in the substation area and the planned operation authorization in-and-out time; The plan determination module is used to determine unplanned operators in the current substation on-site area and send out a heat source analysis instruction; The heat source area screening module, based on the received heat source analysis instruction, screens out the heat source aggregation functional areas in the substation on-site area and sends out a risk prediction signal; The risk prediction module is used to predict and analyze the risk level of unplanned operations in the heat source aggregation functional area after receiving the risk prediction signal; The early warning module is used to determine whether there is a risk in the unplanned operation carried out in the current heat source aggregation functional area, so as to obtain a risk early warning instruction of the corresponding level and execute it.

[0022] In this embodiment, by constructing a distributed perception and intelligent analysis system including a pre-acquisition module, a plan determination module, a heat source area screening module, a risk prediction module, and an early warning module, the accuracy, response speed, and risk prevention and control capabilities of the substation in on-site operation behavior monitoring are significantly improved. Different from the traditional safety management method centered on manual inspection and static monitoring, its innovation lies in the introduction of artificial intelligence technologies such as behavior recognition, infrared heat source aggregation analysis, and system disturbance modeling, realizing the dynamic recognition and accurate determination of unplanned operation behaviors from the source. Through the coupled calculation between the activity period of the operation personnel and the authorized access record, the system can quickly judge whether it is an unplanned operation. On this basis, by using the heat source state analysis method, the abnormal heat source aggregation area is screened in real time, and a heat source density model is constructed to realize the screening of the spatial area of abnormal operation behaviors. Further, by introducing operation behavior indicators and bus voltage disturbance factors to construct a risk prediction of unplanned operations, the system can quantify the potential operation risk level in advance and implement a differential risk early warning response mechanism. This integrated multi-dimensional monitoring method not only effectively makes up for the deficiencies of the traditional system in unplanned operation identification, heat source screening, and operation risk assessment, but also realizes the intelligent linkage and data closed-loop of operation behaviors, environmental parameters, and system states, significantly improving the safety control level and risk disposal initiative in the substation operation and maintenance process.

[0023] Embodiment 2 Please refer to Figure 1 and Figure 3 , specifically: The pre-acquisition module includes a deployment unit and a time period acquisition unit; The deployment unit is used to divide the substation site area according to the monitoring requirements of substation on-site operations, combine the substation GIS structure diagram and the functional uses of the substation site area, construct several functional areas, use the inspection robot integrated with multiple groups of sensors to conduct all-day inspections on the constructed several functional areas, deploy voltage sensors on the bus nodes of the substation site area, and deploy an access control system at the substation entrance and exit, and upload relevant data information to the central data platform in combination with wireless communication technology; The functional uses of the substation site area include high-voltage main equipment operation area, secondary control area, access control supervision area, tool storage area, and computer room area; the multiple groups of sensors include high-definition optical cameras, infrared thermal imagers, timers, audible and visual alarms, edge computing modules, and wireless communication modules; the access control system is used to store the authorized entry and exit times of substation planned operations and the identity verification of entry and exit personnel; the central data platform is used to preprocess, perform AI edge computing, and store the uploaded relevant data information.

[0024] Specifically, the time period collection unit is used to monitor the activity time status of the operators in the substation site area according to the high-definition optical camera and timer integrated in the inspection robot, and in combination with the human body recognition algorithm, to obtain the set Tj of the operator activity time periods. And according to the access control system deployed at the entrance and exit of the substation, it collects the planned operation authorization in-and-out time stored in the access control system to obtain the set Tc of the operation authorization in-and-out time periods, and uploads the obtained set Tj of the operator activity time periods and the set Tc of the operation authorization in-and-out time periods to the central data platform for storage.

[0025] It should be noted that the set Tj of the operator activity time periods refers to the set of time intervals in which the system, through the high-definition optical camera and timer integrated in the inspection robot, and in combination with the human body recognition algorithm, real-time identifies and records the actual appearance and stay time of the operators in the substation site area, reflecting the real operation time periods of the personnel; while the set Tc of the operation authorization in-and-out time periods refers to the set of personnel authorization in-and-out time information corresponding to the planned operation tasks extracted by the system from the access control system, representing the operation time range permitted after approval; the core function of both is to judge whether the current operation behavior of the personnel belongs to the planned operation through the comparison and coupling analysis in the time domain, so as to provide a judgment basis for identifying unplanned operations later, and it is the basic data support for realizing operation compliance analysis and dynamic deviation detection.

