Safety risk early warning method, device, equipment, medium and product for petroleum catalytic cracking device

Abstract

The invention discloses a safety risk early warning method, device and equipment for a petroleum catalytic cracking device, a medium and a product. The method comprises the following steps: acquiring multi-modal data of a petroleum catalytic cracking device, the multi-modal data comprising temperature, pressure, circulation rate and oxygen content; determining a reconstruction error, an energy balance index, an oxygen balance coefficient and a change rate correlation coefficient corresponding to the multi-modal data; if the reconstruction error, the energy balance index, the oxygen balance coefficient and the change rate correlation coefficient corresponding to the multi-modal data meet coupling fault alarm conditions, coupling fault alarm is carried out, and by means of the technical scheme, the safety and reliability of petrochemical industry production can be improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of petrochemical technology, and in particular to a safety risk early warning method, device, equipment, medium and product for a petroleum catalytic cracking unit. Background Art

[0002] The catalytic cracking unit (FCC) is a core production unit in the petroleum refining and chemical industry. Characterized by high temperatures, high pressures, complex processes, and a high degree of continuity, it also involves a large number of flammable, explosive, toxic, and hazardous media, posing a high safety risk. During operation, the unit is susceptible to safety incidents such as leaks, fires, and explosions due to factors such as process fluctuations, equipment aging, and operational errors. Existing petrochemical safety early warning algorithms have been widely used, but they typically have fixed warning parameter settings and are unable to dynamically adjust them in the presence of multiple coupled parameters. Consequently, they suffer from delayed response times and weak early warning capabilities, making them difficult to meet the inherent safety requirements of modern plants.

[0003] Therefore, building a scientific and efficient safety risk early warning method can achieve real-time monitoring and trend analysis of key parameters, identify potential risks in advance, and improve accident prevention and control capabilities, which is of great significance to ensuring the safe and stable operation of catalytic cracking units. Summary of the Invention

[0004] The embodiments of the present invention provide a safety risk early warning method, device, equipment, medium and product for a petroleum catalytic cracking unit, which can improve the safety and reliability of petrochemical production.

[0005] According to one aspect of the present invention, a safety risk early warning method for a petroleum catalytic cracking unit is provided, comprising:

[0006] Acquiring multimodal data of a petroleum catalytic cracking unit, wherein the multimodal data includes: temperature, pressure, circulation rate, and oxygen content;

[0007] Determining a reconstruction error, an energy balance index, an oxygen balance coefficient, and a rate of change correlation coefficient corresponding to the multimodal data;

[0008] If the reconstruction error, energy balance index, oxygen balance coefficient and change rate correlation coefficient corresponding to the multimodal data meet the coupling fault alarm condition, a coupling fault alarm is issued.

[0009] According to another aspect of the present invention, a safety risk early warning device for a petroleum catalytic cracking unit is provided, the device comprising:

[0010] A multimodal data acquisition module, configured to acquire multimodal data of a petroleum catalytic cracking unit, wherein the multimodal data includes temperature, pressure, circulation rate, and oxygen content;

[0011] a determination module, configured to determine a reconstruction error, an energy balance index, an oxygen balance coefficient, and a change rate correlation coefficient corresponding to the multimodal data;

[0012] The early warning module is configured to issue a coupling fault alarm if the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data meet coupling fault alarm conditions.

[0013] According to another aspect of the present invention, an electronic device is provided, comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the safety risk early warning method for a petroleum catalytic cracking unit described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the safety risk warning method for a petroleum catalytic cracking unit described in any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, a computer program product is provided. When executed by a processor, the computer program implements the safety risk early warning method for a petroleum catalytic cracking unit as described in any one of the embodiments of the present invention.

[0019] The embodiment of the present invention first obtains multimodal data of a petroleum catalytic cracking unit; then determines a reconstruction error, an energy balance index, an oxygen balance coefficient, and a change rate correlation coefficient corresponding to the multimodal data; and finally, when the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data meet coupling fault alarm conditions, performs a coupling fault alarm. This enables early source tracing and early warning, achieves the purpose of active prevention and control, and further improves the safety and reliability of petrochemical production.

[0020] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a flow chart of a safety risk early warning method for a petroleum catalytic cracking unit according to an embodiment of the present invention;

[0023] Figure 2 This is a schematic structural diagram of a security risk early warning system according to an embodiment of the present invention;

[0024] Figure 3 is a structural diagram of a main control module in an embodiment of the present invention;

[0025] Figure 4 is a schematic structural diagram of a sensor in an embodiment of the present invention;

[0026] Figure 5 is a structural diagram of a data processing module in an embodiment of the present invention;

[0027] Figure 6 is a schematic structural diagram of an environment adaptation module in an embodiment of the present invention;

[0028] Figure 7 This is a schematic structural diagram of an early warning execution module in an embodiment of the present invention;

[0029] Figure 8 is a schematic structural diagram of a power supply in an embodiment of the present invention;

[0030] Figure 9 is a schematic structural diagram of a communication module in an embodiment of the present invention;

[0031] Figure 10 This is a schematic structural diagram of a safety risk early warning device for a petroleum catalytic cracking unit according to an embodiment of the present invention;

[0032] Figure 11 It is a structural diagram of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0034] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] It is understandable that before using the technical solutions disclosed in the various embodiments of this disclosure, the type, scope of use, usage scenarios, etc. of the personal information involved in this disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0036] Example 1

[0037] Figure 1 This is a flow chart of a safety risk warning method for a petroleum catalytic cracking unit provided in an embodiment of the present invention. This embodiment is applicable to situations where a safety risk warning is provided for a petroleum catalytic cracking unit. This method can be executed by a safety risk warning device for a petroleum catalytic cracking unit in an embodiment of the present invention. The device can be implemented in software and / or hardware. Figure 1 As shown, the method specifically includes the following steps:

[0038] S110, acquiring multimodal data of a petroleum catalytic cracking unit.

