Rural power grid transformer anti-theft monitoring and alarming system and method based on multi-source information fusion

Abstract

The invention provides a rural power grid transformer anti-theft monitoring and alarm system and method based on multi-source information fusion, and relates to the technical field of power systems, and the system comprises a multi-source heterogeneous sensor module, a data preprocessing module, an information fusion decision module, a communication positioning module and an alarm execution module. Collecting multi-dimensional monitoring data of the transformer body and the periphery; performing noise filtering and standardization processing on the multi-dimensional monitoring data; carrying out collaborative verification on the preprocessed data; according to the method, multi-source heterogeneous data such as vibration, displacement, sound and images are fused, environmental noise is filtered through Kalman filtering, cooperative verification is carried out through the D-S evidence theory, the false alarm rate and the missing report rate are lower, man-made theft, environmental interference and animal touch can be accurately distinguished, manpower consumption caused by invalid alarm is avoided, and meanwhile the behavior of missing real theft is completely eradicated.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a rural power grid transformer anti-theft monitoring and alarm system and method based on multi-source information fusion. Background Technology

[0002] Rural power grid transformers are core equipment for rural power transmission, widely distributed in plains, mountains, and suburbs. They are characterized by dispersed installation points, complex surrounding environments, and the difficulty of manual inspection. In recent years, theft of rural power grid transformers has been frequent, causing not only direct economic losses but also regional power outages, affecting agricultural production and residents' lives. Traditional anti-theft methods rely heavily on manual inspections or single-sensor monitoring. Manual inspections are limited by long cycles and limited coverage, making real-time protection difficult. Single sensors are susceptible to environmental interference; for example, vibrations of the transformer casing caused by wind and rain, electromagnetic interference from lightning, and contact with wild animals can all trigger false alarms. Statistics show that the false alarm rate of existing systems is as high as 15%-30%, and the sensitivity to minor theft is insufficient, with a false alarm rate exceeding 10%, seriously affecting the reliability of the anti-theft system.

[0003] In existing technologies, some anti-theft systems attempt to introduce two types of sensors for collaborative monitoring, but they lack an effective information fusion mechanism. They rely solely on "OR logic" to determine alarms, failing to distinguish between interference signals and actual theft. Furthermore, most systems use GPRS communication modules, resulting in low data transmission rates, alarm information upload delays exceeding 30 seconds, and a lack of precise location capabilities. This makes it difficult for security personnel to quickly reach the scene of the theft, leading to untimely intervention and the theft of core transformer components, further exacerbating losses. Simultaneously, existing systems lack a tiered alarm mechanism, only providing on-site audible and visual alarms, offering limited deterrence and failing to preserve evidence of theft, hindering subsequent case investigation. Summary of the Invention

[0004] To address the aforementioned problems, this invention proposes a rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion. This invention solves the problems of high false alarm, missed alarm, and delayed alarm rates in existing technologies.

[0005] Another objective of this invention is to provide a method for monitoring and alarming theft prevention of rural power grid transformers based on multi-source information fusion.

[0006] To achieve the objectives of this invention, the invention is implemented through the following technical solutions:

[0007] A rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion includes:

[0008] Multi-source heterogeneous sensor modules are used to collect multi-dimensional monitoring data of the transformer body and its surroundings;

[0009] The information fusion decision module uses a multi-source information fusion algorithm to collaboratively verify the preprocessed data and determine whether theft has occurred.

[0010] The communication and positioning module is used to upload the judgment results and transformer positioning information to the monitoring center and the terminal of the power protection personnel in real time.

[0011] The alarm execution module initiates alarm operations of the corresponding level based on the judgment result.