[0026] In this embodiment, through the collaborative design of the deployment unit and the time period collection unit, the global pre-coverage and dynamic collection mechanism of the substation on-site operation behavior data is realized, which effectively breaks through the limitations of the traditional system that relies on single-point monitoring and manual statistics, and has the advantages of highly intelligent, regionalized and structured front-end perception; the deployment unit divides the substation into fine areas based on the substation GIS structure diagram and functional use zoning logic, and constructs multiple functional areas such as high-voltage main equipment operation area, secondary control area, access control supervision area, tool storage area and computer room, and combines the high-definition optical camera, infrared thermal imager, timer, sound and light alarm and other groups of sensors integrated in the inspection robot to realize all-weather, blind spot-free inspection; especially based on the configuration of deploying voltage sensors at bus nodes and access control systems at entrances and exits, the system can realize multi-source information on operation behavior, equipment operation and personnel passage. Synchronous perception and seamless upload build a complete front-end information flow closed loop; the time period collection unit uses the human body recognition algorithm and the timer linkage recognition to accurately extract the personnel activity time period set Tj in each functional area, and synchronously compares it with the operation authorization entry and exit time period set Tc recorded by the access control system to achieve the time coupling modeling between personnel operation behavior and operation permission; by uniformly uploading these time period data to the central data platform for structured processing and storage, it not only provides a high-quality data foundation for subsequent unplanned operation identification, heat source aggregation screening and risk prediction, but also significantly enhances the system's real-time perception capability of sudden behavior, unauthorized operation and unauthorized operation, reflecting a forward-looking intelligent perception mechanism that integrates spatial information, behavior recognition and operation plan verification, and has obvious technical advantages in improving the on-site operation safety and data integrity of substations.

[0027] Example 3 Please refer to Figure 1 ,Specifically: the plan determination module includes a coupling degree analysis unit and a determination unit; The coupling degree analysis unit is used to identify the characteristics of the relevant data information of the central data platform, associate the extracted personnel activity time period set Tj and the operation authorization entry and exit time period set Tc, analyze the coupling degree between the operation personnel activity time period and the authorized operation personnel entry time period in the substation site area, and obtain the time coupling coefficient Xoh, which is specifically calculated and obtained by the following formula; ; In the formula, It is represented as the intersection of the personnel activity time period set Tj and the operation authorization entry and exit time period set Tc. It is represented as the union period of the personnel activity time period set Tj and the operation authorization entry and exit time period set Tc, and A is represented as the first correction constant; It should be noted that the formula in the coupling degree analysis unit performs a ratio operation on the intersection and union of the set Tj of personnel activity time periods and the set Tc of job authorization entry and exit time periods, and introduces the first correction constant A, which reflects the degree of overlap between the personnel behavior time and the authorization time; this formula is closely related to the deficiencies proposed in the background technology and effectively solves the problem that it is difficult for traditional systems to determine whether the operating personnel are working within the authorized time; traditional monitoring means often only rely on access control records and it is difficult to identify whether personnel are working in key areas during unauthorized times, while this formula can quantitatively determine whether the operating behavior is consistent with the authorization plan by constructing a coupling degree index in the time dimension, and further identify the occurrence of unplanned operating behaviors; the core role of the time coupling degree coefficient Xoh is to provide a computable and decidable quantitative basis for behavioral compliance, which is the key trigger condition for starting the subsequent heat source analysis and risk prediction processes.

[0028] The determination unit is used to determine whether the operating personnel in the substation site area are unauthorized unplanned operating personnel after comparative analysis based on the obtained value of the time coupling degree coefficient Xoh. The specific analysis process is as follows; If 0 ≤ time coupling degree coefficient Xoh < 1, it means that the operating personnel in the current substation site area are unauthorized unplanned operating personnel, and at this time, a heat source analysis instruction is sent outwards; If the time coupling degree coefficient Xoh = 1, it means that the operating personnel in the current substation site area are not unauthorized unplanned operating personnel, and at this time, no additional heat source analysis instruction is sent.

[0029] In this embodiment, through the collaborative action of the coupling degree analysis unit and the determination unit, an intelligent recognition mechanism for unplanned operations based on the cross-correlation of time series and job authorization data is constructed, significantly improving the system's discrimination efficiency and response accuracy for abnormal operation behaviors. Among them, the coupling degree analysis unit constructs a time coupling degree coefficient Xoh by performing intersection and union operations on the set Tj of personnel activity time periods and the set Tc of job authorization entry and exit time periods uploaded to the central data platform, and reflects the time-domain matching degree between personnel behaviors and planned permissions through this index. Compared with the traditional single access control comparison and manual ticket review mode, this method can not only dynamically identify illegal stays or sudden behaviors outside the authorized time period, but also has good fault tolerance and adaptability, and is particularly suitable for dealing with atypical behaviors such as temporary emergency repairs and missed report operations. By introducing the first correction constant A, the formula has the ability to mathematically balance extreme situations, thus ensuring the stability of the determination result. The determination unit sets judgment rules according to the numerical range of the time coupling degree coefficient Xoh, realizes the automatic identification and classification of unplanned operation personnel, and uses this as a trigger condition to link the downstream heat source analysis module, realizing the closed-loop starting point of operation risk analysis. The significant advantage of this mechanism is that through the quantitative modeling of the time coupling degree, the process of whether it is a planned operation, which originally relied on manual judgment, is transformed into a quantifiable index that can be calculated, compared, and traced, truly realizing the dynamic matching and deviation detection between the substation operation plan and on-site behaviors, improving the intelligent level of the entire system in abnormal recognition, rapid response, and automatic disposal, and providing a reliable pre-judgment basis for subsequent risk assessment and early warning.

