Video monitoring based filling station control method and system

Abstract

The application discloses a filling station control method and system based on video monitoring, and particularly relates to the technical field of industrial process control, and is used for solving the problem of logical dead loop caused by video verification failure due to equipment shutdown in the existing filling station control; a standby judgment signal is generated when the sensor is abnormally overlimited, a thermodynamic entropy increase trajectory is generated based on real-time video stream calculation of fault area pixel entropy value, when the entropy increase slope is over the threshold value, a standby power supply independent power supply video system is started, the quantum state fidelity attenuation characteristics of video light quantum distribution and the chaos index of fluid motion are synchronously analyzed under the standby power supply, when the fidelity attenuation conforms to the quantum effect characteristic curve and the chaos index is over the critical threshold value, a shutdown instruction is generated, otherwise, a main power supply recovery instruction is generated, and finally the corresponding instruction is executed; the method realizes decoupling of fault verification and equipment shutdown, avoids the risk of secondary accidents, and improves fault discrimination accuracy through a physical characteristic double verification mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial process control, and in particular to a filling station control method and system based on video monitoring. BACKGROUND

[0002] In the automation control of mine filling stations, video monitoring systems are often used for equipment state review and safety confirmation. The existing technology adopts a hierarchical control strategy: when a sensor detects an equipment anomaly (such as a sudden drop in pipeline pressure), a shutdown instruction needs to be executed after verifying the authenticity of the fault through video to avoid misoperation. This strategy is highly dependent on the final confirmation function of the video system.

[0003] However, since the core equipment of the filling station (such as industrial pumps and lighting systems) often shares a power supply circuit, executing a shutdown instruction will simultaneously interrupt the working conditions of the video monitoring, resulting in the inability to complete fault verification, which forms a logical dead loop of "needing video verification to trigger shutdown" and "shutdown causing video failure", significantly delaying fault disposal and increasing the risk of secondary accidents. SUMMARY

[0004] The present application provides a filling station control method and system based on video monitoring to solve the technical problems in the prior art.

[0005] The technical solution of the present application to solve the above technical problems is as follows:

[0006] The present application provides the following technical solution:

[0007] The filling station control method based on video monitoring comprises:

[0008] S1, acquiring a sensor anomaly signal of a filling station equipment and a real-time video stream of a video monitoring system;

[0009] S2, maintaining the current state when the sensor anomaly signal does not exceed a preset threshold, otherwise generating a backup judgment signal;

[0010] S3, after generating the backup judgment signal, calculating a pixel entropy value of a fault area based on the real-time video stream to generate a thermodynamic entropy increase trajectory, and starting a backup power supply to independently power the video monitoring system when the slope of the entropy increase trajectory is greater than a set threshold;

[0011] S4, analyzing the fidelity decay characteristics of the quantum state evolution trajectory based on the real-time video quantum distribution characteristics under the backup power supply state, and reconstructing the turbulent dynamics characteristics and extracting the chaos index based on the real-time video stream motion characteristics;

[0012] S5, generating a shutdown instruction when the fidelity decay characteristics meet the quantum effect characteristic curve and the chaos index reaches a preset chaos critical threshold; otherwise, generating a main power supply recovery instruction;

[0013] S6, executing a shutdown instruction or a main power supply resuming instruction.

[0014] Further, the sensor abnormal signal of the filling station device and the real-time video stream of the video monitoring system are acquired, including:

[0015] The pipeline pressure sensor signal, the slurry flow sensor signal and the pump body vibration sensor signal of the industrial pump are acquired as the sensor abnormal signal;

[0016] The filling pipeline connection monitoring video stream corresponding to the pipeline pressure sensor signal and the pump valve interface monitoring video stream corresponding to the pump body vibration sensor signal are acquired as the real-time video stream.

[0017] Further, when the sensor abnormal signal does not exceed the preset threshold, the current state is maintained, otherwise a backup judgment signal is generated, including:

[0018] It is judged whether the pipeline pressure sensor signal of the industrial pump exceeds the pipeline pressure mutation rate threshold;

[0019] It is judged whether the slurry flow sensor signal exceeds the flow deviation threshold;

[0020] It is judged whether the pump body vibration sensor signal exceeds the vibration frequency domain energy threshold;

[0021] When the pipeline pressure sensor signal exceeds the pipeline pressure mutation rate threshold, or the slurry flow sensor signal exceeds the flow deviation threshold, or the pump body vibration sensor signal exceeds the vibration frequency domain energy threshold, the backup judgment signal is generated; otherwise, the current running state of the filling station is maintained.

[0022] Further, after the backup judgment signal is generated, the pixel entropy value of the fault area is calculated based on the real-time video stream to generate a thermodynamic entropy increase trajectory, and when the slope of the entropy increase trajectory is greater than a set threshold, the backup power supply is started to independently supply power to the video monitoring system, including:

[0023] According to the sensor abnormal signal type corresponding to the backup judgment signal, the spatial position of the fault area in the real-time video stream is located:

[0024] The pixel gray value of a plurality of continuous video frames in the fault area is extracted, and the pixel entropy value of each video frame is calculated;

[0025] The pixel entropy values are connected in the video frame time sequence to form a thermodynamic entropy increase trajectory;

[0026] The average slope of the thermodynamic entropy increase trajectory in the latest preset number of video frame time windows is calculated;

[0027] When the average slope is greater than the entropy increase slope threshold, a start instruction is sent to the uninterruptible power supply to independently supply power to the video monitoring system.

[0028] Further, when the standby judgment signal is triggered by the pipeline pressure sensor signal, the fault area is the pipeline connection area in the filling pipeline connection monitoring video stream; when the standby judgment signal is triggered by the pump body vibration sensor signal, the fault area is the pump valve sealing area in the pump valve interface monitoring video stream.

[0029] Further, in the standby power supply state, the fidelity decay characteristic of the quantum state evolution trajectory is analyzed based on the real-time video light quantum distribution characteristic, and the turbulent flow dynamics characteristic is reconstructed and the chaos index is extracted based on the real-time video flow motion characteristic, including:

[0030] In the standby power supply state, the infrared photon intensity sequence of the real-time video stream of the fault area is collected to form sequence data of the change of photon intensity with time;

[0031] According to the photon intensity sequence data, the change process of the quantum state probability distribution is determined to generate a quantum state evolution trajectory;

[0032] Quantize the state maintenance ability change value of the quantum state evolution trajectory in a fixed time interval as the fidelity decay characteristic;

[0033] At the same time, the pixel displacement of the real-time video stream of the fault area is analyzed to obtain the fluid motion velocity distribution;

[0034] According to the fluid motion velocity distribution, the fluid dynamic behavior characteristic is reconstructed;

[0035] Calculate the chaos degree quantization value of the fluid dynamic behavior characteristic as the chaos index.

[0036] Further, when the fidelity decay characteristic meets the quantum effect characteristic curve and the chaos index reaches the preset chaos critical threshold, a shutdown instruction is generated; otherwise, a main power supply recovery instruction is generated, including:

[0037] Compare the consistency of the change law of the fidelity decay characteristic and the quantum effect characteristic curve;

[0038] Determine whether the chaos index exceeds the preset chaos critical threshold;

[0039] When the fidelity decay characteristic meets the change law of the quantum effect characteristic curve and the chaos index exceeds the preset chaos critical threshold, a shutdown instruction is sent to the filling station main controller;

[0040] When the fidelity decay characteristic does not meet the change law of the quantum effect characteristic curve or the chaos index does not exceed the preset chaos critical threshold, a main power supply recovery instruction is sent to the power supply switching device.

[0041] Further, the quantum effect characteristic curve is obtained by the following method:

[0042] Acquiring infrared photon intensity sequences of typical industrial fault scenes under standard light source conditions using quantum image sensors;

[0043] Analyzing typical decay rules of quantum state fidelity over time according to the infrared photon intensity sequences;

[0044] Constructing the change mode of the typical decay rules in a fixed time interval as a quantum effect characteristic curve.

[0045] Further, executing a shutdown instruction or a main power supply recovery instruction, comprising:

[0046] When receiving the shutdown instruction, controlling the filling station main controller to execute an industrial pump shutdown operation, while maintaining the standby power supply to the video monitoring system and recording the shutdown state;

[0047] When receiving the main power supply recovery instruction, controlling the power supply switching device to switch the video monitoring system to the main power supply loop, while restoring the normal operation state of the filling station main controller and generating a power supply loop recovery log.