[0012] Further improvements include: the multi-source heterogeneous sensor module includes a vibration sensor, a displacement sensor, a sound sensor, an image sensor, an electrical parameter sensor, and a tilt sensor; the vibration sensor is used to monitor the amplitude and duration of vibration of the transformer casing; the displacement sensor is used to monitor the positional changes of the transformer terminals or casing; the electrical parameter sensor is used to monitor the changes in current and voltage at the transformer input terminals; the image sensor is used to acquire image information around the transformer; and the tilt sensor is used to monitor the overall tilt angle of the transformer.

[0013] Further improvements include a data preprocessing module that receives multi-dimensional monitoring data and performs noise filtering and standardization.

[0014] The data preprocessing module uses a Kalman filter algorithm to remove noise. The filtering formula is as follows:

[0015] X k =A k X k-1 +B k U k +W k ,

[0016] Z k =H k Xk+Vk,

[0017] Among them, X k Let X be the estimated state of the system at time k. k-1 Let A be the estimated state of the system at time k-1. k Let B be the state transition matrix at time k. k U is the control input matrix at time k. k W is the input quantity to control at time k. k Z represents the process noise at time k. k Let H be the sensor observation at time k. k Let V be the observation matrix at time k. k The noise observed at time k is denoted as .

[0018] A further improvement is that the data preprocessing module also employs a data normalization algorithm, the formula of which is:

[0019]

[0020] Where x′ represents the normalized data, with a value range of [0,1]; x represents the original sensor data; x min x is the minimum value of data collected by this type of sensor. max This represents the maximum value of the data collected by this type of sensor.

[0021] A further improvement lies in the fact that the information fusion decision module uses DS evidence theory to achieve multi-source data fusion, and the basic probability allocation formula is:

[0022]

[0023] in,

[0024] m(A) represents the basic probability of proposition A, indicating the degree of confidence that A is true; A is an evidentiary proposition, including suspected theft, confirmed theft, and environmental interference; m i (A i Let be the i-th sensor pair for proposition A. i The basic probability; K is the conflict coefficient, used to measure the degree of conflict among multiple sources of evidence, K∈[0,1]); Ω is the identification frame, the set of all possible propositions.

[0025] A further improvement is made in that the judgment logic of the information fusion decision module is as follows: when the vibration sensor detects a vibration with an amplitude of more than 0.5g and lasts for more than 10 seconds, and the displacement sensor detects a position change of more than 2mm, or the electrical parameter sensor detects a sudden drop in current of more than 90%, it is judged as a suspected theft; when the image sensor simultaneously recognizes a person approaching and tool operation, it is judged as a confirmed theft.

[0026] Further improvements include: the alarm execution module includes an audible and visual alarm, an SMS push module, and a video capture unit; the alarm execution module provides tiered alarms including: triggering a suspected theft to activate a level one alarm: on-site audible and visual alarm; triggering a confirmed theft to activate a level two alarm: on-site audible and visual alarm + SMS push to management personnel + video capture.

[0027] The method for anti-theft monitoring and alarm of rural power grid transformers based on multi-source information fusion includes the following steps:

[0028] Collect multi-dimensional monitoring data of the transformer body and its surroundings;

[0029] Collaborative verification of multi-dimensional monitoring is used to determine whether theft has occurred.

[0030] The judgment results and transformer location information are uploaded to the monitoring center and the terminal of the power protection personnel in real time.

[0031] Based on the judgment result, the corresponding alarm operation will be initiated.

[0032] The beneficial effects of this invention are as follows:

[0033] 1. This invention breaks through the limitations of single-sensor monitoring, integrates multi-source heterogeneous data such as vibration, displacement, sound, and images, filters out environmental noise through Kalman filtering, and then conducts collaborative verification through DS evidence theory. Compared with existing systems, this invention has a lower false alarm rate and false alarm rate, can accurately distinguish between human theft, environmental interference and animal touch, avoids invalid alarms that waste manpower, and at the same time prevents the omission of real theft.