[0030] Embodiment 4 Please refer to Figure 1 , specifically: The heat source area screening module includes a heat source analysis unit and a screening unit; The heat source analysis unit is used to, after receiving a heat source analysis instruction, based on a number of constructed functional areas and according to the infrared thermal imager integrated in the inspection robot, take infrared images of the shooting areas for each functional area, extract the heat source target area using the temperature threshold segmentation algorithm, map the extracted heat source target area into the pseudo-color image space, construct the color scale map for each functional area, analyze the red channel of the color scale map by pixel traversal, and use the local extreme value extraction algorithm to identify the maximum color scale value of each functional area, so as to obtain the heat source area color scale peak Ssj in each functional area; based on the number of pixels in the color scale map of each functional area and using the physical area conversion model to convert it into the actual heat source area, obtain the total heat source area Smj in each functional area, and through the connected component analysis method, number and identify different heat source areas in each functional area to obtain the number Nss of independent heat source areas in each functional area; construct the heat source area color scale peak Ssj, the total heat source area Smj and the number Nss of independent heat source areas in each functional area into relevant heat source status data information; It should be noted that the color scale peak Ssj represents the color scale value corresponding to the strongest temperature response in the heat source area extracted through the analysis of the red channel pixels of the pseudo-color map in the infrared thermal imaging image, reflecting the upper limit of the heat source intensity; the total heat source area Smj represents the number of pixels identified as the heat source target area, which is converted into the actual physical area in combination with the spatial resolution of the image, reflecting the heat source coverage; the number Nss of independent heat source areas identifies heat source individuals that are not connected to each other in space through the connected component analysis method, representing the degree of dispersion of the heat source in space; these three indicators jointly constitute a multi-dimensional expression system of the heat source status in the functional area, which can provide basic support for constructing the heat source density coefficient subsequently, and is used to determine whether there are multi-point abnormal aggregations, high-intensity heat sources and multi-person collaborative behaviors in the current area, so as to realize the spatial focus and risk identification of unplanned operation behaviors, and improve the sensitivity and response accuracy of the system to the abnormal operation situation.

[0031] According to the dimensions of each functional area in the substation CAD design drawing, obtain the area Sqy of each functional area, associate it with the relevant heat source status data information, and after dimensionless processing, analyze the heat source distribution of each functional area to obtain the heat source density coefficient Xmj of each functional area, which is specifically calculated through the following formula; ; In the formula, represents the heat source density coefficient of the i-th functional area, represents the area of the i-th functional area, represents the heat source area color scale peak of the i-th functional area, represents the total heat source area of the i-th functional area, The number of independent heat source regions represented as the i-th functional region, where , and are all represented as weight values.

[0032] It should be noted that the formula for calculating the heat source density coefficient Xmj takes into account the color scale peak Ssj of the heat source regions, the total area Smj of the heat source regions, and the number Nss of independent heat source regions within each functional region, and is constructed through normalization processing by the regional area Sqy, reflecting the heat source intensity, distribution range, and aggregation degree per unit area, which is a multi-dimensional quantitative expression of the heat source activity density in the functional region; this formula is closely related to the deficiencies proposed in the background technology, and is specifically optimized for the problem that traditional systems are difficult to identify regions of abnormal heat source aggregation and quantify the spatial distribution of heat source risks; by combining the infrared image perception data with the spatial scale in the CAD drawing, the system can evaluate the heat source aggregation degree based on the real size, and screen out the regions of abnormal heat source aggregation through statistical analysis of the heat source density coefficient Xmj, so as to identify potential multi-person operation behaviors and power equipment heating conditions in advance. The heat source density coefficient Xmj not only provides an accurate positioning basis for spatial thermal anomalies, but also significantly improves the spatial recognition ability and intelligent judgment level of the substation operation monitoring system.

[0033] Specifically, the screening unit calculates the average value of the heat source density coefficients Xmj of each functional region in the substation site area based on the obtained heat source density coefficients Xmj of each functional region and according to the statistical mean algorithm, and obtains the average heat source density coefficient of the substation site area ; After comparing and analyzing the heat source density coefficients Xmj of each functional region in the substation site area with the average heat source density coefficient of the substation site area respectively, the functional regions with heat source aggregation are screened out and a risk warning is sent outwards. The specific content of the comparison and analysis is as follows: After comparing and analyzing the heat source density coefficients Xmj of each functional region in the substation site area with the average heat source density coefficient of the substation site area respectively, if the heat source density coefficient Xmj of the current functional region exceeds the average heat source density coefficient , the corresponding functional region is designated as a functional region with heat source aggregation and marked at the same time. At this time, a risk prediction signal is sent outwards. If the heat source density coefficient Xmj of the current functional region does not exceed the average heat source density coefficient , the corresponding functional region is designated as a normal heat source functional region, and no additional risk prediction signal is sent out at this time.