[0048] In another aspect, the present application provides a filling station control system based on video monitoring, comprising:

[0049] A signal acquisition module for acquiring sensor abnormal signal of the filling station equipment and real-time video stream of the video monitoring system;

[0050] A signal judgment module for maintaining the current state when the sensor abnormal signal does not exceed the preset threshold, otherwise generating a standby judgment signal;

[0051] An entropy increase analysis module for generating a pixel entropy value of the fault area based on the real-time video stream to generate a thermodynamic entropy increase trajectory after generating the standby judgment signal, and starting the standby power supply to independently supply power to the video monitoring system when the slope of the entropy increase trajectory is greater than the set threshold;

[0052] A chaos analysis module for analyzing the fidelity decay characteristics of the quantum state evolution trajectory based on the real-time video quantum distribution characteristics under the standby power supply state, and reconstructing the turbulent dynamics characteristics and extracting the chaos index based on the real-time video stream fluid motion characteristics;

[0053] An instruction generation module for generating a shutdown instruction when the fidelity decay characteristics meet the quantum effect characteristic curve and the chaos index reaches the preset chaos critical threshold; otherwise, generating a main power supply recovery instruction;

[0054] An instruction execution module for executing the shutdown instruction or the main power supply recovery instruction.

[0055] The beneficial effects of the present application are:

[0056] 1. By establishing a video analysis driven pre-judgment-verification-execution cascade mechanism, the power supply dead loop of device shutdown and video verification is effectively broken, the pre-judgment switching of standby power supply is triggered by entropy increase trajectory analysis, and the quantum state fidelity and fluid chaos index are verified in the independent power supply environment, which not only guarantees the integrity of the fault criterion, but also ensures the continuous operation of video monitoring in the key decision-making stage, so that the system has completed automatic verification before the main power of the device is cut off, avoiding the risk of video verification failure due to power failure in the traditional scheme, and significantly shortening the fault response time.

[0057] 2. By dual physical modeling of quantum state evolution trajectory and turbulent flow dynamics, a more reliable fault confirmation system is built, which uses the distribution characteristics of light quantum to capture the material phase change process, and combines the fluid motion chaos index to identify the turbulent instability state, forming a multi-dimensional criterion for abnormal molten slurry. This verification mechanism based on the physical nature can more accurately distinguish between real faults and sensor false alarms than traditional manual video review, while maintaining key equipment monitoring, avoiding unnecessary shutdown of industrial pumps causing production line interruptions, and preventing secondary accidents such as pipe bursts caused by delayed handling. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 Flowchart of the filling station control method based on video monitoring of the present application;

[0059] Figure 2 Structure diagram of the filling station control system based on video monitoring of the present application. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application.

[0061] Embodiment 1: Figure 1 The filling station control method based on video monitoring of the present application is given, which includes:

[0062] S1, acquiring the sensor abnormal signal of the filling station device and the real-time video stream of the video monitoring system;

[0063] S2, maintaining the current state when the sensor abnormal signal does not exceed the preset threshold, otherwise generating a standby judgment signal;

[0064] S3, after generating the standby judgment signal, generating a thermodynamic entropy increase trajectory based on the pixel entropy value of the fault area calculated from the real-time video stream, and starting the standby power supply to independently power the video monitoring system when the slope of the entropy increase trajectory is greater than a set threshold value;

[0065] S4, in the standby power supply state, analyzing the fidelity decay characteristics of the quantum state evolution trajectory based on the real-time video quantum distribution characteristics, and reconstructing the turbulent dynamics characteristics and extracting the chaos index based on the real-time video stream motion characteristics;

[0066] S5, generating a shutdown instruction when the fidelity decay characteristics meet the quantum effect characteristic curve and the chaos index reaches a preset chaos critical threshold value; otherwise, generating a main power supply recovery instruction;

[0067] S6, executing the shutdown instruction or the main power supply recovery instruction.

[0068] S1, acquiring the sensor abnormal signal of the filling station equipment and the real-time video stream of the video monitoring system, specifically including:

[0069] The sensor abnormal signal of the filling station equipment is acquired by a special sensor installed at a key position of the industrial pump. The pipe pressure sensor signal of the industrial pump is collected by a piezoresistive pressure sensor fixed to the flange connection at a position 0.5 meters upstream of the filling pipe. The sensor is connected to a data acquisition module through a current loop, and the pipe pressure value is sampled every 50 milliseconds. The slurry flow sensor signal is detected by an electromagnetic flowmeter installed at a vertical pipe section 1.2 meters downstream of the industrial pump outlet. The flowmeter converts the induced electromotive force generated when the slurry flows through the electrodes into an instantaneous flow value, and the flow data is transmitted to the control cabinet at a frequency of 20 times per second. The pump body vibration sensor signal is acquired by a piezoelectric acceleration sensor rigidly fixed to the industrial pump drive end bearing seat, which detects the axial vibration signal. The vibration signal is converted into a voltage signal by a charge amplifier and then digitized at a sampling rate of 1024 points per second.

[0070] The real-time video stream of the video monitoring system is acquired using a strategy linked to the physical position of the sensor. The filling pipe connection monitoring video stream corresponding to the pipe pressure sensor signal is captured by an infrared network camera installed 1.5 meters away from the flange connection point. The camera optical axis is directed at the flange sealing surface, and the viewing angle coverage ensures that all flange bolts are visible. The pump valve interface monitoring video stream corresponding to the pump body vibration sensor signal is captured by an explosion-proof infrared camera installed 0.8 meters above the pump body base. The camera uses a 30-degree downward angle to ensure that the pump shaft mechanical seal surface is completely imaged in the central area of the screen. The above monitoring video streams are transmitted to the video processing server through industrial Ethernet. The video stream encoding format is compressed, and the resolution is set to 1280x1024 pixels.

[0071] The filling pipe joint monitoring video stream and the pump valve interface monitoring video stream each include an infrared video image sequence of no less than 30 frames per second, which is generated by a built-in refrigeration type detector of the camera. The detector works in a wave band of 8 to 14 microns, and the intensity of thermal radiation is detected by rotating a filter wheel, and the gray value of each pixel point corresponds to the temperature distribution of the object surface. The time synchronization mechanism of the video image sequence adopts a precise time protocol, and the deviation between the video frame timestamp and the sensor data timestamp is controlled within plus or minus 2 milliseconds, so that the time domain is aligned.

[0072] In the process of acquiring the pipe pressure sensor signal of the industrial pump, the installation position of the pressure sensor is determined according to the principle of fluid mechanics. The installation at the upstream 0.5-meter straight pipe section of the flange joint can avoid turbulent interference, and this position meets the arrangement requirements of the pressure measurement point. The measurement of the slurry flow sensor signal adopts a non-contact electromagnetic detection technology. The excitation coil of the electromagnetic flowmeter generates an alternating magnetic field. When the fluid containing the conductive slurry cuts the magnetic force line, an electromotive force proportional to the flow rate is generated between the electrodes. The pump body vibration sensor signal is collected according to the mechanical vibration measurement specification. The sensor is directly and rigidly coupled to the bearing seat shell through a stainless steel base, avoiding resonance frequency interference, and the frequency response range covers 10 Hz to 5000 Hz.

[0073] The spatial correspondence of the filling pipe joint monitoring video stream is established by camera calibration. During the equipment installation stage, a checkerboard calibration board is placed at the position of the flange sealing surface, and the camera parameters are calculated by a calibration method to establish the mapping relationship between the image pixel coordinates and the physical space coordinates. The correspondence of the pump valve interface monitoring video stream adopts laser assisted positioning, and a visible laser mark is projected at the center point of the pump valve sealing ring. The camera focal length is adjusted so that the mark is imaged at the center pixel coordinate position of the picture.

[0074] The generation process of the infrared video image sequence involves thermal radiation energy conversion. The detector converts the received infrared photons into electronic signals, and the analog signals of each pixel point are converted to form digital gray values. The conversion relationship between the gray value and the temperature value is calibrated by a radiation source. In the calibration process, a standard surface source thermal radiation device is used to set multiple temperature points in the range of 30 degrees Celsius to 300 degrees Celsius, a gray-temperature correspondence table is established, and the table is fixed to the camera firmware. The acquisition rate of 30 frames per second is realized by the frame buffer control of the camera processor, and a double buffer structure is adopted to ensure continuous video stream without frame loss.