[0034] 2. This invention uses a 4G / 5G wireless communication module, which has a small delay in uploading alarm information. Combined with GPS positioning function, the positioning error is less than 10 meters. Power protection personnel can obtain the transformer location and on-site captured images in real time through a mobile APP, quickly plan routes to reach the site, and have a shorter average response time, effectively preventing theft and reducing equipment damage and power outage time.

[0035] 3. This invention adopts a multi-level alarm mechanism. The first-level alarm uses on-site sound and light to deter thieves, delaying or stopping the theft. The second-level alarm simultaneously sends text messages to management personnel and captures on-site images, ensuring timely intervention and preserving evidence of theft to support subsequent case investigation and solve the problems of weak deterrence and lack of evidence in existing systems. Attached Figure Description

[0036] Figure 1 This is a system configuration block diagram of the present invention.

[0037] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0038] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0039] Example 1

[0040] according to Figure 1 As shown in the figure, this embodiment proposes a rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion, including a multi-source heterogeneous sensor module, a data preprocessing module, an information fusion decision module, a communication and positioning module, and an alarm execution module. The multi-source heterogeneous sensor module is used to collect multi-dimensional monitoring data of the transformer body and its surroundings; the data preprocessing module receives the multi-dimensional monitoring data and performs noise filtering and standardization processing; the information fusion decision module uses a multi-source information fusion algorithm to collaboratively verify the preprocessed data and determine whether theft has occurred.

[0041] The communication and positioning module is used to upload the judgment results and transformer location information to the monitoring center and the terminal of the power protection personnel in real time; the alarm execution module initiates the corresponding level of alarm operation according to the judgment results. Through multi-module collaboration, a closed-loop management system is achieved from data acquisition to alarm execution, overcoming the "monitoring-decision-alarm" disconnect problem caused by the lack of module integration in traditional systems, and improving the overall reliability of the anti-theft system; the combination of multi-dimensional data acquisition and intelligent decision-making avoids the limitations of single-module functions, providing comprehensive technical support for transformer anti-theft in different scenarios.

[0042] The multi-source heterogeneous sensor module includes vibration sensors, displacement sensors, sound sensors, image sensors, electrical parameter sensors, and tilt sensors. The vibration sensors monitor the amplitude and duration of transformer casing vibration; the displacement sensors monitor changes in the position of transformer terminals or casing; the electrical parameter sensors monitor changes in current and voltage at the transformer's input terminals; the image sensors acquire images of the transformer's surroundings; and the tilt sensors monitor the overall tilt angle of the transformer. These multiple sensor types cover key dimensions of "vibration-displacement-electrical parameters-image-tilt," solving the blind spots of traditional single-sensor monitoring and comprehensively capturing multi-feature signals of theft. Targeted design of each sensor's monitoring objectives (such as terminal displacement and input terminal electrical parameters) precisely focuses on vulnerable parts of the transformer, improving the early detection capability of theft.

[0043] The image sensor supports color imaging during the day and black-and-white imaging at night, with nighttime imaging illumination not exceeding 0.001 Lux. It also features facial recognition. The electrical parameter sensor monitors current and voltage, with current drop trigger thresholds above 90% and voltage drop trigger thresholds above 80%. The 0.001 Lux low-light imaging capability eliminates blind spots in nighttime anti-theft monitoring, ensuring 24-hour uninterrupted monitoring in unlit areas such as plains and mountains. Facial recognition can identify thieves, providing crucial clues for case investigation and enhancing the post-incident traceability value of the anti-theft system. Dual-parameter monitoring of current and voltage can distinguish between "theft-induced power outages" (simultaneous current and voltage drops) and "normal power outages" (voltage drops first, current drops later), reducing false alarms. Targeted drop threshold settings (90% current, 80% voltage) can filter out interference from temporary power outages due to construction in suburban areas, ensuring that only genuine theft-induced power outage alarms are triggered.