[0034] In this embodiment, through the deep cooperation between the heat source analysis unit and the screening unit, an intelligent heat source identification and aggregation analysis mechanism based on the combination of infrared thermal sensing information and spatial image features is constructed, significantly improving the spatial positioning accuracy of unplanned operation behaviors at the substation site and the screening efficiency of high-risk areas; after receiving the heat source analysis instruction, the heat source analysis unit can, based on the constructed functional area division, call the infrared thermal images taken by the inspection robot and combine with the temperature threshold segmentation algorithm to accurately extract the heat source target area; by constructing a pseudo-color map and extracting the color scale peak Ssj, the total heat source area Smj is obtained through physical area conversion, and the connected domain analysis technology is used to identify the number of heat sources Nss, forming a heat source state data structure integrating heat intensity, distribution range, and target quantity; in particular, the system obtains the area Sqy of the region based on the substation CAD drawing, organically correlates the structural dimensions with the heat source characteristics, and establishes the heat source density coefficient Xmj through dimensionless processing, forming a quantitative and comparable regional heat source characteristic index; in the screening stage, based on the statistical mean algorithm, the heat source density coefficient Xmj is analyzed globally to effectively identify abnormal areas where the heat source density is significantly higher than the average level, and automatically label them as heat source aggregation functional areas, and send an analysis signal to the risk prediction module in a timely manner; the outstanding advantage of this mechanism is that it establishes a heat anomaly aggregation active screening mechanism centered on infrared sensing, which can not only identify hidden operations that are difficult to detect in unplanned operations, but also efficiently focus on suspected multi-person collaborative behaviors and abnormal heating of power equipment, truly realizing the quantitative identification and risk focusing of heat anomalies in the operation space, and improving the risk early warning preposition ability of the substation in multi-person and high-density operation scenarios.

[0035] Embodiment 5 Please refer to Figure 1 , specifically: The risk prediction module includes a behavior risk analysis unit, a perturbation analysis unit, and a prediction unit; The behavior risk analysis unit is used to monitor the behavior status of unplanned operation personnel in the heat source aggregation functional area during the personnel activity period after receiving the risk prediction signal, according to the high-definition optical camera and RFID reader integrated in the inspection robot, and obtain relevant personnel behavior status data information. Among them, the relevant personnel behavior status data information includes the number of times of climbing Ccs, the climbing duration Tsc, and the proportion of safety-wearing personnel Bcd in the heat source aggregation functional area; It should be noted that the number of times of climbing height Ccs represents the number of times that personnel's climbing behavior is detected by the system through the high-definition optical camera integrated in the inspection robot in combination with the human body posture recognition algorithm during the monitoring period; the climbing duration Tsc is the cumulative time that personnel are in the climbing state in the heat source aggregation functional area, which can be calculated by combining continuous frame recognition and a timer; the proportion of personnel wearing safety equipment Bcd refers to the proportion of the number of personnel wearing complete safety equipment and confirmed through visual recognition and protective label reading among the personnel working in the current area; the above indicators are all obtained through real-time analysis by the robot edge computing unit and uploaded to the central data platform as the core input parameters for judging the risk level of operation behavior; By performing feature recognition on the obtained data information on the behavior states of relevant personnel, the number of times of climbing height Ccs, the climbing duration Tsc, and the proportion of personnel wearing safety equipment Bcd in the heat source aggregation functional area are correlated. After dimensionless processing, the risk degree of the behavior of unplanned operation personnel in the heat source aggregation functional area is analyzed to obtain the behavior risk coefficient Xxw, which is specifically calculated through the following formula; ; In the formula, 、 and respectively represent the weight values of the number of times of climbing height Ccs, the climbing duration Tsc, and the proportion of personnel wearing safety equipment Bcd.

[0036] It should be noted that the formula for calculating the behavior risk coefficient Xxw quantifies and evaluates the specific behavior states of unplanned operation personnel in the heat source aggregation functional area, and combines the high-risk behavior indicators of the number of times of climbing height Ccs, the climbing duration Tsc, and the proportion of personnel wearing safety equipment Bcd of the operation personnel through weighting to form a comprehensive risk parameter reflecting the danger degree of operation behavior; it is closely related to the deficiencies proposed in the background technology and is designed to address the problems that the traditional monitoring system is difficult to identify the risk degree of operation behavior and lacks the deep perception ability of operation actions and violation states; the role of the behavior risk coefficient Xxw is to provide a quantitative basis for the danger level assessment of the behavior of operation personnel in unplanned operations and is one of the core input parameters in the risk prediction module for determining whether to trigger an early warning response; Specifically, the perturbation analysis unit is used to monitor the bus voltage status of the heat source aggregation functional area in real time according to the voltage sensors deployed on the bus nodes during the personnel activity period, and obtain relevant voltage status data information. Among them, the relevant voltage status data information includes the voltage value Vss at each monitoring time point in the heat source aggregation functional area. By extracting features from the historical data stored in the central data platform and combining the statistical mean algorithm, the historical voltage mean value Vjz of the heat source aggregation functional area is obtained. By correlating the voltage value Vss at each monitoring time point in the heat source aggregation functional area with the historical voltage mean value Vjz, after dimensionless processing, the bus voltage perturbation situation in the heat source aggregation functional area during the personnel activity period is analyzed, and the bus voltage perturbation coefficient Xrd is obtained, which is specifically calculated through the following formula; ; In the formula, represents the voltage value at the j-th monitoring time point, j = 1, 2, 3,..., n, and n represents the number of monitoring time points during the personnel activity period, represents the degree of fluctuation of the voltage value Vss at the j-th monitoring time point relative to the historical voltage mean value Vjz.