[0075] The time and space synchronization of the sensor signal and the video stream is realized by a hardware timestamp mechanism. A time chip is built into the data acquisition module, which receives a time signal as a reference once per second. The video encoder is synchronized with the time source through a clock synchronization protocol. In the data transmission layer, both the sensor data packet and the video stream data packet carry high-precision timestamps, and the central processing unit establishes the correspondence between the sensor abnormal events and the video frames through a timestamp alignment algorithm.

[0076] The effectiveness verification of the filling pipe connection monitoring video stream adopts a dynamic range detection method. During daily equipment self-checking, a constant temperature reference source is placed in the camera field of view to check whether the gray value fluctuation of the reference source area in the video image is less than the set threshold. The availability of the pump valve interface monitoring video stream is guaranteed by motion blur detection, and the sharpness of the sealing surface edge is analyzed in the pump shaft rotation state. When the edge sharpness is lower than a certain gradient, the camera exposure parameters are automatically adjusted.

[0077] The reliability of the pipe pressure sensor signal of the industrial pump is ensured by a double redundancy mechanism. Two identical type pressure sensors are installed in parallel at the key pipe node, and when the difference between the readings of the two sensors exceeds the threshold, the sensor cross-checking program is started. The accuracy of the slurry flow sensor signal is maintained using zero point calibration technology, and automatic zero point calibration is performed periodically in the pump stopped state. The integrity detection of the pump body vibration sensor signal is realized by background noise monitoring, and when the sensor detects that the vibration effective value is greater than the reference value in the equipment stopped state, the sensor check alarm is triggered.

[0078] The quality control of the infrared video image sequence contains three dimensions. The spatial resolution is verified periodically by testing the graph to ensure that a certain width of thermal radiation stripes can be distinguished. The temperature resolution is detected by a temperature difference test board, which can distinguish small temperature difference targets at an ambient temperature of 25 degrees Celsius. The time resolution adopts a rotating verification device to ensure that the video image clearly presents the moving target at a standard speed. All quality control data are recorded in the equipment maintenance log.

[0079] The transmission path of the sensor abnormal signal adopts an industrial level communication protocol to ensure. The pipe pressure sensor signal is transmitted through the bus, and the bus terminal matches the impedance resistance to avoid signal reflection. The slurry flow sensor signal adopts a check transmission protocol, and each data packet contains a check code. The pump body vibration sensor signal is directly connected to a high-precision data acquisition card through a cable. The network transmission of real-time video stream is configured with a quality of service strategy to ensure that the video stream transmission delay meets the real-time analysis requirements.

[0080] S2, when the sensor abnormal signal does not exceed the preset threshold, the current state is maintained, otherwise a standby judgment signal is generated, the specific implementation includes:

[0081] The process of determining whether the pipeline pressure sensor signal of the industrial pump exceeds the pipeline pressure mutation rate threshold is achieved by real-time calculation of the pressure change rate. The pipeline pressure mutation rate threshold is determined according to the safety operation requirements of the industrial pump, and the threshold represents the maximum allowable change amplitude of the pressure per unit time. In specific calculation, the difference between the pressure values of two adjacent sampling time points is divided by the sampling interval time, for example, if the difference between the current pressure value and the previous pressure value divided by the time interval exceeds the preset mutation rate threshold, it is determined to be abnormal. The method for obtaining the pipeline pressure mutation rate threshold is: under the normal working condition of the industrial pump, the historical pressure data is counted, the distribution range of the pressure change rate is calculated, and the upper limit value of the normal range, for example, 1.6 times, is taken as the threshold. The calculation process is executed every 50 milliseconds in the programmable logic controller, and the pressure data is derived from the pipeline pressure sensor signal sampled every 50 milliseconds in step S1.

[0082] The determination of whether the slurry flow sensor signal exceeds the flow deviation threshold is achieved by comparing the current flow with the reference flow. The determination method of the flow deviation threshold is: during the stable production stage, the flow data is continuously recorded, the average value of the flow is calculated as the reference value, and the allowed deviation range is set to be, for example, plus or minus 15% of the reference value. In real-time determination, the current flow sensor reading is compared with the dynamically updated reference value, the reference value is recalculated every 30 minutes, and the moving average algorithm is used to eliminate the influence of short-term fluctuations. The flow data is derived from the slurry flow sensor signal sampled 20 times per second in step S1, and the arithmetic mean of the 20 sampling values per second is taken as the flow value of the current second in calculation.

[0083] The determination of whether the pump body vibration sensor signal exceeds the vibration frequency energy threshold requires signal spectrum analysis. The vibration frequency energy threshold is set based on the mechanical structure characteristics of the equipment, and the characteristic frequency band reflecting bearing wear is focused on. In specific implementation, the time domain signal collected by the vibration sensor is subjected to fast Fourier transform, and the energy integral value in a specific frequency range is extracted as an evaluation index, for example, for the drive end bearing of the industrial pump, the vibration energy in the frequency band of 100 Hz to 500 Hz is analyzed. The method for obtaining the vibration frequency energy threshold is: during the debugging stage of the new equipment, the vibration spectrum during normal operation is measured, and the threshold is taken as, for example, 1.6 times of the energy value of the characteristic frequency band. In vibration data processing, each data block of 1024 sampling points is used for frequency spectrum calculation, and the sampling frequency is 1024 points per second, i.e. the frequency spectrum is updated once per second.

[0084] A standby judgment signal is generated when the pipe pressure sensor signal exceeds the pipe pressure mutation rate threshold, or the slurry flow sensor signal exceeds the flow deviation threshold, or the pump body vibration sensor signal exceeds the vibration frequency domain energy threshold. The judgment logic is implemented through the program of the programmable logic controller, and the three judgment conditions are connected in a logical OR relationship. The standby judgment signal is a digital output signal, and a high level indicates that the standby judgment process needs to be started, and a low level indicates a normal state. After the signal is generated, it is transmitted to the video analysis server through the industrial Ethernet.

[0085] When all the judgment conditions are not met, the current running state of the filling station is maintained. At this time, the control program continuously monitors the sensor signals but does not change the state of any actuator, and at the same time sets the system state code to normal running mode. During the maintenance state, the sensor readings are recorded periodically as a running log, and the log data includes time stamp, pressure value, flow value, vibration energy value and judgment result flag bit.

[0086] The preprocessing in the pipe pressure mutation rate calculation includes signal filtering. The original pressure signal is first filtered by a low-pass filter to remove high-frequency noise, and the cutoff frequency is set to, for example, 10 Hz. The filtered signal is then subjected to mutation rate calculation to avoid false triggering caused by noise interference. For signal abnormality, there is a protection mechanism, when the pressure sensor is detected to be disconnected or out of range, the mutation rate calculation of this sampling point is automatically ignored and a sensor fault alarm is issued.

[0087] The reference value update in the flow deviation judgment has conditional constraints. Only when the following conditions are met, the reference value is updated: the current flow fluctuation coefficient is less than, for example, 0.05; and the continuous stable running time exceeds, for example, 15 minutes; and there is no external operation instruction intervention. The reference value calculation uses a weighted average algorithm, and the new reference value is equal to a certain weight of the historical reference value plus a certain weight of the current average flow, for example, updated by 70% historical reference value plus 30% current average.

[0088] The vibration frequency domain energy analysis includes data validity verification. Before processing each data block, the peak-to-peak value of the time domain signal is calculated, and if the peak-to-peak value exceeds, for example, 80% of the range, the data block is discarded. The frequency spectrum calculation uses a windowing function to reduce frequency spectrum leakage, and the frequency resolution is set to, for example, 1 Hz, and the energy integration excludes the power frequency interference frequency.

[0089] The generation of the standby judgment signal includes anti-jitter processing. When a single sensor triggers an exception, it needs to continuously meet the exception condition for a certain time to confirm the generation of the signal, for example, the pipe pressure mutation needs to exceed the limit for 3 consecutive calculations to be determined valid. When multiple sensors trigger at the same time, a priority response strategy is adopted, the vibration signal triggers an immediate response, and the pressure and flow signals trigger a confirmation for, for example, 500 milliseconds.