[0044] The data preprocessing module uses a Kalman filter algorithm to remove noise. The filtering formula is as follows:

[0045] X k =A k X k-1 +B k U k +W k,

[0046] Z k =H k Xk+Vk,

[0047] Among them, X k Let X be the estimated state of the system at time k. k-1 Let A be the estimated state of the system at time k-1. k Let B be the state transition matrix at time k. k U is the control input matrix at time k. k W is the input quantity to control at time k. k Z represents the process noise at time k. k Let H be the sensor observation at time k. k Let V be the observation matrix at time k. k Let k be the observation noise at time k. By quantizing process noise and observation noise, environmental noise such as wind, rain, vibration, lightning, and electromagnetic interference is effectively filtered out, avoiding distortion of raw data caused by noise and providing a reliable data foundation for subsequent decision-making; the system state estimate (X) is dynamically updated. k It can track sensor data changes in real time, improving its ability to detect slow theft behaviors (such as low-amplitude prying).

[0048] The data preprocessing module also employs a data normalization algorithm, the formula of which is:

[0049]

[0050] Where x′ represents the normalized data, with a value range of [0,1]; x represents the original sensor data; x min x is the minimum value of data collected by this type of sensor. max This represents the maximum value of the data collected by this type of sensor. Normalization eliminates the dimensional differences in vibration (g), displacement (mm), and current (A) data, solving the problem of the inability to directly fuse multi-source data; the data values ​​are unified to the [0,1] range, reducing the computational complexity of the information fusion decision module and improving the efficiency and accuracy of theft detection.

[0051] The information fusion decision module uses DS evidence theory to achieve multi-source data fusion, and the basic probability allocation formula is as follows:

[0052]

[0053] in,

[0054] m(A) represents the basic probability of proposition A, indicating the degree of confidence that A is true; A is an evidentiary proposition, including suspected theft, confirmed theft, and environmental interference; m i (A iLet be the i-th sensor pair for proposition A. i The basic probability (m(A)) is used to measure the degree of conflict among multiple sources of evidence (K∈[0,1]). Ω is the identification frame, which is the set of all possible propositions. Quantifying the degree of contradiction among multiple sources of evidence using the conflict coefficient (K) can effectively handle sensor data conflict scenarios (such as vibration triggering but no people in the image), avoiding misjudgments by traditional "OR logic". The trust update mechanism based on the basic probability (m(A)) can gradually strengthen the evidence weight of real theft behavior and improve the differentiation accuracy of "environmental interference-animal touch-human theft".

[0055] The decision-making logic of the information fusion module is as follows: when the vibration sensor detects a vibration with an amplitude exceeding 0.5g and lasting for more than 10 seconds, and the displacement sensor detects a positional change exceeding 2mm, or the electrical parameter sensor detects a sudden drop in current exceeding 90%, it is determined to be a suspected theft. When the image sensor simultaneously identifies personnel approaching and tool operation, it is determined to be a confirmed theft. The dual-path suspected theft determination of "vibration + displacement" and "electrical parameters" covers two types of theft modes: "physical damage" and "power outage theft," avoiding missed detections due to a single path. The dual confirmation condition of "personnel + tools" from the image sensor provides intuitive visual evidence of theft, eliminating false alarms triggered by non-human factors (such as animal collisions).

[0056] The communication and positioning module uses a 4G or 5G wireless communication module, and the positioning information is obtained through a GPS module, with a positioning error of less than 10 meters. Compared with traditional GPRS, 4G / 5G communication reduces the alarm delay to within 5 seconds, and the GPS positioning error is controlled within 10 meters, ensuring that power protection personnel can quickly locate the scene and reduce the time for theft to be carried out.