[0039] In this embodiment, the multimodal data includes temperature, pressure, circulation rate, and oxygen content. The temperature refers to the reactor temperature, the pressure refers to the regenerator pressure, and the circulation rate refers to the catalyst circulation rate. In the catalytic cracking process, the circulation rate specifically refers to the mass flow rate of the catalyst circulating between the reactor and the regenerator per unit time and is a key control parameter for maintaining reaction equilibrium. Where C is the cycle rate, Δm cis the mass of catalyst circulating in the time interval Δt. Oxygen content is the oxygen content of the flue gas.

[0040] In this embodiment, for the scenario of cracking heavy oil into light fuel, the normal range of reactor temperature is 480-520°C, the normal range of regenerator pressure is 0.25-0.35 MPa, the normal range of catalyst circulation rate is 800-1200 tons / hour, and the normal range of flue gas oxygen content is 1.5-3.0 vol%.

[0041] In this embodiment, the parameter coupling relationship can be: temperature ↑ → reaction rate ↑ → catalyst consumption ↑ → circulation rate needs to be synchronized ↑; circulation rate abnormality ↑ → regenerator pressure drop ↑ → pressure fluctuation ↑; oxygen content ↓ → incomplete combustion → reaction temperature ↓.

[0042] S120, determining a reconstruction error, an energy balance index, an oxygen balance coefficient, and a change rate correlation coefficient corresponding to the multimodal data.

[0043] In this embodiment, the reconstruction error corresponding to the multimodal data may be determined by calculating the reconstruction error using the Adaptive Windowing (ADWIN) algorithm in Concept Drift Detection. For example, the multimodal data may be input into an autoencoder to obtain reconstructed data, and the reconstruction error may be determined based on the difference between the reconstructed data and the multimodal data.

[0044] In a specific example, the reconstruction error is determined based on the following formula:

[0045]

[0046] Where,∈ is the reconstruction error, the T t is the measured temperature corresponding to time t (temperature in multimodal data), is the reconstruction temperature corresponding to time t; P t is the measured pressure corresponding to time t (pressure in multimodal data), is the reconstruction pressure corresponding to time t, C t is the measured cycle rate corresponding to time t (cycle rate in multimodal data), is the reconstruction cycle rate corresponding to time t, O 2t is the measured oxygen content corresponding to time t (oxygen content in multimodal data), is the reconstructed oxygen content corresponding to time t.

[0047] In this embodiment, the energy balance index is equal to the ratio of the product of temperature and circulation rate to pressure. For example, the energy balance index can be determined based on the following formula: Among them, E bal is the energy balance index, T is the temperature, C is the circulation rate, and P is the pressure.

[0048] In this embodiment, the oxygen balance coefficient is determined according to the pressure, circulation rate, and oxygen content. For example, the oxygen balance coefficient may be determined based on the following formula: Among them, O2 is the oxygen content.

[0049] In this embodiment, the change rate correlation coefficient is determined based on the temperature and the cycle rate. For example, the change rate correlation coefficient can be determined based on the following formula: in, is the derivative of temperature with respect to time, is the derivative of the cycle rate and time, and corr() is a function used to calculate the correlation coefficient.

[0050] In this embodiment, the reconstruction error corresponding to the multimodal data is determined by the difference between the multimodal data and the reconstructed data of the multimodal data, the reconstructed data of the multimodal data is obtained by inputting the multimodal data into a concept drift detection model, the energy balance index is equal to the ratio of the product of temperature and circulation rate to pressure, the oxygen balance coefficient is equal to the product of oxygen content and a characteristic parameter, the characteristic parameter is equal to the square root of the ratio of pressure to circulation rate, and the change rate correlation coefficient is equal to the correlation coefficient between the derivative of temperature and time and the derivative of circulation rate and time.

[0051] S130: If the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data meet coupling fault alarm conditions, a coupling fault alarm is issued.

[0052] In this embodiment, the coupling fault alarm conditions include: the reconstruction error corresponding to the multimodal data is greater than the target value, wherein the target value is determined by the mean and standard deviation of the target reconstruction error; the energy balance index is less than the energy balance index threshold; the oxygen balance coefficient is greater than the first value, or less than the second value, wherein the second value is less than the first value; the rate of change correlation coefficient is less than the third value, or greater than the fourth value, wherein the third value is less than the fourth value.

[0053] Optionally, the method further includes: if the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data do not meet the coupling fault alarm condition, determining the detection rules corresponding to each modal data, and determining the detection result of the petroleum catalytic cracking unit based on the detection rules and each modal data, wherein the reconstruction error corresponding to the multimodal data is determined by the difference between the multimodal data and the reconstructed data of the multimodal data, and the reconstructed data of the multimodal data is obtained by inputting the multimodal data into a concept drift detection model; the energy balance index is equal to the ratio of the product of temperature and circulation rate to pressure; the oxygen balance coefficient is equal to the product of oxygen content and characteristic parameter; the characteristic parameter is equal to the square root of the ratio of pressure to circulation rate; and the change rate correlation coefficient is equal to the correlation coefficient between the derivative of temperature and time and the derivative of circulation rate and time;

[0054] The coupling fault alarm condition includes: a reconstruction error corresponding to the multimodal data is greater than a target value, wherein the target value is determined by a mean and a standard deviation of the target reconstruction error.

[0055] In this embodiment, the target value is determined based on the formula: target value = μ + 3σ, where μ is the mean of the target reconstruction error and σ is the standard deviation of the target reconstruction error. The target reconstruction error is determined based on historical normal data, the mean of the target reconstruction error can be the mean of the reconstruction error within a sliding window, and the standard deviation of the target reconstruction error can be the standard deviation of the reconstruction error within the same window. The window size can be 30 minutes.