[0090] There is an implicit monitoring mechanism during the running state maintenance period. The system background continuously tracks the trend changes of the three sensor signals. When any signal approaches its threshold value, for example, 80%, the data cache is started in advance to prepare for the possible trigger of the backup judgment. At the same time, the threshold value is automatically adjusted regularly, and the original threshold value is compensated and corrected according to the device running time and the environmental temperature.

[0091] All data processing of the judgment process is executed in the real-time operating system environment of the industrial controller. The pressure mutation rate calculation task priority is set to the highest, and it is executed by forced interruption every 50 milliseconds; the flow deviation judgment task is executed every 1 second; the vibration analysis task is triggered to execute after 1024 points of data are ready. Data is exchanged between tasks through shared memory.

[0092] The history record of the sensor signal overrun judgment forms an event log. Each time a backup judgment signal is generated, the triggering sensor type, overrun value, duration, and original data snapshot of the previous 60 seconds are recorded. The judgment results when the signal is not triggered generate a statistical report, and the maximum overrun proportion and duration distribution of each sensor are output regularly.

[0093] The transmission of the backup judgment signal uses a reliable transmission protocol with a retransmission mechanism. Each signal is encapsulated as a data frame after generation, including frame number, generation timestamp, and check code. The receiver returns an acknowledgement signal for each successfully received frame, and automatically retransmits if it is not acknowledged within a timeout period. The maximum number of retransmissions is set to, for example, 3. The signal transmission path is physically isolated from the sensor data transmission channel described in step S1.

[0094] In the maintenance of the current running state stage, the system continuously outputs the running health index. The index is calculated by the three sensor overrun degrees, and the pressure mutation rate overrun proportion accounts for, for example, 40% of the weight, the flow deviation overrun proportion accounts for, for example, 30% of the weight, and the vibration energy overrun proportion accounts for, for example, 30% of the weight. The health index is updated regularly and displayed in real time through the human-machine interface.

[0095] S3, after generating the backup judgment signal, the pixel entropy value of the fault area is calculated based on the real-time video stream to generate a thermodynamic entropy increase trajectory. When the entropy increase trajectory slope is greater than the set threshold value, the backup power supply is started to independently power the video monitoring system, including:

[0096] After generating the backup judgment signal, the spatial position of the fault area in the real-time video stream is determined according to the sensor abnormal signal type corresponding to the signal. When the backup judgment signal is triggered by the pipeline pressure sensor signal, the fault area is located in the pipeline connection area in the filling pipeline connection monitoring video stream, which refers to a rectangular range with a flange sealing surface as the center, for example, a width of 50 pixels and a height of 50 pixels. The positioning method is to read the camera calibration parameters, convert the physical coordinates of the flange connection point to image pixel coordinates, and define a fixed size area with the coordinates as the center point. When the backup judgment signal is triggered by the pump body vibration sensor signal, the fault area is located in the pump valve sealing area in the pump valve interface monitoring video stream, which refers to the projection area of the mechanical seal moving ring in the video screen, and the analysis area is a range with a width of, for example, 20 pixels and a height of, for example, 40 pixels around the marker point.

[0097] The pixel gray value of a plurality of continuous video frames in the fault area is extracted. For each video frame, the gray value of all pixels in the fault area is obtained, and the gray value ranges from 0 to 255. The gray value of each pixel is derived from the original data output by the infrared detector of the camera. The number of continuous video frames is set to a fixed value, for example, 30 continuous frames correspond to about 1 second, and the time interval between frames is strictly kept fixed.

[0098] The pixel entropy value of each video frame is calculated. The pixel entropy value represents the degree of disorder of the gray scale distribution in the fault area of the frame, and the calculation method is to calculate the negative value of the sum of the product of the proportion of each gray scale and the logarithm base 2. In specific implementation, first, the number of pixels of each gray scale is counted, and the frequency of each gray scale is calculated, and then the sum operation is performed according to the entropy value formula. The calculation is completed on the video processing server, and the processing time of each frame is controlled within a certain time.

[0099] The pixel entropy values are connected in the time sequence of the video frames to form a thermodynamic entropy increase trajectory. The continuously calculated pixel entropy values are arranged in time sequence, with the horizontal coordinate being the video frame timestamp and the vertical coordinate being the entropy value, to form a time sequence curve. The timestamp accuracy is millisecond level, which is synchronized with the sensor timestamp of S1 step. The curve is drawn in the memory buffer, and only the data of the recent specific number of frames are retained for subsequent analysis.

[0100] The average slope of the thermodynamic entropy increase trajectory in the recent preset number of video frame time windows is calculated. The time window size is set to a fixed number of frames, for example, the last 5 frames. The average slope calculation method is to perform linear regression analysis on all data points in the time window, with time as the independent variable and entropy value as the dependent variable, and to calculate the slope coefficient of the regression straight line. The regression analysis is realized by using the least square method, and the calculation result is updated every time a new frame arrives.

[0101] When the average slope is greater than the entropy increase slope threshold, a start instruction is sent to the uninterruptible power supply. The entropy increase slope threshold is set according to the thermodynamic characteristics of the equipment, and the threshold represents a safe limit of the entropy increase rate per unit time. The threshold is obtained by collecting video data when the equipment is normally running, calculating the entropy increase slope of each period, and taking a specific multiple of the upper limit value of the statistical distribution as the threshold, for example, 1.25 times of the maximum observed slope. The start instruction is a direct current pulse signal, which is connected to the control port of the uninterruptible power supply through hardwiring.

[0102] The coordinate conversion in the fault area positioning process uses a transformation matrix. The matrix is determined by calibration when the equipment is installed: a calibration board of known size is placed in the target area, and the camera imaging model parameters are solved. Each time the physical coordinates are directly substituted into the matrix equation to calculate the pixel coordinates.

[0103] The pixel gray value extraction is provided with an abnormal processing mechanism. When it is detected that more than a certain proportion of pixel gray values in the fault area are limit values, it is determined that the image is abnormal, and the analysis area is automatically expanded to re-extract. If it is still invalid after expansion, the valid data of the previous frame is used and an adjustment instruction is triggered.

[0104] The entropy value calculation process uses histogram acceleration. The counting units are pre-allocated, and when each pixel in the area is traversed, the corresponding counting unit of its gray value is incremented by 1, and finally the distribution is obtained by dividing the counting unit value by the total number of pixels.

[0105] The storage of the thermodynamic entropy increase trajectory uses a circular buffer structure. The buffer capacity is a specific group of data, and the newest data covers the oldest data. Each time new data is added, the data subset required for slope calculation within the time window is updated synchronously.

[0106] The average slope calculation includes numerical stability processing. During the calculation process, the time variable is zero-meaned: all time stamps within the time window are subtracted from the time stamp of the center point of the time window. In the regression coefficient calculation formula, the slope is equal to the covariance of time and entropy value divided by the variance of time, and the covariance and variance are updated in real time through the incremental algorithm.

[0107] The dynamic adjustment of the entropy increase slope threshold includes environmental compensation. The threshold is corrected according to the readings of the environmental temperature sensor, for example, when the environmental temperature rises, the threshold is adjusted proportionally. The correction parameter is determined by testing, and a preset temperature compensation table is stored in the memory.

[0108] A safety confirmation step is provided before the start instruction is sent. After the average slope first exceeds the limit, it needs to be continuously calculated for a specific number of times before the instruction is actually sent. The video frame sequence and entropy value data at a specific time before the trigger time are recorded when the instruction is sent.

[0109] The power supply state verification after the UPS starts is realized by current detection. After a specific time of instruction sending, the current value of the video monitoring system power supply loop is detected. If the current rises to the normal working range, it is determined that the independent power supply is successful. If the current does not change, the switching device is started and an alarm is sent.

[0110] The triggering condition of fault area relocation includes equipment displacement monitoring. A displacement sensor is installed on the pump body base. When it is detected that the equipment position deviates more than a certain value, the camera calibration process is automatically re-executed.

[0111] The logarithmic operation in the pixel entropy value calculation process is optimized by using a lookup table. The logarithmic value of the probability value is calculated in advance and stored as a lookup table. During actual calculation, the probability value is converted into an index to query the table.