[0057] The alarm execution module includes an audible and visual alarm, an SMS push module, and a video capture unit. The sound pressure level of the audible and visual alarm is no less than 110dB, and the image resolution of the video capture unit is no less than 1080P. The alarm execution module's tiered alarm system includes: triggering a suspected theft to activate a Level 1 alarm: on-site audible and visual alarm; triggering a confirmed theft to activate a Level 2 alarm: on-site audible and visual alarm + SMS push to management personnel + video capture. The tiered alarm system both immediately deters thieves through on-site audible and visual means and achieves simultaneous "notification-evidence" through SMS and video, solving the problems of weak deterrence and lack of traceability in traditional systems. The 110dB high sound pressure level audible and visual alarm can effectively disperse thieves, preventing the dismantling and theft of core transformer components (such as copper coils), thus reducing economic losses; the 1080P high-definition video capture ensures that the crime process, tools, and the appearance of the perpetrators are clearly identifiable, providing a complete chain of evidence for judicial evidence collection and preventing evidence from becoming invalid.

[0058] Example 2

[0059] according to Figure 1As shown, this embodiment proposes a rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion, including its application in rural power grid transformers in plain areas:

[0060] Sensor layout:

[0061] Vibration sensor: A piezoelectric vibration sensor (model: YZ-801) is used, which is installed on the side wall of the transformer tank. The sampling frequency is 100Hz, the amplitude threshold is 0.5g, and the duration threshold is 10 seconds.

[0062] Displacement sensor: Laser displacement sensor (model: LD-602), installed at the high voltage terminal, measuring range 0-10mm, displacement threshold 2mm;

[0063] Electrical parameter sensor: Hall current sensor (model: HL-50A), connected in series at the transformer input terminal, sampling period of 1 second, current drop threshold of 90%;

[0064] Image sensor: High-definition network camera (model: IPC-625), installed on a 1.5-meter-high pole around the transformer, covering the transformer body, and supporting daytime color / nighttime black and white modes;

[0065] Tilt sensor: Dual-axis tilt sensor (model: QT-306), installed on the top of the transformer tank, with a tilt angle threshold of 5°.

[0066] Test and verification:

[0067] Ten 10kV rural power grid transformers in a plain area were selected and continuously tested for 30 days, simulating 100 scenarios (including 50 instances of human theft, 30 instances of environmental disturbance, and 20 instances of animal contact). The results are as follows:

[0068] Human theft: 49 correct identifications (1 missed detection was due to slight prying of the fuel tank, amplitude 0.45g), identification accuracy rate 98%;

[0069] Environmental interference: 29 cases were correctly eliminated (1 false alarm was caused by rainstorm, resulting in vibration lasting 12 seconds with an amplitude of 0.52g), with a false alarm rate of 3.3%;

[0070] Animal touch: All 20 attempts were correctly ruled out, with no false alarms;

[0071] Average alarm message upload time: 3.2 seconds; GPS positioning error: 8.5 meters.

[0072] Example 3

[0073] according to Figure 1 As shown, this embodiment proposes a rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion, including its application in mountainous rural power grid transformers:

[0074] Sensor optimization:

[0075] Communication module: Adopts 5G enhanced module (model: 5G-720) to solve the problem of weak signal in mountainous areas, with an upload speed of ≥10Mbps;

[0076] Tilt sensor: Improved sensitivity, tilt angle threshold reduced to 3° (adapting to minor impacts from animal collisions and landslides in mountainous areas);

[0077] Sound sensor: A new microphone array sound sensor (model: SM-401) has been added to monitor the knocking sound of burglary tools (frequency 200-500Hz) and help distinguish between animal sounds and tool sounds.

[0078] Test and verification:

[0079] Eight 10kV rural power grid transformers in mountainous areas were selected and continuously tested for 30 days, simulating 80 scenarios (including 40 instances of human theft, 25 instances of environmental disturbance, and 15 instances of animal contact). The results are as follows:

[0080] Human theft: 40 correct identifications in 40 attempts, with a 100% accuracy rate;

[0081] Environmental interference: 23 correct eliminations (2 false alarms were caused by wild boar collision sensors, which were eliminated after verification by the sound sensor that there was no tool knocking sound), with a false alarm rate of 8%;

[0082] Animal touch: All 15 attempts were correctly ruled out, with no false alarms;

[0083] Average alarm message upload time: 4.5 seconds; GPS positioning error: 9.2 meters.