[0056] The energy balance index is less than the energy balance index threshold.

[0057] In this embodiment, the energy balance index threshold may be a preset value, for example, the energy balance index threshold may be 1.2×10 6 .

[0058] The oxygen balance coefficient is greater than a first value, or less than a second value, wherein the second value is less than the first value.

[0059] In this embodiment, the normal range of the oxygen balance coefficient is greater than the second value and less than the first value. Therefore, if the oxygen balance coefficient is greater than the first value or less than the second value, the oxygen balance coefficient is determined to be abnormal. The first value is a preset value, for example, the first value may be 0.25, and the second value is a preset value less than the first value, for example, the second value may be 0.15.

[0060] The change rate correlation coefficient is smaller than a third value or larger than a fourth value, wherein the third value is smaller than the fourth value.

[0061] In this embodiment, the normal range of the change rate correlation coefficient is greater than the third value and less than the fourth value. The third value can be a pre-set value greater than the first value, for example, the third value can be 0.6, and the fourth value can be a pre-set value greater than the third value, for example, the fourth value can be 0.8.

[0062] In a specific example, the fault detection process is analyzed as follows: t = 30s (early fault); single parameter: only temperature / oxygen content slightly exceeds the limit (yellow alarm); normal. Slightly abnormal. TC = 0.25, the correlation breaks down. The reconstruction error suddenly increases by 30% → the predicted failure probability is 72%. At t = 60s (development period); Abnormal decline. Abnormal rise. TC = -0.05, inverse correlation. Reconstruction error exceeds 4σ → failure probability 98%, triggering emergency response.

[0063] System response measures include automatic adjustments and alarm grading. Automatic adjustments include reducing the feed rate by 20% and activating a backup catalyst delivery line. Alarm grading includes: console: audible and visual alarms with fault location map; mobile phone: push notifications of parameter coupling anomalies.

[0064] Cause analysis: Temperature↓→catalyst circulation↑→pressure↑→oxygen sensor fluctuation→control system misadjustment→circulation rate out of control.

[0065] In a specific example, the input data of the adaptive sliding window algorithm (ADWIN) is a real-time temperature sequence T = [T1, T2, ..., T n]. The initial window W contains the first 1000 samples (about 16 minutes of data). Divide the window W into two sub-windows W0 (old data) and W1 (new data). Calculate the mean μ0 and μ1 and variance σ02 and σ12 of W0 and W1. Determine whether |μ0-μ1|> the significance threshold. The significance threshold can be calculated by the Hoeffding inequality. The Hoeffding inequality is used to describe the upper bound of the probability of deviation between the sample mean of an independent random variable and the true mean. For example, in the normal stage: the mean of W0 is 65°C, the mean of W1 is 66°C → no significant difference. In the aging stage: the mean of W0 is 68°C, the mean of W1 is 72°C → a difference is detected. If |μ0-μ1|> the significance threshold, it is determined that concept drift has occurred, the old data W0 is deleted, and W1 is retained as the starting point of the new window. If there is no drift, the window is gradually expanded to include new data. When ADWIN detects drift, the following actions are triggered: collect recent data (such as the latest 1-hour window). Incrementally train the model: Use online gradient descent on the autoencoder to fine-tune the encoder layer weights. Reset dynamic thresholds: Compute the mean and standard deviation of the reconstruction error based on new data.

[0066] It should be noted that the simulation data generation process includes: Phase 1 (normal): In the first 8 hours, the temperature obeys N(65,5 2 Phase 2 (gradual drift): Over the next 8 hours, the average temperature rises linearly to 75°C (simulating device aging). Phase 3 (abrupt drift): A fault is suddenly introduced, causing the temperature to jump to 90°C.

[0067] ADWIN detection results: 1. Gradual drift detection: ADWIN detected drift when the mean temperature rose to 70°C (approximately the 10th hour), shortening the window from 1000 to 300 samples. This triggered an incremental model update, adjusting the autoencoder reconstruction error threshold from 65±15 to 70±20. 2. Sudden drift detection: When the temperature jumped to 90°C, ADWIN detected drift within 10 seconds, triggering emergency full model training.

[0068] Optionally, the multimodal data further includes: an infrared image, the infrared image being used to characterize the surface temperature distribution of a storage tank, and the petroleum catalytic cracking unit includes: a storage tank;

[0069] Determining a detection rule corresponding to the infrared image, and determining a detection result of the storage tank based on the detection rule and the infrared image, including:

[0070] If the local hotspot temperature of the infrared image is greater than a target temperature threshold, the tank is determined to be overheated, wherein the target temperature threshold is determined in the following manner: if the temperature inside the tank does not rise, the initial temperature threshold is used as the target temperature threshold; if the temperature inside the tank rises, the difference between the initial temperature threshold and the first temperature value is used as the target temperature threshold.

[0071] In this embodiment, the temperature inside the storage tank can be acquired by a temperature sensor. The initial temperature threshold can be a preset temperature, for example, the initial temperature threshold can be 150°C.

[0072] It should be noted that if the local hotspot temperature in the infrared image exceeds the initial temperature threshold, if the temperature sensor reading inside the tank is normal, a manual review will be initiated, possibly due to interference from external impurities. If the temperature inside the tank rises simultaneously, a shutdown command will be immediately triggered, and the initial temperature threshold will be lowered.

[0073] In this embodiment, the first temperature value may be a preset temperature adjustment value, for example, the first temperature value may be 20° C. In a specific example, if the initial temperature threshold is 150° C. and the first temperature value is 20° C., the adjusted threshold is 130° C.

[0074] Optionally, the multimodal data further includes: corrosion detection data, wherein the corrosion detection data includes: pipeline wall thickness, real-time corrosion rate and chemical property parameters of the liquid in the pipeline, and the petroleum catalytic cracking unit further includes: pipeline.