[0112] The smoothing processing of the thermodynamic entropy increase trajectory uses moving average filtering. Before storing in the buffer, the entropy value data is processed by multiple point moving average. The filtered data is used for slope calculation, and the original data is reserved for diagnosis.

[0113] The record of the average slope overrun event contains multi-dimensional data. The original video frame, sensor reading, and environmental information in this time window are saved to form a complete event snapshot. The event data is stored in a compressed format.

[0114] The communication protocol for sending start instructions to the UPS uses a standard format. The instruction data packet includes a start symbol, device address code, function code, instruction type, check code, and end symbol. The transmission medium is twisted pair, and the baud rate is set to a specific value. Each instruction is sent repeatedly a certain number of times, and the receiver needs to return an acknowledgment frame.

[0115] The system state monitoring during the independent power supply of the backup power supply includes power quality analysis. The output voltage fluctuation range is detected in real time, and when the voltage deviation exceeds the allowed value, it is automatically switched to the redundant module. At the same time, the power supply duration and load current curve are recorded.

[0116] S4, in the backup power supply state, based on real-time video light quantum distribution characteristic analysis quantum state evolution trajectory fidelity decay characteristic, and based on real-time video flow motion characteristic reconstruction turbulence dynamics characteristic and extract chaos index, specific implementation includes:

[0117] In the standby power supply state, the infrared photon intensity sequence of the real-time video stream of the fault area is collected. The specific implementation is: from the infrared video image sequence described in step S1, which is not less than 30 frames per second, the gray value of all pixels in the fault area is extracted, and the gray value directly corresponds to the infrared photon intensity received by the detector. Summing up the gray values of all pixels in the region as a single-frame photon intensity value, arranging the video frame time sequence to form a sequence data of the change of photon intensity with time. The time resolution of the sequence data is consistent with the frame interval, and the data of the last certain number of frames are retained for subsequent analysis. The physical unit of the photon intensity value is relative intensity unit, and the corresponding relationship with the number of photons is established through camera calibration.

[0118] According to the photon intensity sequence data, the change process of the quantum state probability distribution is determined. The quantum state probability distribution is represented as the state distribution of the thermal radiation field of the fault area, and the calculation method is: regarding the continuous multiple frames of photon intensity sequence as an observation window, the quantum state evolution is described by mathematics. In the specific implementation, a description matrix of a two-dimensional state space is constructed, and the matrix elements are calculated from the photon intensity change value of adjacent frames. The quantum state corresponding to each video frame is updated by the evolution equation, and the initial state is set as the thermal equilibrium state. The evolution trajectory is composed of the sequence of description matrices of continuous video frames.

[0119] The change value of the state maintenance ability of the quantum state evolution trajectory in a fixed time interval is quantized as the fidelity decay characteristic. The fixed time interval is set as a continuous certain number of video frames. The calculation method of the fidelity decay characteristic is: taking the quantum state description matrices of the starting frame and the ending frame of the time interval, calculating the state similarity value between them, and then dividing the time interval to obtain the decay rate. The state similarity value is realized by mathematical operation. The decay rate unit is per second.

[0120] Meanwhile, the real-time video stream of the fault area is analyzed for pixel displacement. The position change amount of each pixel in adjacent video frames is calculated by using a motion detection method, and the position change amount is obtained by comparing the gray difference between the target pixel and the surrounding pixels. The analysis area is completely consistent with the quantum state analysis area, and the calculation interval is every two frames. The horizontal and vertical components of the position change constitute a two-dimensional velocity field, and the velocity unit is converted into physical velocity value through calibration parameters.

[0121] According to the fluid motion velocity distribution, the fluid dynamic behavior characteristics are reconstructed. The reconstruction method is: a two-dimensional velocity distribution map is established in the analysis area, and the flow characteristics of each position are calculated by mathematics. The fluid dynamic behavior characteristics include: average kinetic energy (average value of the square of the speed of all positions), rotation intensity peak value, and maximum value of the gradient of the speed change. The reconstruction process is executed periodically, and the generated characteristic parameter sequence constitutes a dynamic behavior characteristic vector.

[0122] The chaotic degree quantization value of the calculated fluid dynamic behavior characteristics is taken as the chaos index. The chaos index is represented by the divergence rate index, and the calculation method is as follows: a plurality of groups of dynamic behavior characteristic vectors are taken, trajectories are constructed in the characteristic space, and the separation rate of adjacent trajectories is calculated by an algorithm. Specifically, the nearby point pairs are located in the characteristic space, the distance growth rate of the point pairs with time is measured, and the average value of the growth rates of all point pairs is taken as the chaos index.

[0123] Noise suppression in the photon intensity sequence acquisition process uses filtering processing. The filtering strength is dynamically adjusted according to the gray level fluctuation degree of the background area, and spatial filtering is enabled when the fluctuation degree is greater than a certain threshold. The filtered data is stored in the buffer area, and the quantum state analysis is triggered when the buffer area is full.

[0124] The initialization of the description matrix in the quantum state probability distribution calculation is based on the ambient temperature. The temperature value near the fault area is read by a temperature sensor, and the initial state is set according to the temperature distribution. When the temperature change exceeds a certain value, the initialization is reinitialized. The evolution step length is fixed when the matrix is updated, and is consistent with the video frame interval.

[0125] The time interval division of the fidelity decay characteristic uses a sliding window mechanism. The window size is fixed, and the window slides forward by one frame every time a new frame arrives, and the decay rate in the latest interval is recalculated. The calculation result is stored with the timestamp.

[0126] The speed field calibration of pixel displacement analysis includes image distortion correction. The correction parameters obtained in the camera calibration stage are used to correct the edge pixel displacement amount. Coordinate correction is performed before displacement calculation.

[0127] The reconstruction of fluid dynamic behavior characteristics includes data validity verification. When it is detected that more than a certain proportion of pixel motion amounts are unreliable, an alternative algorithm is enabled. The calculation resources consumed by the reconstruction process are monitored in real time, and the spatial resolution is automatically reduced when the single reconstruction time exceeds a threshold.

[0128] The trajectory construction in the chaos index calculation uses time delay technology. The characteristic space trajectory is constructed by mathematical method, which ensures that the original system characteristics are retained. The distance calculation uses the geometric distance formula, and the nearby point search is accelerated by the optimization algorithm.

[0129] The spatiotemporal alignment of the quantum state evolution trajectory and the fluid behavior characteristics is realized by timestamp matching. The same video frame sequence is used for the two types of analysis, and the processing start time deviation is controlled within 1 frame time. The result data is stored with a unified time reference mark.

[0130] The environmental compensation of the fidelity decay characteristic includes pressure correction. The decay rate is linearly compensated by the pressure sensor reading of the pipeline, and the correction coefficient is adjusted in proportion when the pressure changes.

[0131] The sampling area of fluid motion analysis can be dynamically adjusted. When the chaos index exceeds the warning threshold for a certain number of consecutive times, the analysis area is automatically expanded, and the sampling point density is increased. The data after expansion is processed in parallel with the original data.

[0132] Real-time visualization of chaos index is achieved through human-machine interface. The current chaos index value, historical trend curve and warning level are displayed on the monitoring screen. The display update frequency is a certain value.

[0133] The calculation of quantum state analysis is accelerated by parallel processing. Matrix operations are decomposed into multiple parallel threads. State similarity calculation is achieved through a dedicated function.

[0134] Temperature monitoring of the fault area is linked with quantum state analysis. When the average temperature change rate of the area exceeds a certain value, the quantum state evolution trajectory calculation is suspended, and then reinitialized after the temperature stabilizes.

[0135] The verification of chaos index is completed by test data. In the system self-checking stage, the video sequence of the preset flow pattern is injected, and the deviation of the output chaos index from the reference value is verified.

[0136] The storage of analysis results adopts a hierarchical structure. The original photon intensity sequence, quantum state trajectory, fluid feature vector and chaos index are stored respectively. The data retention policy is set according to the state classification.

[0137] The monitoring of standby power supply state is linked with the analysis process. When the output voltage fluctuation of the power supply exceeds the allowed range, the quantum state analysis process is automatically suspended, and then resumed after the power supply stabilizes. The power supply monitoring data and video analysis results are stored synchronously.