[0037] It should be noted that the voltage value Vss at each monitoring time point in the heat source aggregation functional area refers to the instantaneous bus voltage value collected in real time at a fixed time interval by the voltage sensors deployed on the bus nodes in this area during the actual operation period of the personnel, which reflects the operation state fluctuation of the electrical system under the intervention of the operation behavior; while the historical voltage mean value Vjz is obtained by extracting statistical features and calculating the mean value from the voltage historical data sequence of this functional area stored in the central data platform under non-operation interference conditions, which represents the steady-state voltage reference value of the regional bus under normal operation conditions; the acquisition methods of the two complement each other. The former relies on the online monitoring ability and time synchronization mechanism of the sensor, and the latter depends on the historical data mining and statistical modeling ability of the big data platform; by correlating the voltage value Vss at each monitoring time point in the heat source aggregation functional area with the historical voltage mean value Vjz, it is identified whether the unplanned operation behavior has actually disturbed the voltage stability in the area, so as to provide a reliable perception basis for the power operation side for subsequent calculation of the voltage perturbation coefficient and implementation of risk prediction; The formula in the perturbation analysis unit is used to calculate the bus voltage perturbation coefficient Xrd, which normalizes and evaluates the relative fluctuation degree between the voltage value Vss at each monitoring time point in the heat source aggregation functional area and the historical voltage average value Vjz during the occurrence of the operation behavior. By traversing all monitoring time points, this formula processes the relative deviation of the voltage at each moment in an average form, reflecting the overall level of voltage perturbation during the operation. It is mainly used to solve the problems that it is difficult for traditional monitoring systems to associate the causal relationship between operation behavior and power grid operation perturbation, and it is difficult to quantitatively reflect the impact of operation behavior on the stability of the electrical system. By introducing the bus voltage perturbation coefficient Xrd, it effectively realizes the real-time quantitative judgment of whether unplanned operation behavior causes voltage fluctuation, and establishes an accurate linkage criterion between the operation state on the electrical side and the operation behavior. The role of the bus voltage perturbation coefficient lies not only in being used as a key system perturbation index input in risk prediction, but also in providing an objective criterion based on power operation data for identifying high-risk operation behaviors, realizing the cross-layer linkage from behavior identification to electrical perturbation analysis, and effectively enhancing the system's perception ability of the depth of operation impact and risk level.

[0038] Specifically, the prediction unit is used to correlate the heat source density coefficient Xmj of each functional area in the heat source aggregation functional area with the behavior risk coefficient Xxw and the bus voltage perturbation coefficient Xrd of the heat source aggregation functional area. After dimensionless processing, it predicts and analyzes the degree of unplanned operation risk in the heat source aggregation functional area to obtain the unplanned operation risk prediction index Zyc of the heat source aggregation functional area, which is specifically obtained through the following formula: ; In the formula, represents the heat source density coefficient of the z-th functional area in the heat source aggregation functional area, z = 1, 2, 3,..., m, and m represents the number of functional areas in the heat source aggregation functional area. represents the behavior risk coefficient of the heat source aggregation functional area. represents the bus voltage perturbation coefficient of the heat source aggregation functional area, where , and all represent weight values, and R represents the second correction constant.

[0039] It should be noted that the formula in the prediction unit is used to calculate the unplanned operation risk prediction index Zyc. Its core purpose is to perform weighted fusion on the heat source density coefficient Xmj of each functional area in the heat source aggregation functional area, the behavior risk coefficient Xxw of the heat source aggregation functional area, and the bus voltage disturbance coefficient Xrd, and construct a composite prediction index reflecting the overall risk level of unplanned operations; this formula closely corresponds to the prominent deficiencies of the traditional system pointed out in the background technology in terms of lagging unplanned operation risk assessment, untimely response, lack of multi-source data fusion, and risk quantification model; by introducing the unplanned operation risk prediction index Zyc, the spatial heat source anomaly, the degree of behavior violation, and the electrical disturbance level are organically unified, and a linkage quantification logic between behavior, heat source, and power grid status is established, effectively realizing the closed-loop analysis of unplanned operation behavior from identification to risk prediction; specifically, the calculation result of the unplanned operation risk prediction index Zyc can be used to trigger the hierarchical response mechanism of the early warning module and is an important basis for judging whether to execute the first-level early warning and the second-level reminder; it provides a unified judgment quantity integrating multi-dimensional risk factors for the system, not only effectively solving the problems of scattered perception and isolated judgment of the traditional system, but also significantly enhancing the ability to identify and intervene in high-risk unplanned operation behaviors in advance, and improving the intelligence and effectiveness of substation operation management.