[0084] Example 4

[0085] according to Figure 1 As shown, this embodiment proposes a rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion, including its application in rural power grid transformers in urban-rural fringe areas:

[0086] Sensor optimization:

[0087] Image sensor: Upgraded to a starlight-level camera (model: IPC-830), with a nighttime illumination of 0.001 Lux, and supports face recognition;

[0088] Electrical parameter sensor: Added voltage monitoring function, monitoring voltage drop threshold of 80% (adapting to temporary power outage interference during construction in urban and suburban areas);

[0089] Data sampling period: shortened to 0.5 seconds (to address the characteristics of high population mobility and rapid theft in urban and suburban areas).

[0090] Test and verification:

[0091] Twelve 10kV rural power grid transformers in suburban areas were selected and continuously tested for 30 days, simulating 120 scenarios (including 60 instances of human theft, 35 instances of environmental disturbance, and 25 instances of animal contact). The results are as follows:

[0092] Human theft: 58 correct identifications (2 missed detections were due to slow disassembly at night, which were subsequently recorded after video playback), with an identification accuracy rate of 96.7%;

[0093] Environmental interference: 34 cases were correctly resolved (1 false alarm was caused by a power outage during construction, resulting in a sudden current drop of 92%, which was resolved after voltage monitoring confirmed that no personnel were near), with a false alarm rate of 2.9%;

[0094] Animal touch: All 25 attempts were correctly ruled out, with no false alarms;

[0095] Average alarm message upload time: 2.8 seconds; GPS positioning error: 7.6 meters.

[0096] Validation data:

[0097]

[0098]

[0099] This rural power grid transformer anti-theft monitoring and alarm system, based on multi-source information fusion, breaks through the limitations of single-sensor monitoring. It integrates heterogeneous data from multiple sources, including vibration, displacement, sound, and images. Environmental noise is filtered out using Kalman filtering, and further verification is performed using DS evidence theory. Compared to existing systems, this invention has a lower false alarm rate and a lower false alarm rate. It can accurately distinguish between human theft, environmental interference, and animal contact, avoiding unnecessary alarms and wasting manpower, while also preventing the omission of genuine theft. Furthermore, it employs a 4G / 5G wireless communication module, resulting in low alarm information upload latency. Combined with GPS positioning, the positioning error is less than 10 meters. Power protection personnel can obtain the transformer location and on-site captured images in real time via a mobile app, quickly planning routes to reach the scene. The average response time is shorter, effectively stopping theft and reducing equipment damage and power outage duration. Simultaneously, a multi-level alarm mechanism is adopted. The first-level alarm uses on-site sound and light to deter thieves, delaying or stopping theft. The second-level alarm simultaneously pushes SMS messages to management personnel and captures on-site images, ensuring timely intervention and preserving evidence for subsequent case investigation, solving the problems of weak deterrence and lack of evidence in existing systems.

[0100] Example 5

[0101] like Figure 2As shown, the rural power grid transformer anti-theft monitoring and alarm method based on multi-source information fusion of the present invention, based on embodiment 1, includes the following steps: collecting multi-dimensional monitoring data of the transformer body and its surroundings; receiving the multi-dimensional monitoring data and performing noise filtering and standardization processing; performing collaborative verification on the pre-processed data to determine whether theft has occurred; uploading the determination result and transformer location information to the monitoring center and the terminal of the power protection personnel in real time; and initiating the corresponding level of alarm operation according to the determination result.

[0102] In the data noise filtering step, the Kalman filter algorithm is used to filter out noise, and the filtering formula is:

[0103] X k =A k X k-1 +B k U k +W k ,

[0104] Z k =H k Xk+Vk,

[0105] Among them, X k Let X be the estimated state of the system at time k. k-1 Let A be the estimated state of the system at time k-1. k Let B be the state transition matrix at time k. k U is the control input matrix at time k. k W is the input quantity to control at time k. k Z represents the process noise at time k. k Let H be the sensor observation at time k. k Let V be the observation matrix at time k. k The noise observed at time k is denoted as .