[0138] S5, when the fidelity decay characteristic meets the quantum effect characteristic curve and the chaos index reaches the preset chaos critical threshold, a shutdown instruction is generated; otherwise, a main power supply recovery instruction is generated, which includes:

[0139] The consistency of the change rule of the fidelity decay characteristic and the quantum effect characteristic curve is achieved by time series matching algorithm. The quantum effect characteristic curve is a reference curve stored in the database in advance, which represents the benchmark mode of quantum state fidelity decay under typical fault conditions. The specific comparison process is as follows: take the fidelity decay rate sequence of a certain number of consecutive video frames as the comparison sequence, and calculate the similarity with the quantum effect characteristic curve in the same time length. The similarity calculation uses dynamic time warping algorithm, which allows nonlinear alignment of time axis, and finds the optimal matching path by constructing cumulative distance matrix. When the normalized distance of the optimal path is less than a certain threshold, it is determined that the change rule is consistent, and the threshold is determined by historical data verification. The data of quantum effect characteristic curve comes from the output results of quantum state analysis method in typical fault scenarios.

[0140] The real-time threshold comparison mechanism is used to determine whether the chaos index exceeds the preset chaos critical threshold. The chaos critical threshold is set to a specific value, for example, 0.75, which represents the critical point at which the fluid motion enters a strong turbulent state. The judgment logic is as follows: take the average value of the chaos index calculation results of the latest specific number of times, and determine that the critical value is exceeded when the average value is greater than the threshold. The average value calculation excludes the interference of abnormal points, and if the deviation of a certain value from the previous value exceeds a certain proportion, it is considered as an abnormal point and discarded. The method for obtaining the chaos critical threshold is as follows: collect multiple groups of data under simulated fault conditions in the laboratory, take the high percentile of the chaos index distribution as the reference value, and multiply it by a safety factor to determine the final threshold.

[0141] When the fidelity decay characteristic meets the variation law of the quantum effect characteristic curve and the chaos index exceeds the preset chaos critical threshold, a stop command is sent to the filling station main controller. The generation of the stop command needs to meet two conditions at the same time: the normalized distance of the time series matching algorithm is less than a specific threshold and the average value of the chaos index is greater than a specific critical value. After the command is generated, it is packaged into a communication protocol data frame, which includes device address code, function code, instruction code and check code. It is transmitted to the specified register address of the filling station main controller through industrial Ethernet. The main controller executes the preset stop program immediately after receiving the instruction code.

[0142] When the fidelity decay characteristic does not meet the variation law of the quantum effect characteristic curve or the chaos index does not exceed the preset chaos critical threshold, a resume main power supply instruction is sent to the power supply switching device. The generation condition of the resume instruction is that the normalized distance of the time series matching algorithm is greater than or equal to a specific threshold or the average value of the chaos index is less than or equal to a specific critical value. The data format of the instruction is the same as that of the stop instruction, and a specific instruction code is set. After receiving the instruction, the power supply switching device first detects the stability of the main power grid voltage, and when the voltage fluctuation is less than the allowed range and lasts for a certain time, it performs the switching operation. The switching process uses the first-on and then-off method to ensure uninterrupted power supply for the video monitoring system.

[0143] The construction of the quantum effect characteristic curve is realized through the following process: using a quantum image sensor to collect infrared photon intensity sequences of typical industrial fault scenes under standard light source conditions. The standard light source condition refers to a test environment with stable radiation source temperature. The typical industrial fault scenes include two types: pipeline rupture scene simulating pipeline pressure drop state; pump valve failure scene simulating mechanical seal leakage state. Multiple infrared photon intensity sequences are collected for each scene, and each sequence lasts for a certain time.

[0144] According to the infrared photon intensity sequence, the typical decay law of quantum state fidelity with time is analyzed. The analysis method is the same as the quantum state evolution trajectory calculation method in step S4: convert the photon intensity sequence into a state description matrix sequence, and calculate the fidelity decay rate in a fixed time interval. Each sequence generates a decay rate curve, which is used for subsequent analysis.

[0145] The change pattern of the typical attenuation law in a fixed time interval is constructed as a quantum effect characteristic curve. The construction method is: time alignment processing is performed on multiple attenuation rate curves, and the statistical value of the attenuation rate at each time point is taken as the characteristic curve reference value. The fixed time interval is set to a specific length, and the curve data is stored in a specific density array format. The characteristic curve is attached with confidence interval data, and the upper limit of the confidence interval is the high percentile of the attenuation rate at each time point, and the lower limit is the low percentile.

[0146] The data preprocessing in the fidelity comparison process includes trend item elimination. The current fidelity attenuation rate sequence is subjected to difference processing, and after removing the low-frequency trend component, it is compared with the characteristic curve. The mean value of the processed data is zero, which enhances the matching degree of high-frequency change characteristics.

[0147] The environmental compensation of the chaos index judgment considers the influence of fluid characteristics. The chaos critical threshold is corrected through the slurry density sensor reading, and the threshold is adjusted in proportion when the density changes. The correction parameter is determined through rheological experiment, and a parameter reference table is established and stored in the control system.

[0148] A safety interlocking mechanism is set before sending the shutdown instruction. After meeting the double conditions, it needs to be continuously monitored for a certain time, and if any condition is not met during this period, the instruction sending is cancelled. The instruction sending activates multi-level confirmation at the same time: the main controller returns the receiving confirmation; the actuator feedbacks the state signal; the video analysis system verifies the device state change.

[0149] The execution of the recovery main power supply instruction includes power quality monitoring. After switching, the main power supply loop parameters are monitored in real time, including voltage harmonic distortion rate, frequency fluctuation and phase imbalance. After continuous monitoring for a certain time, the system state is switched to normal operation mode.

[0150] Periodic updating mechanism of quantum effect characteristic curve. Periodically reacquire typical fault scene data, and when the similarity between the newly constructed characteristic curve and the old curve is lower than a certain threshold, update the curve data in the database. The update process is selected during the device maintenance period.

[0151] Redundant design of instruction transmission channel. In addition to the main communication channel, a backup channel is set. When the main channel fails to transmit continuously for a certain number of times, the channel is automatically switched, and a communication check instruction is triggered at the same time.

[0152] Deep analysis of chaos index over-limit event. Each time the chaos index exceeds the threshold, the fluid motion backtracking analysis is automatically started: the velocity field data of the previous certain time is extracted, the turbulent flow starting position is located through the recognition algorithm, and the analysis report is generated and stored in the knowledge base.

[0153] Data synchronization after power recovery. After main power recovery, all analysis data during backup power supply period will be transmitted to central server in batch. Transmission adopts breakpoint resume protocol, and transmission progress will be automatically saved when network is interrupted.

[0154] Scene matching optimization of quantum effect characteristic curve. Automatically select corresponding curve according to current fault type: use pump valve failure curve when vibration sensor triggers; use pipeline rupture curve when pressure sensor triggers. Matching process is realized through fault diagnosis result.

[0155] Short-term prediction function of chaos index. Based on recent multiple chaos index values, estimate the trend of index change at a specific time in the future through a prediction model. When the predicted value exceeds the threshold, an early warning signal is sent. The prediction model parameters are recalibrated regularly.

[0156] Timestamp alignment mechanism for all decision-making processes. A high-precision clock source is used to mark a unified time for key events: video frame acquisition, fidelity calculation completion, chaos index output, and instruction sending. The time synchronization error is controlled within the allowed range.

[0157] Closed-loop verification of instruction execution effect. After sending the shutdown instruction, verify that the industrial pump speed decreases to the safe range within a specific time; after sending the restore power supply instruction, verify that the main power supply circuit current reaches the normal working range within a specific time. If the verification fails, the backup safety protocol is started.

[0158] Visual debugging interface for decision logic. Engineers can view the following through the human-machine interface in real time: matching degree display of fidelity decay sequence and characteristic curve, history trend of chaos index and threshold line, and identification of double condition satisfaction. The interface is refreshed regularly, and supports historical data playback analysis.