[0040] In this embodiment, through the collaborative linkage of the behavior risk analysis unit, the disturbance analysis unit, and the prediction unit, an intelligent risk quantification mechanism integrating personnel behavior perception, electrical disturbance analysis, and multi-source information fusion is constructed, realizing the forward-looking modeling and active prediction of the risks caused by unplanned operation behaviors; among them, the behavior risk analysis unit, based on the high-definition video and RFID identification information collected by the inspection robot, real-time obtains the climbing behavior, climbing duration, and safety protection status of the operating personnel in the heat source aggregation area, and constructs the behavior risk coefficient Xxw through weighted modeling and dimensionless processing, accurately reflecting the degree of violation of the current personnel behavior and the potential operation danger; the disturbance analysis unit, with the help of the voltage sensors deployed at the bus nodes, continuously obtains the voltage status Vss during the operation, and dynamically compares its real-time fluctuation characteristics with the historical average value Vjz to quantitatively generate the bus voltage disturbance coefficient Xrd, enabling the system to have the real-time judgment ability on whether the unplanned operation has disturbed the electrical operation; the prediction unit performs multi-dimensional fusion on the heat source density Xmj, the behavior risk coefficient Xxw, and the voltage disturbance coefficient Xrd to form the unplanned operation risk prediction index Zyc, realizing the horizontal risk assessment from "behavior" to "equipment" and the vertical level output; improving the comprehensive risk prediction ability and risk disposal efficiency of the substation for unplanned operations.

[0041] Embodiment 6 Please refer to Figure 1, specifically: The early warning module is used to compare and analyze the unplanned operation risk prediction index Zyc of the heat source aggregation functional area with a preset risk prediction threshold Y to determine whether there is a risk in the unplanned operation currently carried out in the heat source aggregation functional area, so as to obtain a risk early warning instruction of the corresponding level and execute it. The specific content is as follows: If the unplanned operation risk prediction index Zyc of the heat source aggregation functional area ≥ the risk prediction threshold Y, it means that there is a risk in the unplanned operation currently carried out in the heat source aggregation functional area. Generate a first-level risk early warning instruction and execute it. The execution content is: Guide the inspection robot to move to the heat source aggregation functional area, and emit a sound and light alarm through the integrated sound and light alarm, and at the same time play the voice prompt of "There is a risk in the current unplanned operation, please stop the operation immediately"; Link the access control system to prohibit unplanned operation personnel from entering again, and notify the power supervisor to verify the on-site situation; If the unplanned operation risk prediction index Zyc of the heat source aggregation functional area < the risk prediction threshold Y, it means that there is no risk in the unplanned operation currently carried out in the heat source aggregation functional area. Generate a second-level risk early warning instruction and execute it. The execution content is: Guide the inspection robot to move to the heat source aggregation functional area to play the voice prompt of "The current operation behavior is unauthorized, please pay attention to operation safety", and continuously monitor the unplanned operation behavior in the heat source aggregation functional area. At the same time, mark the current operation event as an unplanned operation behavior without risk and record it in the central data platform.

[0042] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence, characterized in that: It includes a pre - acquisition module, a plan determination module, a heat source area screening module, a risk prediction module, and an early warning module; The pre - acquisition module is used to construct several functional areas according to the monitoring requirements of substation on - site operations, and pre - collect the activity time of operating personnel in the substation on - site area and the planned operation authorized entry and exit time; The plan determination module is used to determine unplanned operating personnel in the current substation on - site area and send out a heat source analysis instruction; Based on the received heat source analysis instruction, the heat source area screening module screens out the heat source - gathering functional areas in the substation on - site area and sends out a risk prediction signal; The risk prediction module is used to predict and analyze the risk level of unplanned operations in the heat source - gathering functional areas after receiving the risk prediction signal; The early warning module is used to determine whether there is a risk in the unplanned operations carried out in the current heat source - gathering functional area, so as to obtain a risk early - warning instruction of the corresponding level and execute it.

2. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 1, characterized in that: The pre - acquisition module includes a deployment unit and a time - period acquisition unit; The deployment unit is used to divide the substation on - site area according to the monitoring requirements of substation on - site operations, combine the substation GIS structure diagram and the functional use of the substation on - site area, construct several functional areas, use the inspection robot integrated with multiple groups of sensors to conduct all - day inspections on the constructed functional areas, deploy voltage sensors on the bus nodes of the substation on - site area, and deploy an access control system at the substation entrance and exit. Combining with wireless communication technology, upload relevant data information to the central data platform; The functional use of the substation on - site area includes the high - voltage main equipment operation area, the secondary control area, the access control supervision area, the tool storage area, and the computer room area; multiple groups of sensors include high - definition optical cameras, infrared thermal imagers, timers, audible and visual alarms, edge computing modules, and wireless communication modules; the access control system is used to store the planned operation authorized entry and exit time of the substation and verify the identity of the entering and exiting personnel; the central data platform is used to pre - process, perform AI edge computing, and store the uploaded relevant data information.

3. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 2, characterized in that: The time - period acquisition unit is used to monitor the activity time status of operating personnel in the substation on - site area according to the high - definition optical camera and timer integrated in the inspection robot, and combine with the human body recognition algorithm to obtain the set of personnel activity time periods Tj, and collect the planned operation authorized entry and exit time stored in the access control system according to the access control system deployed at the substation entrance and exit to obtain the set of authorized operation entry and exit time periods Tc, and upload the obtained set of personnel activity time periods Tj and the set of authorized operation entry and exit time periods Tc to the central data platform for storage.

4. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 3, characterized in that: The plan determination module includes a coupling - degree analysis unit and a determination unit; The coupling - degree analysis unit is used to perform feature recognition on the relevant data information of the central data platform, correlate the extracted set of personnel activity time periods Tj and the set of authorized operation entry and exit time periods Tc, analyze the coupling degree between the activity time period of operating personnel in the substation on - site area and the entry time period of authorized operating personnel, and obtain the time coupling - degree coefficient Xoh, which is specifically calculated through the following formula; ； Wherein, represents the intersection period of the set Tj of personnel activity time periods and the set Tc of authorized operation entry and exit time periods; represents the union period of the set Tj of personnel activity time periods and the set Tc of authorized operation entry and exit time periods, and A represents a first correction constant; The determination unit is used to determine whether the operating personnel in the substation site area are unauthorized unplanned operating personnel after comparative analysis according to the obtained value of the time coupling coefficient Xoh. The specific analysis process is as follows; If 0 ≤ time coupling coefficient Xoh < 1, it means that the operating personnel in the current substation site area are unauthorized unplanned operating personnel, and at this time, a heat source analysis instruction is sent outwards; If the time coupling coefficient Xoh = 1, it means that the operating personnel in the current substation site area are not unauthorized unplanned operating personnel, and at this time, no additional heat source analysis instruction is sent.

5. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 4, characterized in that: The heat source area screening module includes a heat source analysis unit and a screening unit; The heat source analysis unit is used to, after receiving the heat source analysis instruction, based on a number of constructed functional areas, and according to the infrared thermal imager integrated in the inspection robot, take infrared images of the shooting areas of each functional area, use the temperature threshold segmentation algorithm to extract the heat source target area, map the extracted heat source target area into the pseudo-color image space, construct the color scale map of each functional area, perform pixel traversal analysis on the red channel in the color scale map, and use the local extreme value extraction algorithm to identify the maximum color scale value of each functional area to obtain the heat source area color scale peak Ssj of each functional area; based on the number of pixels in the color scale map of each functional area, and use the physical area conversion model to convert it into the actual heat source area to obtain the total heat source area Smj of each functional area, and through the connected component analysis method, number and identify different heat source areas in each functional area to obtain the number Nss of independent heat source areas in each functional area; construct the relevant heat source status data information with the heat source area color scale peak Ssj, the total heat source area Smj and the number Nss of independent heat source areas in each functional area; According to the dimensions of each functional area in the substation CAD design drawing, obtain the area Sqy of each functional area, associate it with the relevant heat source status data information, and after dimensionless processing, analyze the heat source distribution of each functional area to obtain the heat source density coefficient Xmj of each functional area, which is specifically calculated by the following formula; ； wherein, represents the heat source density coefficient of the i-th functional area, represents the area of the i-th functional area, represents the peak value of the heat source area color scale of the i-th functional area, represents the total area of the heat source area of the i-th functional area, represents the number of independent heat source areas of the i-th functional area, where , and all represent weight values.

6. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 5, characterized in that: The screening unit calculates the mean value of the heat source density coefficients Xmj of each functional area in the substation site area based on the obtained heat source density coefficients Xmj of each functional area and according to the statistical mean algorithm, so as to obtain the average heat source density coefficient of the substation site area ; The heat source density coefficients Xmj of each functional area in the substation site area are respectively compared with the average heat source density coefficient of the substation site area After comparative analysis, the heat source aggregation functional areas are screened out and a risk warning is sent out. The specific comparative analysis content is as follows: Respectively compare the heat source density coefficient Xmj of each functional area in the substation site area with the average heat source density coefficient of the substation site area for comparative analysis. If the heat source density coefficient Xmj of the current functional area exceeds the average heat source density coefficient , then classify the corresponding functional area as a heat source aggregation functional area, and at the same time mark it. At this time, send out a risk prediction signal. If the heat source density coefficient Xmj of the current functional area does not exceed the average heat source density coefficient , then classify the corresponding functional area as a normal heat source functional area, and no additional risk prediction signal is sent at this time.