[0075] In this embodiment, the chemical property parameter of the liquid in the pipeline may be the pH value of the liquid in the pipeline.

[0076] Determining a detection rule corresponding to the pipeline wall thickness, and determining a pipeline detection result based on the detection rule and the pipeline wall thickness, including: if the pipeline wall thickness is less than a target pipeline wall thickness threshold, determining that the pipeline wall thickness is abnormal, wherein the target pipeline wall thickness threshold is determined by: determining an initial pipeline wall thickness threshold based on a design wall thickness value.

[0077] In this embodiment, the initial pipe wall thickness threshold value may be determined according to the design wall thickness value by taking 80% of the design wall thickness value as the initial pipe wall thickness threshold value.

[0078] The target pipeline wall thickness threshold is determined according to the real-time corrosion rate, the chemical property parameters of the liquid in the pipeline, and the initial pipeline wall thickness threshold.

[0079] In this embodiment, the target pipeline wall thickness threshold is determined based on the real-time corrosion rate, the chemical property parameters of the liquid in the pipeline, and the initial pipeline wall thickness threshold. The method may be as follows: determining the remaining pipeline life based on the real-time corrosion rate; if the pipeline is determined to be in a stable corrosion state based on the chemical property parameters of the liquid in the pipeline, determining the target pipeline wall thickness threshold based on the attenuation frequency and the remaining pipeline life; if the pipeline is determined to be in an accelerated corrosion state based on the chemical property parameters of the liquid in the pipeline, determining the target pipeline wall thickness threshold based on the first multiplier, the attenuation frequency, and the remaining pipeline life.

[0080] Optionally, a target pipe wall thickness threshold is determined based on the real-time corrosion rate, chemical property parameters of the liquid in the pipe, and the initial pipe wall thickness threshold, including:

[0081] Determine remaining pipeline life based on real-time corrosion rates.

[0082] In this embodiment, the remaining life of the pipeline may be determined based on the real-time corrosion rate by taking the ratio of the difference between the current pipeline wall thickness and the initial pipeline wall thickness threshold to the implemented corrosion rate as the remaining life of the pipeline.

[0083] If the pipeline is determined to be in a stable corrosion state based on the chemical properties of the liquid in the pipeline, the target pipeline wall thickness threshold is determined based on the attenuation frequency, the remaining life of the pipeline, and the initial pipeline wall thickness threshold.

[0084] In this embodiment, if the chemical property parameter of the liquid in the pipeline is greater than the first pH value and less than the second pH value, it is determined that the pipeline is in a stable corrosion state.

[0085] In this embodiment, the target pipeline wall thickness threshold can be determined based on the attenuation frequency, the remaining pipeline life, and the initial pipeline wall thickness threshold by multiplying the attenuation frequency and the remaining pipeline life as the corrosion allowance, and taking the sum of the corrosion allowance and the initial pipeline wall thickness threshold as the target pipeline wall thickness threshold. The target pipeline wall thickness threshold can also be determined based on the attenuation frequency, the remaining pipeline life, and the initial pipeline wall thickness threshold by obtaining the current time and automatically adjusting the initial pipeline wall thickness threshold downward by the wall thickness corresponding to the attenuation frequency each month based on the current time to obtain the target pipeline wall thickness threshold.

[0086] In a specific example, the remaining life is dynamically calculated based on the real-time corrosion rate (eg, 0.1 mm / year), and the threshold is automatically lowered every month (eg, attenuated by 0.5%).

[0087] If the pipeline is determined to be in an accelerated corrosion state based on the chemical property parameters of the liquid in the pipeline, the target pipeline wall thickness threshold is determined based on the first multiple, the attenuation frequency, the remaining life of the pipeline, and the initial pipeline wall thickness threshold.

[0088] In this embodiment, if the chemical property parameter of the liquid in the pipeline is less than the first pH value, or greater than the second pH value, it is determined that the pipeline is in an accelerated corrosion state. The first pH value is less than the second pH value. The first pH value can be 6, and the second pH value can be 11.

[0089] In this embodiment, the first multiple is an accelerated decay multiple. It should be noted that if the pipeline is in an accelerated corrosion state, the threshold decay rate is doubled.

[0090] Optionally, the multimodal data also includes: gas concentration collected by each gas concentration sensor; determining a detection rule corresponding to the gas concentration, and determining a detection result of the petroleum catalytic cracking unit based on the detection rule and the gas concentration, including: if the gas concentration collected by at least one gas concentration sensor is greater than the target concentration threshold corresponding to the gas concentration sensor, then determining a gas leakage, wherein the method for determining the target concentration threshold corresponding to the gas concentration sensor includes: obtaining an initial concentration threshold; if the difference in gas concentration collected by at least three adjacent sensors is greater than the difference threshold, then determining the leakage diffusion direction based on the gas concentration; adjusting the initial concentration threshold according to the leakage diffusion direction to obtain the target concentration threshold corresponding to each gas concentration sensor.

[0091] In this embodiment, the difference threshold is a preset concentration value, for example, 10 ppm.

[0092] In this embodiment, the initial concentration threshold is adjusted according to the leakage diffusion direction, and the target concentration threshold corresponding to each gas concentration sensor is obtained as follows: the gas concentration sensor in the downwind area is determined according to the leakage diffusion direction, and the target concentration threshold corresponding to the gas concentration sensor in the downwind area is lowered.

[0093] In this embodiment, the gas includes at least one of H 2 S, CO and VOC.

[0094] In this embodiment, the initial concentration threshold may be a preset value, for example, the initial concentration threshold may be 10 ppm.

[0095] In a specific example, if the difference in gas concentration collected by three adjacent sensors is greater than 5 ppm / m, the leakage diffusion direction is determined based on the gas concentration collected by the three adjacent sensors, and the downwind sensor threshold is automatically lowered to 8 ppm.