[0106] In the data standardization process, a data normalization algorithm is used to standardize the data, and the formula is:

[0107]

[0108] Where x′ represents the normalized data, with a value range of [0,1]; x represents the original sensor data; x min x is the minimum value of data collected by this type of sensor. max This represents the maximum value of the data collected by this type of sensor.

[0109] In the data collaborative verification step, DS evidence theory is used to achieve multi-source data fusion, and the basic probability allocation formula is as follows:

[0110]

[0111] in,

[0112] m(A) represents the basic probability of proposition A, indicating the degree of confidence that A is true; A is an evidentiary proposition, including suspected theft, confirmed theft, and environmental interference; m i (A i Let be the i-th sensor pair for proposition A. i The basic probability; K is the conflict coefficient, used to measure the degree of conflict among multiple sources of evidence, K∈[0,1]); Ω is the identification frame, the set of all possible propositions.

[0113] In the data collaborative verification step, the judgment logic is as follows: when a vibration with an amplitude exceeding 0.5g and lasting for more than 10 seconds is detected, and a positional change exceeding 2mm is detected, or a sudden drop in current exceeding 90% is detected, it is judged as a suspected theft; when personnel approach and tool operation are simultaneously detected, it is judged as a confirmed theft.

[0114] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion, characterized in that, include: Multi-source heterogeneous sensor modules are used to collect multi-dimensional monitoring data of the transformer body and its surroundings; The information fusion decision-making module collaboratively verifies multi-dimensional monitoring data to determine whether theft has occurred. The communication and positioning module is used to upload the judgment results and transformer positioning information to the monitoring center and the terminal of the power protection personnel in real time. The alarm execution module initiates alarm operations of the corresponding level based on the judgment result.

2. The rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion as described in claim 1, characterized in that: The multi-source heterogeneous sensor module includes a vibration sensor, a displacement sensor, a sound sensor, an image sensor, an electrical parameter sensor, and a tilt sensor. The vibration sensor is used to monitor the vibration amplitude and duration of the transformer casing. The displacement sensor is used to monitor the positional changes of the transformer terminals or casing. The electrical parameter sensor is used to monitor the changes in current and voltage at the transformer's input terminals. The image sensor is used to acquire image information around the transformer. The tilt sensor is used to monitor the overall tilt angle of the transformer.

3. The rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion as described in claim 1, characterized in that: It also includes a data preprocessing module, which receives multi-dimensional monitoring data and performs noise filtering and standardization; the data preprocessing module uses the Kalman filter algorithm to filter noise, and the filtering formula is: X k =A k X k-1 +B k U k +W k , Z k =H k Xk+Vk, Among them, X k Let X be the estimated state of the system at time k. k-1 Let A be the estimated state of the system at time k-1. k Let B be the state transition matrix at time k. k U is the control input matrix at time k. k W controls the input at time k. k Z represents the process noise at time k. k Let H be the sensor observation at time k. k Let V be the observation matrix at time k. k The noise observed at time k is denoted as .

4. The rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion according to claim 3, characterized in that: The data preprocessing module uses a data normalization algorithm to standardize the data, and the formula is: Where x′ represents the normalized data, with a value range of [0,1]; x represents the original sensor data; x min x is the minimum value of data collected by this type of sensor. max This represents the maximum value of the data collected by this type of sensor.

5. The rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion according to claim 1, characterized in that: The information fusion decision module uses DS evidence theory to achieve multi-source data fusion, and the basic probability allocation formula is as follows: in, m(A) represents the basic probability of proposition A, indicating the degree of confidence that A is true; A is an evidentiary proposition, including suspected theft, confirmed theft, and environmental interference; m i (A i Let be the i-th sensor pair for proposition A. i The basic probability; K is the conflict coefficient, used to measure the degree of conflict among multiple sources of evidence (K∈[0,1]); Ω is the identification frame, the set of all possible propositions.