7. The multi-dimensional monitoring system for unscheduled substation operations based on artificial intelligence according to claim 6, characterized in that: The risk prediction module includes a behavior risk analysis unit, a disturbance analysis unit and a prediction unit; The behavior risk analysis unit is used to, after receiving the risk prediction signal, monitor the behavior status of unplanned operating personnel in the heat source aggregation functional area during the personnel activity period according to the high-definition optical camera and RFID reader integrated in the inspection robot, and obtain the relevant personnel behavior status data information. Among them, the relevant personnel behavior status data information includes the number of times Ccs of personnel climbing heights, the climbing duration Tsc and the proportion Bcd of personnel wearing safety equipment in the heat source aggregation functional area; By performing feature recognition on the obtained relevant personnel behavior status data information, associate the number of times Ccs of personnel climbing heights, the climbing duration Tsc and the proportion Bcd of personnel wearing safety equipment in the heat source aggregation functional area, and after dimensionless processing, analyze the behavior risk degree of unplanned operating personnel in the heat source aggregation functional area to obtain the behavior risk coefficient Xxw.

8. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 7, characterized in that: The disturbance analysis unit is used to monitor the bus voltage status of the heat source aggregation functional area during the personnel activity period in real time according to the voltage sensors deployed on the bus nodes, and obtain relevant voltage status data information. Among them, the relevant voltage status data information includes the voltage value Vss at each monitoring time point in the heat source aggregation functional area. By extracting features from the historical data stored in the central data platform and combining the statistical mean algorithm, the historical voltage mean value Vjz of the heat source aggregation functional area is obtained. By correlating the voltage value Vss at each monitoring time point in the heat source aggregation functional area with the historical voltage mean value Vjz, after dimensionless processing, the bus voltage disturbance situation in the heat source aggregation functional area during the personnel activity period is analyzed, and the bus voltage disturbance coefficient Xrd is obtained, which is specifically calculated by the following formula; ； In the formula, represents the voltage value at the j-th monitoring time point, where j = 1, 2, 3, ..., n, and n represents the number of monitoring time points within the personnel activity time period.

9. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 8, wherein: The prediction unit is used to correlate the heat source density coefficient Xmj of each functional area in the heat source aggregation functional area with the behavior risk coefficient Xxw and the bus voltage disturbance coefficient Xrd of the heat source aggregation functional area. After dimensionless processing, the risk degree of unplanned operations in the heat source aggregation functional area is predicted and analyzed, and the unplanned operation risk prediction index Zyc of the heat source aggregation functional area is obtained, which is specifically obtained by the following formula: ； In the formula, represents the heat source density coefficient of the z-th functional area regarded as the heat source aggregation functional area, where z = 1, 2, 3, ..., m, and m represents the number of functional areas of the heat source aggregation functional area. represents the behavioral risk coefficient of the heat source aggregation functional area. represents the bus voltage disturbance coefficient of the heat source aggregation functional area, where , and all represent weight values, and R represents the second correction constant.

10. The multi-dimensional monitoring system for unplanned substation operations based on artificial intelligence according to claim 9, characterized in that: The warning module is used to compare and analyze the unplanned operation risk prediction index Zyc of the heat source aggregation functional area with the preset risk prediction threshold Y to determine whether there is a risk in the unplanned operation currently carried out in the heat source aggregation functional area, so as to obtain a risk warning instruction of the corresponding level and execute it. The specific content is as follows: If the unplanned operation risk prediction index Zyc of the heat source aggregation functional area ≥ the risk prediction threshold Y, it means that there is a risk in the unplanned operation currently carried out in the heat source aggregation functional area, and a first-level risk warning instruction is generated and executed. The execution content is: guiding the inspection robot to move to the heat source aggregation functional area, and emitting a sound and light alarm through the integrated sound and light alarm, and at the same time playing the voice prompt of "There is a risk in the current unplanned operation, please stop the operation immediately"; linking the access control system to prohibit unplanned operation personnel from entering again, and notifying the power supervisor to check the on-site situation; If the unplanned operation risk prediction index Zyc of the heat source aggregation functional area < the risk prediction threshold Y, it means that there is no risk in the unplanned operation currently carried out in the heat source aggregation functional area, and a second-level risk warning instruction is generated and executed. The execution content is: guiding the inspection robot to move to the heat source aggregation functional area to play the voice prompt of "The current operation behavior is unauthorized, please pay attention to operation safety", and continuously monitoring the unplanned operation behavior in the heat source aggregation functional area. At the same time, mark the current operation event as an unplanned operation behavior without risk and record it in the central data platform.

Images