[0096] Optionally, adjusting the initial concentration threshold according to the leakage diffusion direction to obtain a target concentration threshold corresponding to each gas concentration sensor includes:

[0097] If it is determined that the gas concentration sensor is located in the downwind area according to the leakage diffusion direction, the initial concentration threshold is lowered by a set value to obtain a target concentration threshold corresponding to the gas concentration sensor in the downwind area.

[0098] In this embodiment, the set value is a pre-set concentration value, for example, the set value may be 2 ppm.

[0099] In this embodiment, when the pressure sensor shows an abnormality but the vibration data is normal, redundant sensor comparison is started (such as 2 out of 3 voting); if there is still a conflict, high-frequency sampling (increased from 1Hz to 100Hz) is triggered and re-judgment is made after 10 seconds.

[0100] In this embodiment, if a sensor falsely reports three times in a row, its weight is automatically reduced by 50% and marked as "low confidence". After manual confirmation of the false alarm, a threshold rollback (restoration to the previous stable version) is triggered.

[0101] In a specific example, for the temperature runaway warning of the hydrogenation reactor (high temperature runaway risk): the hydrogenation reactor operates under high pressure and high temperature. If the temperature is out of control (temperature runaway), it may cause catalyst deactivation, equipment damage or even explosion. It is necessary to integrate multiple sources of data such as internal temperature, pressure, circulating hydrogen flow, and external infrared thermal imaging to dynamically adjust the warning threshold. Obtain internal thermocouples (continuous type, sampling frequency 1Hz, accuracy ±1℃), pressure transmitters (4-20mA signal, range 0-20MPa) and infrared thermal imagers (external surface temperature, resolution 640×480, frame rate 30fps); internal temperature threshold = design value (such as 300℃) × 90% = 270℃ (safety margin 10%). When the pressure change rate is >5MPa / min and the circulating hydrogen flow rate drops by >20% → it is determined that insufficient hydrogen supply has caused heat storage, and the temperature threshold is immediately lowered to 250℃. Infrared cross-validation: If the internal temperature is >270℃, but the external infrared shows a normal temperature gradient (temperature difference <5℃ / m 2 ) → Trigger the sensor calibration process; if the external infrared shows a local hot spot (temperature difference > 20℃ / m 2) → Ignore internal sensor delays and directly trigger emergency pressure relief. Temperature > 270°C + normal pressure change rate → Primary alarm (notify the central control room); Temperature > 270°C + pressure change rate > 5MPa / min → Advanced alarm (initiate quench hydrogen injection); Infrared hotspot + internal temperature > 250°C → Emergency shutdown (interlock the SIS system); Pressure sensor fault tolerance: If pressure data is lost, automatically switch to a pure temperature-flow model, and tighten the threshold to 260°C. Historical data playback verification: Inject historical temperature fluctuation event data (such as a 10-second curve of temperature rising from 250°C to 400°C) to test whether the rule response delay is less than the time threshold (for example, it can be 2 seconds). For liquefied hydrocarbon storage tank leak monitoring (VOC gas + video fusion): Liquefied hydrocarbons (such as liquefied petroleum gas (LPG)) quickly vaporize after leakage, forming a low-temperature white mist. Traditional gas sensors have delays, and early warning requires a combination of VOC concentration, video smoke recognition, and pressure drop. Phase 1 (gas concentration trigger): VOC concentration > 20% LEL (lower explosion limit) → start video analysis and call the YOLOv8 model to detect white fog. Phase 2 (dynamic threshold adjustment): If the video recognizes white fog and the coverage area is greater than the area threshold (for example, 1m 2) → Lower the VOC alarm threshold from 20%LEL to 10%LEL and activate surrounding fans (to reduce local concentration). If the pressure sensor indicates a pressure drop >0.1MPa / min within the tank → Ignore the concentration threshold and directly trigger the emergency shut-off valve. Video false alarm filtering: White mist recognition must last for three frames (0.1 seconds) and be synchronized with the rising trend of gas concentration; eliminate interference such as steam and dust (through texture analysis: white mist edge blur >60). For long-distance pipeline corrosion leak warning (ultrasonic + corrosion rate model): Corrosion monitoring data: ultrasonic thickness gauge (accuracy ±0.1mm), online pH value, and medium flow rate. Long-term trend prediction: Based on historical corrosion data (e.g., wall thickness from 12mm to 11.5mm / year), fit an exponential decay model: Automatically update the safe wall thickness threshold monthly (e.g., from 10mm to 9.8mm). Short-term mutation detection: If the wall thickness drops >0.05mm per day (exceeding three times the average rate), trigger high-frequency detection mode (increasing sampling from 1 time / hour to 1 time / minute). Corrosion model training: An LSTM network is used to predict wall thickness for the next seven days, using input features including pH, temperature, and flow rate. A manual review is triggered when the model error exceeds 0.2mm. High-risk pipe sections (red areas near the threshold) are marked on the digital twin map to guide inspection routes. Fire early warning for constant and distillation units (infrared thermal imaging + flame recognition): Flames in high-temperature environments can be misidentified (e.g., normal combustion in the furnace), requiring differentiation between normal fire sources and abnormal fires. Infrared image zoning detection: If a hot spot >200°C appears in a non-combustion area (e.g., a pump room), an immediate alarm is issued. Normal combustion flames: High intensity in the near-infrared band and low intensity in the ultraviolet band; oil leak fires: Sudden increase in ultraviolet band intensity: Adjust the alarm priority to the highest. If the flame recognition + combustible gas concentration exceeds 15% LEL, the foam fire extinguishing system is automatically activated, closing the fuel valve. High-temperature cameras: Installed in a 350°C explosion-proof shield, the mirrors are regularly purged with nitrogen. The flame recognition model runs in real time using NVIDIA Jetson AGX Orin. For coking monitoring of catalytic cracking units (vibration + acoustic emission fusion): Catalyst coking causes abnormal vibration of the internal components of the reactor, but traditional vibration sensors are easily interfered by noise. Normal operating baseline: The main frequency band of the vibration spectrum is 80-120Hz, and the amplitude is <2mm / s. Coking feature identification: If the vibration energy migrates to the high frequency band (>200Hz) and the acoustic emission count rate is >100 times / second → it is determined to be in the early stage of coking, and the vibration alarm threshold is lowered from 2mm / s to 1.5mm / s. Within 24 hours after the unit is restarted, the threshold is relaxed to 2.5mm / s to avoid false alarms caused by startup shock. The original vibration signal is decomposed into 5 layers to remove mechanical noise <20Hz. Feature fusion model: Input: vibration spectrum entropy + acoustic emission energy → output coking probability (0-1); if the probability is >0.7 → trigger the decoking procedure recommendation.