[0159] S6, execute the shutdown instruction or the restore main power supply instruction, the specific implementation includes:

[0160] When receiving the shutdown instruction, control the filling station main controller to execute the industrial pump shutdown operation. The specific process is as follows: after the filling station main controller receives the shutdown instruction data frame from the S5 step, it analyzes the instruction code to confirm its validity, and then executes a multi-stage shutdown program. The first stage linearly reduces the power supply frequency of the industrial pump drive motor to a safe speed within a specific time, and the linear reduction curve slope is calculated by a preset function. The second stage closes the feed valve, and the valve closing rate is set to a specific value. The valve opening degree is fed back in real time through a position sensor, and when the opening degree is less than a specific percentage, the third stage operation is triggered. The third stage cuts off the main motor power supply and activates the mechanical brake, and the brake clamping force increases in segments over time. The entire shutdown process is completed within a specific time, and the pump shaft speed is verified to be reduced below the safety threshold through the speed sensor during the process.

[0161] The standby power supply maintains the power supply of the video monitoring system while performing the shutdown operation. The standby power supply maintenance mechanism includes: continuously monitoring the output voltage of the uninterruptible power supply, and automatically enabling the redundant power supply unit when the voltage fluctuation exceeds the allowed range; real-time calculation of the video monitoring system load power, when the power exceeds a certain proportion of the rated capacity of the standby power supply, automatically shut down non-critical function units to reduce power consumption. The power supply maintenance state records power supply parameters including output voltage, load current and remaining power percentage periodically, and the data is stored in a special log.

[0162] Recording the shutdown state includes multi-dimensional data acquisition. The shutdown transient vibration waveform is collected by a vibration sensor, the sampling frequency is set to a certain value, and the recording is continued for a certain time. At the same time, the change curve of the quantum state fidelity attenuation characteristic and the chaos index in S4 step before and after the shutdown is recorded, and the data time range covers a certain time before the shutdown instruction is triggered to a certain time after the shutdown is completed. All data are time stamped to form a complete data package of the shutdown event.

[0163] When receiving the main power supply recovery instruction, the power supply switching device switches the video monitoring system to the main power supply loop. The switching operation execution process is: after the power supply switching device receives the recovery instruction of S5 step, the stability of the three-phase voltage of the main power grid is detected, and the voltage deviation, frequency deviation and harmonic distortion rate are required to be less than the allowed value. After the detection is qualified for a certain time, the main power supply contactor is closed first, and then the standby power supply contactor is disconnected after a certain delay. Real-time monitoring of the voltage of the video monitoring system power bus during the switching process, when the voltage drop exceeds a certain proportion or the duration exceeds the threshold, automatically switch back to the standby power supply and generate an alarm.

[0164] The gradual startup strategy is adopted to restore the normal operation state of the filling station controller. First, the communication function of the controller is restored, and the connection with the upper system is established and the state data is synchronized. Then the data acquisition channels are started step by step: the pressure sensor channel is enabled after a certain delay, the flow sensor channel is enabled after a certain delay, and the vibration sensor channel is enabled after a certain delay. Finally, the actuator control authority is restored, and the control mode is set to automatic operation state, and the initial set value is a certain proportion of the running parameter before shutdown, and it is linearly restored to the normal value within a certain time.

[0165] Generating a power supply loop recovery log includes recording electrical parameters and system state. The electrical parameter records the voltage drop depth, recovery time and switching impact current peak value at the moment of main power supply recovery. The system state records the time consumed by the controller to start each function, the deviation value of the sensor channel recalibration, and the position reset accuracy of the actuator. The log data is arranged in chronological order, a trend chart is generated, and the time marker of the switching operation is attached.

[0166] The power supply optimization of the video monitoring system in shutdown state includes dynamic frame rate adjustment. When the remaining power of the backup power supply is lower than a certain proportion, the infrared video frame rate is automatically reduced. The frame rate switching adopts a smooth transition mode to avoid picture jumping. At the same time, the video compression rate is improved, and the code stream control adopts the constant quality mode.

[0167] The electrical protection design of the power supply switching device includes multiple protections. The main circuit is provided with a current limiting protection device with a rated current of a certain multiple of the maximum load current. A transient voltage suppression device is installed at the output end of the backup power supply. The control signal circuit adopts isolation technology. All metal shells are grounded through wires.

[0168] The data storage of the shutdown event adopts differential encoding technology. The data of a certain time before and after the shutdown moment is stored completely as a key frame, and the differential data of the remaining period is stored periodically. The storage format includes a time reference area, a sensor data area, a video metadata area, and a check code area.

[0169] The system self-check after the main power supply is restored includes multiple stages. The first stage detects the communication state of all sensors. The second stage performs sensor zero point calibration: the pressure sensor verifies the zero point drift in the no pressure state; the flow sensor verifies the output signal at zero flow rate. The third stage performs actuator stroke test: the valve executes opening degree to and fro movement; the pump motor runs at no load to rated speed.

[0170] The generation rule of the power supply circuit recovery log includes intelligent classification. According to the key indicators score of the switching process, the voltage stability accounts for a certain weight in the scoring algorithm, the switching time accounts for a certain weight, and the impact current accounts for a certain weight. The log file header includes the score result and the explanation.

[0171] The priority management of the controller state recovery adopts a task scheduling algorithm. The start task is divided into multiple priority queues: communication recovery and safety monitoring belong to the highest priority; data acquisition channel start belongs to the medium priority; historical data synchronization and report generation belong to the ordinary priority. The task scheduler scans the ready queue regularly.

[0172] Picture quality guarantee technology in the switching process of the video monitoring system. During the transition period of the contactor switching, the frame buffer compensation mechanism is enabled: the last multiple frames of video before switching are stored in the buffer, and if the video stream interruption is detected, the transition frame is generated by interpolation using the buffer frame. This mechanism ensures that the video stream interruption time does not exceed a certain value.

[0173] Cooperative control of shutdown operation and power supply maintenance. When executing the shutdown instruction, the video monitoring system immediately starts the event recording mode: the video encoding code rate is increased to a certain multiple of the baseline value; at the same time, unnecessary functions are turned off. This mode is maintained until the main power supply is restored after a certain time and automatically exits.

[0174] The digital signature of the recovery log is designed to be tamper-proof. After generating the log file each time, the file feature value is processed with an encryption algorithm to generate a digital signature. The signature is attached at the end of the file. During verification, the feature value is recalculated and compared. If they are inconsistent, an alarm is triggered.

[0175] Fallback strategy for power supply switching failure. When consecutive switching attempts fail for a certain number of times, automatically switch to an emergency solution: maintain backup power supply while starting the backup power generation equipment. The grid connection process is as follows: close the switch after detecting phase synchronization of the output voltage, and complete the power transfer within a certain time.

[0176] Energy efficiency optimization during system recovery phase. After the main power supply is restored, gradually reduce the charging current of the backup power supply within a certain time. Monitor the battery temperature during the charging process. If the temperature exceeds a certain value, suspend charging and start the cooling function.

[0177] Automatic generation of shutdown event analysis report. After each shutdown operation is completed, the system extracts key data to generate a report: including the shutdown process speed curve, valve action timing diagram, vibration spectrum comparison, power supply parameter statistical table, and video key frame screenshot. The report is automatically uploaded through a secure link.

[0178] Embodiment 2: Figure 2 The structure diagram of the filling station control system based on video monitoring is given, and the filling station control system based on video monitoring comprises:

[0179] The signal acquisition module is configured to acquire the sensor abnormal signal of the filling station device and the real-time video stream of the video monitoring system.

[0180] The signal judgment module is configured to maintain the current state when the sensor abnormal signal does not exceed the preset threshold, and otherwise generate a backup judgment signal.

[0181] The entropy increase analysis module is configured to, after generating the backup judgment signal, calculate the pixel entropy value of the fault area based on the real-time video stream to generate a thermodynamic entropy increase trajectory, and start the backup power supply to independently supply power to the video monitoring system when the slope of the entropy increase trajectory is greater than a set threshold.

[0182] The chaos analysis module is configured to, in the backup power supply state, analyze the fidelity decay characteristic of the quantum state evolution trajectory based on the real-time video light quantum distribution characteristic, and reconstruct the turbulent dynamics characteristic based on the real-time video flow motion characteristic and extract the chaos index.

[0183] The instruction generation module is configured to generate a shutdown instruction when the fidelity decay characteristic meets the quantum effect characteristic curve and the chaos index reaches a preset chaos critical threshold, and otherwise generate a main power supply recovery instruction.

[0184] The instruction execution module is configured to execute the shutdown instruction or the main power supply recovery instruction.