6. The rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion according to claim 5, characterized in that: The judgment logic of the information fusion decision module is as follows: when the vibration sensor detects a vibration with an amplitude of more than 0.5g and lasts for more than 10 seconds, and the displacement sensor detects a position change of more than 2mm, or the electrical parameter sensor detects a sudden drop in current of more than 90%, it is judged as a suspected theft; when the image sensor simultaneously recognizes a person approaching and tool operation, it is judged as a confirmed theft.

7. The rural power grid transformer anti-theft monitoring and alarm system based on multi-source information fusion according to claim 1, characterized in that: The alarm execution module includes an audible and visual alarm, an SMS push module, and a video capture unit; the alarm execution module has tiered alarms, including: triggering a suspected theft to activate a level one alarm: on-site audible and visual alarm; triggering a confirmed theft to activate a level two alarm: on-site audible and visual alarm + SMS push to management personnel + video capture.

8. A method for anti-theft monitoring and alarm of rural power grid transformers based on multi-source information fusion, characterized in that, Includes the following steps: Collect multi-dimensional monitoring data of the transformer body and its surroundings; Collaborative verification of multi-dimensional monitoring is used to determine whether theft has occurred. The judgment results and transformer location information are uploaded to the monitoring center and the terminal of the power protection personnel in real time. Based on the judgment result, the corresponding alarm operation will be initiated.

9. The method for anti-theft monitoring and alarm of rural power grid transformers based on multi-source information fusion according to claim 8, characterized in that, It also includes noise filtering and standardization preprocessing steps for the collected multi-dimensional monitoring data. In the data noise filtering step, the Kalman filter algorithm is used to filter out noise, and the filtering formula is: X k =A k X k-1 +B k U k +W k , Z k =H k Xk+Vk, Among them, X k Let X be the estimated state of the system at time k. k-1 Let A be the estimated state of the system at time k-1. k Let B be the state transition matrix at time k. k U is the control input matrix at time k. k W controls the input at time k. k Z represents the process noise at time k. k Let H be the sensor observation at time k. k Let V be the observation matrix at time k. k The noise observed at time k is denoted as .

10. The method for anti-theft monitoring and alarm of rural power grid transformers based on multi-source information fusion according to claim 9, characterized in that, In the data standardization process, a data normalization algorithm is used to standardize the data, and the formula is: Where x′ represents the normalized data, with a value range of [0,1]; x represents the original sensor data; x min x is the minimum value of data collected by this type of sensor. max This represents the maximum value of the data collected by this type of sensor.

11. The method for anti-theft monitoring and alarm of rural power grid transformers based on multi-source information fusion according to claim 8, characterized in that, In the data collaborative verification step, DS evidence theory is used to achieve multi-source data fusion, and the basic probability allocation formula is as follows: in, m(A) represents the basic probability of proposition A, indicating the degree of confidence that A is true; A is an evidentiary proposition, including suspected theft, confirmed theft, and environmental interference; m i (A i Let be the i-th sensor pair for proposition A. i The basic probability; K is the conflict coefficient, used to measure the degree of conflict among multiple sources of evidence (K∈[0,1]); Ω is the identification frame, the set of all possible propositions.

12. The method for anti-theft monitoring and alarm of rural power grid transformers based on multi-source information fusion according to claim 11, characterized in that, In the data collaborative verification step, the judgment logic is as follows: when a vibration with an amplitude exceeding 0.5g and lasting for more than 10 seconds is detected, and a positional change exceeding 2mm is detected, or a sudden drop in current exceeding 90% is detected, it is judged as a suspected theft; when personnel approach and tool operation are simultaneously detected, it is judged as a confirmed theft.

Images