[0102] In another specific example, the safety risk early warning method for a petroleum catalytic cracking unit provided by an embodiment of the present invention is executed by a main controller in a safety risk early warning system. The safety early warning system includes a main controller, sensors, and a power supply. The main controller includes a main control module, a data processing module, an environmental adaptation module, an early warning execution module, and a communication module. The main control module is electrically connected to each module. The sensors include a temperature sensor, a pressure sensor, a gas concentration sensor, a humidity sensor, and a vibration sensor. The data processing module includes a data acquisition unit, a data fusion unit, and an anomaly detection unit. The environmental adaptation module includes a parameter adaptive adjustment unit and a model adaptive learning unit. The early warning execution module includes an alarm unit and a control unit. The power supply includes a main power supply and a backup power supply. The communication module includes a wired communication unit and a wireless communication unit. This method can dynamically adjust early warning parameters and update anomaly detection models based on real-time environmental data, improving early warning accuracy, reducing false alarm rates, and ensuring the safety and reliability of petrochemical production.

[0103] like Figure 2 and 3 As shown, the main control module 1 includes: a central processing unit 11, a data storage 12, a signal transceiver 13 and a power management module 14; the central processing unit 11 is connected to the data storage 12 via a data bus, the signal transceiver 13 is connected to the central processing unit 11, and the power management module 14 is connected to the central processing unit 11. Figure 2 and 4 As shown, the sensor 2 includes a temperature sensor 21, a pressure sensor 22, a gas concentration sensor 23, a humidity sensor 24 and a vibration sensor 25. Figure 2 and 5 As shown, the data processing module 3 includes a data acquisition unit 31, a data fusion unit 32 and an anomaly detection unit 33. The data acquisition unit 31 is connected to the data fusion unit 32 via a bus, the data fusion unit 32 is connected to the anomaly detection unit 33 via a bus, and the anomaly detection unit 33 is connected to the main control module 1 via a signal transceiver 13. Figure 2 and 6As shown, the environment adaptation module 4 includes a parameter adaptive adjustment unit 41 and a model adaptive learning unit 42. The parameter adaptive adjustment unit 41 is connected to the data fusion unit 32 of the data processing module 3 through the signal transceiver 13, and the model adaptive learning unit 42 is connected to the abnormality detection unit 33 of the data processing module 3 through the signal transceiver 13; the parameter adaptive adjustment unit 41 includes an environment parameter monitoring subunit 411, a parameter adjustment subunit 412 and a parameter optimization subunit 413. The environment parameter monitoring subunit 411 is connected to the parameter adjustment subunit 412 through a data bus. The parameter adjustment subunit 412 is connected to the parameter optimization subunit 413 via a data bus, and the parameter optimization subunit 413 is connected to the main control module 1 via a signal transceiver 13; the model adaptive learning unit 42 includes a data preprocessing subunit 421, a model training subunit 422 and a model optimization subunit 423, the data preprocessing subunit 421 is connected to the model training subunit 422 via a data bus, the model training subunit 422 is connected to the model optimization subunit 423 via a data bus, and the model optimization subunit 423 is connected to the main control module 1 via a signal transceiver 13. Figure 2 and 7 As shown, the early warning execution module 5 includes an alarm unit 51 and a control unit 52. The alarm unit 51 is connected to the main control module 1 through the signal transceiver 13, and the control unit 52 is connected to the main control module 1 through the signal transceiver 13. The alarm unit 51 includes an audible and visual alarm 511 and an alarm indicator light 512. The audible and visual alarm 511 is connected to the main control module 1 through the signal transceiver 13, and the alarm indicator light 512 is connected to the main control module 1 through the signal transceiver 13. The control unit 52 includes a relay 521 and a control interface 522. The relay 521 is connected to the main control module 1 through the signal transceiver 13, and the control interface 522 is connected to the main control module 1 through the signal transceiver 13. Figure 2 and 8 As shown, the power supply 6 includes a main power supply 61 and a backup power supply 62, which are connected via a power switching device 63; the power supply 6 also includes an intelligent power management module 64, which includes a power monitoring unit 641, a battery charger 642 and an energy recovery unit 643. The power monitoring unit 641 is connected to the main control module 1 via a signal transceiver 13, the battery charger 642 is connected to the main control module 1 via a data bus, and the energy recovery unit 643 is connected to the main control module 1 via a data bus. Figure 2 and 9As shown, the communication module 7 includes a wired communication unit 71 and a wireless communication unit 72. The wired communication unit 71 is connected to the main control module 1 via a data bus, while the wireless communication unit 72 is connected to the main control module 1 via an antenna. The communication module 7 also includes an intelligent communication management module 73, which includes a communication status monitoring unit 731 and a communication protocol conversion unit 732. The communication status monitoring unit 731 is connected to the main control module 1 via a signal transceiver 13, while the communication protocol conversion unit 732 is connected to the main control module 1 via a data bus. The temperature sensor 21, pressure sensor 22, gas concentration sensor 23, humidity sensor 24, and vibration sensor 25 are installed at key locations in the petrochemical equipment. The main power supply 61 is connected to the mains electricity, while the backup power supply 62 is a large-capacity battery. The power switching device 63 automatically switches to the backup power supply 62 when the main power supply 61 loses power. The wired communication unit 71 is used for wired communication with other on-site equipment, while the wireless communication unit 72 is used for wireless communication with a remote monitoring center.