[0185] The calculations involved in the embodiments are all dimensionless numerical calculations, and the preset parameters and threshold values in the calculations are set by a person skilled in the art according to actual conditions.

[0186] It should be noted that the application can be deployed on a device itself to realize embedded applications, or run on a PC terminal or other terminal with a user interface, thereby meeting various hardware environments and use requirements.

[0187] The above embodiments can be realized wholly or partially by software, hardware, firmware or any other combination. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another through wireless or wired transmission. The wired transmission includes optical fiber, twisted pair, coaxial cable and the like; the wireless transmission includes infrared rays, microwaves and the like. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center and the like containing one or more available medium collections. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD) or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0188] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, device and module can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0189] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and actual implementation can have another division manner, for example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed ones can be indirect coupling or communication connection through some interfaces, devices or modules, and can be electrical, mechanical or other forms.

[0190] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed on multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment of the present application.

[0191] In addition, the functional modules in each embodiment of the present application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0192] The functions, if realized in the form of software functional modules and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part of the prior art or the part of the technical solutions of the present application can be embodied in the form of software products, and the computer software product is stored in a storage medium, including a plurality of instructions for making a computer device (which can be a personal computer, a server or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk and various program code storage media.

[0193] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be limited by the protection scope of the claims.

[0194] Finally: the above is only the preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application, should be included in the protection scope of the present application.

Claims

1. A method for controlling a filling station based on video surveillance, characterized in that Comprising: S1, acquiring sensor abnormal signal of filling station equipment and real-time video stream of video monitoring system; S2, maintaining the current state when the sensor abnormal signal does not exceed the preset threshold, otherwise generating a backup judgment signal; S3, after generating the backup judgment signal, calculating the pixel entropy value of the fault area based on the real-time video stream to generate the thermodynamic entropy increase trajectory: locating the spatial position of the fault area in the real-time video stream according to the sensor abnormal signal type corresponding to the backup judgment signal: extracting the pixel gray value of the continuous multiple video frames in the fault area, and calculating the pixel entropy value of each video frame; Connecting the pixel entropy value in the video frame time sequence to form the thermodynamic entropy increase trajectory; When the slope of the entropy increase trajectory is greater than the set threshold, starting the backup power supply to independently supply power to the video monitoring system; S4, under the backup power supply state, analyzing the fidelity decay characteristics of the quantum state evolution trajectory based on the real-time video quantum distribution characteristics, and reconstructing the turbulent dynamics characteristics and extracting the chaos index based on the real-time video stream motion characteristics, including: Under the backup power supply state, collecting infrared photon intensity sequence of the real-time video stream of the fault area to form sequence data of the change of photon intensity with time; Determine the change process of the quantum state probability distribution according to the photon intensity sequence data, and generate the quantum state evolution trajectory; Quantize the state maintenance ability change value of the quantum state evolution trajectory in a fixed time interval as the fidelity decay characteristics; At the same time, the real-time video stream of the fault area is analyzed for pixel displacement to obtain the fluid motion velocity distribution; Reconstruct the fluid dynamic behavior characteristics according to the fluid motion velocity distribution; Calculate the chaos degree quantization value of the fluid dynamic behavior characteristics as the chaos index; S5, when the fidelity decay characteristics meet the quantum effect characteristic curve and the chaos index reaches the preset chaos critical threshold, generate the shutdown instruction; otherwise, generate the recovery main power supply instruction; The quantum effect characteristic curve is obtained by: using a quantum image sensor to collect infrared photon intensity sequence of a typical industrial fault scene under standard light source conditions; analyzing the typical decay law of quantum state fidelity with time according to the infrared photon intensity sequence; and constructing the change mode of the typical decay law in a fixed time interval as the quantum effect characteristic curve; S6, execute the shutdown instruction or the recovery main power supply instruction.

2. The video monitoring based filling station control method according to claim 1, characterized in that, Acquiring sensor abnormal signal of filling station equipment and real-time video stream of video monitoring system, comprising: Acquiring the pipe pressure sensor signal, slurry flow sensor signal and pump body vibration sensor signal of the industrial pump as the sensor abnormal signal; Acquiring the monitoring video stream corresponding to the pipe pressure sensor signal at the filling pipe connection, and the monitoring video stream corresponding to the pump body vibration sensor signal at the pump valve interface as the real-time video stream.

3. The video monitoring based filling station control method according to claim 1, characterized in that, When the sensor abnormal signal does not exceed the preset threshold, maintaining the current state, otherwise generating a backup judgment signal, including: Judging whether the pipe pressure sensor signal of the industrial pump exceeds the pipe pressure mutation rate threshold; Judging whether the slurry flow sensor signal exceeds the flow deviation threshold; Judging whether the pump body vibration sensor signal exceeds the vibration frequency energy threshold; A standby judgment signal is generated when the pipeline pressure sensor signal exceeds the pipeline pressure mutation rate threshold, or the slurry flow sensor signal exceeds the flow deviation threshold, or the pump body vibration sensor signal exceeds the vibration frequency domain energy threshold; otherwise, the current operation state of the filling station is maintained.

4. The video monitoring based filling station control method according to claim 1, characterized in that, After the standby judgment signal is generated, the standby power supply is started to independently supply power to the video monitoring system when the entropy increase trajectory slope is greater than a set threshold, including: calculating the average slope of the thermodynamic entropy increase trajectory within a recent preset number of video frame time windows; when the average slope is greater than the entropy increase slope threshold, sending a start instruction to the uninterruptible power supply to independently supply power to the video monitoring system.

5. The video monitoring based filling station control method according to claim 4, characterized in that, When the standby judgment signal is triggered by the pipeline pressure sensor signal, the fault area is the pipeline connection area in the monitored video stream at the filling pipeline connection; when the standby judgment signal is triggered by the pump body vibration sensor signal, the fault area is the pump valve sealing area in the monitored video stream at the pump valve interface.

6. The video monitoring based filling station control method according to claim 1, characterized in that, A shutdown instruction is generated when the fidelity decay characteristic meets the quantum effect characteristic curve and the chaos index reaches a preset chaos critical threshold; otherwise, a main power supply recovery instruction is generated, including: comparing the consistency of the change law of the fidelity decay characteristic and the quantum effect characteristic curve; judging whether the chaos index exceeds the preset chaos critical threshold; when the fidelity decay characteristic meets the change law of the quantum effect characteristic curve and the chaos index exceeds the preset chaos critical threshold, sending a shutdown instruction to the filling station main controller; when the fidelity decay characteristic does not meet the change law of the quantum effect characteristic curve or the chaos index does not exceed the preset chaos critical threshold, sending a main power supply recovery instruction to the power supply switching device.

7. The video monitoring based filling station control method according to claim 1, characterized in that, Executing the shutdown instruction or the main power supply recovery instruction, including: when the shutdown instruction is received, controlling the filling station main controller to perform an industrial pump shutdown operation, while maintaining the standby power supply to the video monitoring system and recording the shutdown state; when the main power supply recovery instruction is received, controlling the power supply switching device to switch the video monitoring system to the main power supply loop, while restoring the normal operation state of the filling station main controller and generating a power supply loop recovery log.

8. A video surveillance based filling station control system for implementing the video surveillance based filling station control method according to any one of claims 1 - 7, characterized by including: a signal acquisition module for acquiring sensor abnormal signals of filling station equipment and real-time video streams of a video monitoring system; a signal judgment module for maintaining the current state when the sensor abnormal signal does not exceed the preset threshold, otherwise generating a standby judgment signal; an entropy increase analysis module for generating a thermodynamic entropy increase trajectory based on pixel entropy values of a fault area calculated from real-time video streams after the standby judgment signal is generated, and starting the standby power supply to independently supply power to the video monitoring system when the entropy increase trajectory slope is greater than a set threshold; a chaos analysis module for analyzing the fidelity decay characteristic of a quantum state evolution trajectory based on real-time video quantum distribution characteristics under the standby power supply state, and reconstructing the turbulent dynamics characteristics based on real-time video flow characteristics and extracting the chaos index; an instruction generation module for generating a shutdown instruction when the fidelity decay characteristic meets the quantum effect characteristic curve and the chaos index reaches a preset chaos critical threshold; otherwise, generating a main power supply recovery instruction; an instruction execution module for executing the shutdown instruction or the main power supply recovery instruction.

Images