[0104] It should be noted that the parameter adaptive adjustment unit in the environmental adaptation module monitors changes in environmental parameters in real time through the environmental parameter monitoring subunit. The parameter adjustment subunit dynamically adjusts warning parameters based on changes in environmental parameters. The parameter optimization subunit optimizes the parameter adjustment strategy based on historical and real-time data. The data preprocessing subunit in the model adaptive learning unit preprocesses real-time and historical data. The model training subunit trains an anomaly detection model based on the preprocessed data. The model optimization subunit optimizes the anomaly detection model based on the training results. The main control module, based on the data provided by the parameter adaptive adjustment unit and the model adaptive learning unit, sends instructions to the data fusion unit and anomaly detection unit in the data processing module via a signal transceiver to fuse the real-time collected data and detect anomalies. When an anomaly is detected, the main control module sends instructions to the alarm unit and control unit in the warning execution module via a signal transceiver. The alarm unit emits an audible and visual alarm signal, and the control unit controls the relevant equipment to take safety measures. The intelligent power management module in the power supply monitors the status of the main and backup power supplies in real time through the power monitoring unit. The battery charger charges the backup power supply when the main power supply is functioning properly, and the energy recovery unit replenishes the backup power supply with regenerated energy during equipment operation. The intelligent communication management module in the communication module monitors the communication status of the communication module in real time through the communication status monitoring unit, and the communication protocol conversion unit converts data between different communication protocols. The main control module connects to each module through a signal transceiver and data bus to achieve real-time data transmission and the issuance of control instructions.

[0105] The parameter adaptive adjustment unit and the model adaptive learning unit can dynamically adjust the warning parameters and update the anomaly detection model according to the real-time collected environmental data, thereby improving the adaptability and accuracy of the warning device under complex working conditions and reducing the false alarm rate.

[0106] The data processing module fuses the data of multiple sensors through the data fusion unit, which improves the reliability and accuracy of the data and enhances the precision and stability of anomaly detection.

[0107] The intelligent power management module monitors the status of the main power supply and backup power supply in real time through the power monitoring unit. The battery charger charges the backup power supply when the main power supply is normal. The energy recovery unit uses the regenerated energy during equipment operation to supplement the backup power supply, ensuring the normal operation of all modules when the main power supply fails, thereby improving the reliability and stability of the system.

[0108] The technical solution provided by this embodiment solves the problems of existing early warning devices, such as unstable warning effects under complex operating conditions, high false alarm rates, high system complexity, high maintenance costs, and high energy consumption. The technical solution provided by this embodiment of the invention is more intelligent, efficient, and adaptable to changing environments, thereby effectively improving the safety and reliability of petrochemical production.

[0109] The technical solution of this embodiment is to first obtain multimodal data of the petroleum catalytic cracking unit; then determine the reconstruction error, energy balance index, oxygen balance coefficient and change rate correlation coefficient corresponding to the multimodal data; finally, when the reconstruction error, energy balance index, oxygen balance coefficient and change rate correlation coefficient corresponding to the multimodal data meet the coupling fault alarm conditions, a coupling fault alarm is issued, which can trace the source and provide early warning, achieve the purpose of active prevention and control, and further improve the safety and reliability of petrochemical production.

[0110] Example 2

[0111] Figure 10 This is a schematic diagram of the structure of a safety risk warning device for a petroleum catalytic cracking unit provided by an embodiment of the present invention. This embodiment is applicable to situations where a safety risk warning is provided for a petroleum catalytic cracking unit. The device can be implemented in software and / or hardware. The device can be integrated into any device that provides a safety risk warning function for a petroleum catalytic cracking unit, such as Figure 10 As shown, the safety risk warning device for a petroleum catalytic cracking unit specifically includes: a multimodal data acquisition module 1010 , a determination module 1020 and a warning module 1030 .

[0112] The multimodal data acquisition module is used to acquire multimodal data of the petroleum catalytic cracking unit, wherein the multimodal data includes: temperature, pressure, circulation rate and oxygen content;

[0113] a determination module, configured to determine a reconstruction error, an energy balance index, an oxygen balance coefficient, and a change rate correlation coefficient corresponding to the multimodal data;

[0114] The early warning module is configured to issue a coupling fault alarm if the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data meet coupling fault alarm conditions.

[0115] The above-mentioned product can execute the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0116] Example 3

[0117] Figure 11 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0118] like Figure 11 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0119] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0120] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the safety risk early warning method for a petroleum catalytic cracking unit.

[0121] In some embodiments, the safety risk early warning method for a petroleum catalytic cracking unit may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the safety risk early warning method for a petroleum catalytic cracking unit described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the safety risk early warning method for a petroleum catalytic cracking unit in any other appropriate manner (e.g., by means of firmware).

[0122] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0123] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0124] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0125] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0126] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0127] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0128] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0129] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the safety risk early warning method for a petroleum catalytic cracking unit according to any embodiment of the present invention.

[0130] The computer program product may be implemented by writing computer program code for performing the operations of the present invention in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0131] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A safety risk early warning method for a petroleum catalytic cracking unit, characterized in that: include: Acquiring multimodal data of a petroleum catalytic cracking unit, wherein the multimodal data includes: temperature, pressure, circulation rate, and oxygen content; Determining a reconstruction error, an energy balance index, an oxygen balance coefficient, and a rate of change correlation coefficient corresponding to the multimodal data; If the reconstruction error, energy balance index, oxygen balance coefficient and change rate correlation coefficient corresponding to the multimodal data meet the coupling fault alarm condition, a coupling fault alarm is issued.

2. The method according to claim 1, characterized in that Also includes: If the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data do not meet the coupling fault alarm condition, a detection rule corresponding to each modal data is determined, and a detection result of the petroleum catalytic cracking unit is determined based on the detection rule and each modal data, wherein the reconstruction error corresponding to the multimodal data is determined by the difference between the multimodal data and the reconstructed data of the multimodal data, the reconstructed data of the multimodal data is obtained by inputting the multimodal data into a concept drift detection model, the energy balance index is equal to the product of the temperature and the circulation rate, and the ratio to the pressure, the oxygen balance coefficient is equal to the product of the oxygen content and the characteristic parameter, the characteristic parameter is equal to the square root of the ratio of the pressure to the circulation rate, and the change rate correlation coefficient is equal to the correlation coefficient between the derivative of the temperature and the time and the derivative of the circulation rate and the time; The coupling fault alarm conditions include: The reconstruction error corresponding to the multimodal data is greater than a target value, wherein the target value is determined by a mean and a standard deviation of the target reconstruction error; The energy balance index is less than the energy balance index threshold; The oxygen balance coefficient is greater than a first value, or less than a second value, wherein the second value is less than the first value; The change rate correlation coefficient is smaller than a third value or larger than a fourth value, wherein the third value is smaller than the fourth value.

3. The method according to claim 2, characterized in that The multimodal data further includes: an infrared image, the infrared image being used to characterize the surface temperature distribution of a storage tank, wherein the petroleum catalytic cracking unit includes: a storage tank; Determining a detection rule corresponding to the infrared image, and determining a detection result of the storage tank based on the detection rule and the infrared image, including: If the local hot spot temperature of the infrared image is greater than a target temperature threshold, it is determined that the tank is overheated, wherein the target temperature threshold is determined in the following manner: If the temperature in the tank does not rise, the initial temperature threshold is used as the target temperature threshold; If the temperature in the storage tank rises, the difference between the initial temperature threshold and the first temperature value is used as the target temperature threshold.

4. The method according to claim 3, characterized in that The multimodal data further includes: corrosion detection data, wherein the corrosion detection data includes: pipeline wall thickness, real-time corrosion rate and chemical property parameters of the liquid in the pipeline, and the petroleum catalytic cracking unit further includes: pipelines; Determining a detection rule corresponding to the pipeline wall thickness, and determining a pipeline detection result based on the detection rule and the pipeline wall thickness, including: If the pipe wall thickness is less than the target pipe wall thickness threshold, it is determined that the pipe wall thickness is abnormal, wherein the target pipe wall thickness threshold is determined in the following manner: Determine the initial pipe wall thickness threshold based on the wall thickness design value; The target pipeline wall thickness threshold is determined according to the real-time corrosion rate, the chemical property parameters of the liquid in the pipeline, and the initial pipeline wall thickness threshold.

5. The method according to claim 4, characterized in that Determine the target pipe wall thickness threshold based on the real-time corrosion rate, the chemical properties of the liquid in the pipe, and the initial pipe wall thickness threshold, including: Determine remaining pipeline life based on real-time corrosion rates; If the pipeline is determined to be in a stable corrosion state based on the chemical properties of the liquid in the pipeline, the target pipeline wall thickness threshold is determined based on the attenuation frequency, the remaining life of the pipeline, and the initial pipeline wall thickness threshold; If the pipeline is determined to be in an accelerated corrosion state based on the chemical property parameters of the liquid in the pipeline, the target pipeline wall thickness threshold is determined based on the first multiple, the attenuation frequency, the remaining life of the pipeline, and the initial pipeline wall thickness threshold.

6. The method according to claim 5, characterized in that The multimodal data also includes: gas concentrations collected by each gas concentration sensor; Determining a detection rule corresponding to the gas concentration, and determining a detection result of the petroleum catalytic cracking unit based on the detection rule and the gas concentration, including: If the gas concentration collected by at least one gas concentration sensor is greater than a target concentration threshold corresponding to the gas concentration sensor, a gas leak is determined, wherein the target concentration threshold corresponding to the gas concentration sensor is determined in a manner including: Get the initial concentration threshold; If the difference in gas concentrations collected by at least three adjacent sensors is greater than a difference threshold, the leakage diffusion direction is determined based on the gas concentration; The initial concentration threshold is adjusted according to the leakage diffusion direction to obtain a target concentration threshold corresponding to each gas concentration sensor.

7. The method according to claim 6, characterized in that The initial concentration threshold is adjusted according to the leakage diffusion direction to obtain the target concentration threshold corresponding to each gas concentration sensor, including: If it is determined that the gas concentration sensor is located in the downwind area according to the leakage diffusion direction, the initial concentration threshold is lowered by a set value to obtain a target concentration threshold corresponding to the gas concentration sensor in the downwind area.

8. A safety risk early warning device for a petroleum catalytic cracking unit, characterized in that: include: A multimodal data acquisition module, configured to acquire multimodal data of a petroleum catalytic cracking unit, wherein the multimodal data includes temperature, pressure, circulation rate, and oxygen content; a determination module, configured to determine a reconstruction error, an energy balance index, an oxygen balance coefficient, and a change rate correlation coefficient corresponding to the multimodal data; The early warning module is configured to issue a coupling fault alarm if the reconstruction error, energy balance index, oxygen balance coefficient, and change rate correlation coefficient corresponding to the multimodal data meet coupling fault alarm conditions.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the safety risk early warning method for a petroleum catalytic cracking unit according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the safety risk early warning method for a petroleum catalytic cracking unit according to any one of claims 1 to 7 when executed.

11. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements the safety risk early warning method for a petroleum catalytic cracking unit according to any one of claims 1 to 7.

Images