Intelligent fire extinguishing buffer device for automatic firework and cracker production line

Abstract

The invention discloses an intelligent fire extinguishing buffer device for an automatic firework and cracker production line. A system comprises a sensing recognition module, an airflow modeling module, a fire extinguishing evaluation module, a nozzle correction module and a control decision module. According to the method, the boundary of the fire source is locked through multi-source induction fusion, the vertical airflow velocity is calculated based on the layered pressure difference, the flowing environment around the fire source is dynamically mastered, whether the coverage requirement is met or not is judged by combining the liquid level and the flow velocity of the fire extinguishing agent, and if not, an alarm is triggered and the nozzle orientation is guided to correct the angle; a three-dimensional motion vector is constructed by a flame horizontal propagation trend and vertical airflow, and the spraying direction and force are adjusted in real time, so that medium delivery is dynamically matched with a fire source, the fire extinguishing response speed, the space adaptive capacity and the resource utilization efficiency are improved, the problems of blind spraying, deflected spraying, insufficient dosage or waste and the like are reduced, and the safety guarantee capacity of a risk workshop section is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of spraying devices, and particularly to an intelligent fire extinguishing and buffering device for an automated production line of fireworks and firecrackers. Background Art

[0002] During the production process of fireworks and firecrackers, due to the existence of flammable and explosive substances such as gunpowder and pyrotechnics, fires or explosions are extremely likely to be triggered by friction, static electricity or operational errors. Traditional fire extinguishing methods (such as manual fire extinguishers and fixed sprinkler systems) have problems such as slow response speed and limited coverage, and cannot effectively respond to sudden fires.

[0003] In the prior art, the automated fire extinguishing devices used in fireworks and firecrackers factories detect fires using a single sensor (such as a smoke or temperature sensor), but have disadvantages such as a high false alarm rate and a lag in response. In addition, the existing systems lack a buffering mechanism and cannot quickly suppress the spread of shock waves in the initial stage of an explosion, resulting in the expansion of the accident.

[0004] In the prior art, although there are already induction and spraying response mechanisms for the fire source, there are generally problems such as limited recognition range, rigid response path, and rough parameter control. In traditional systems, fire source positioning is mostly triggered based on a single sensing signal, and is easily affected by smoke shielding, environmental interference, etc., resulting in misjudgment or missed judgment, and then leading to inaccurate fire extinguishing direction. During the fire extinguishing process, the orientation of the spray nozzle depends on static preset or rough adjustment, lacking real-time calculation support for the actual propagation dynamics of the flame, and easily causing lag or off-target spraying of the fire extinguishing agent. Traditional airflow environment assessment is mostly static modeling, ignoring the convective disturbances formed around the fire source with the temperature rise, resulting in the deviation of the fire extinguishing agent delivery path and reducing the delivery efficiency. In addition, most systems only have an opening or closing threshold for the fire extinguishing agent and cannot dynamically adjust the release intensity of the fire extinguishing agent, resulting in over-spraying in some areas and under-spraying in other areas, increasing the consumption of the fire extinguishing agent and reducing the processing efficiency. For example, in continuous production line operations, if a fire source forms at the edge of the conveyor belt, the fixed direction of the traditional nozzle cannot effectively deliver it to the specific fire area, increasing the risk of secondary ignition. The above limitations will seriously affect the production safety guarantee ability in high-density and highly flammable technological processes. Summary of the Invention

[0005] The purpose of the present invention is to solve the disadvantages existing in the prior art, and to propose an intelligent fire extinguishing and buffering device for an automated production line of fireworks and firecrackers.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0007] An intelligent fire extinguishing and buffering device for an automated production line of fireworks and firecrackers includes:

[0008] The sensing and recognition module acquires the smoke concentration, infrared radiation intensity, and ultraviolet radiation intensity obtained by sensors in the flammable section of the fireworks and firecrackers automated production line, performs multi-modal fusion on the three groups of data through the SSD target detection algorithm, generates a fire source bounding box, and transmits it to the airflow modeling module;

[0009] The airflow modeling module acquires the z-axis height information in the fire source bounding box, obtains the stratified pressure data of the z-axis height through a piezoresistive pressure sensor, calculates the difference between the pressure data of adjacent height layers to construct a vertical pressure gradient, and generates a vertical airflow velocity through pneumatic function conversion, and transmits it to the fire extinguishing evaluation module;

[0010] The fire extinguishing evaluation module acquires the liquid level data of ABC dry powder fire extinguishing agent through a capacitive water level sensor, combines and calculates it with the vertical airflow velocity to generate the mass flow rate of the fire extinguishing agent. When the mass flow rate of the fire extinguishing agent is lower than the fire source bounding box, an alarm signal is triggered and transmitted to the nozzle correction module;

[0011] The nozzle correction module, based on the alarm signal, calculates the horizontal propagation rate of the flame through the fire source bounding box, constructs a three-dimensional motion vector in combination with the vertical airflow velocity data, generates a nozzle azimuth correction angle based on the projection angle of the vector in the x-z plane, and performs a fire extinguishing operation;

[0012] The control decision module, based on the nozzle azimuth correction angle, acquires the infrared radiation intensity in the new cycle through a dual-band flame sensor and calculates the attenuation rate. The attenuation rate and the horizontal propagation rate of the flame are processed through a proportional-integral controller to generate an optimized injection angle and output an angle adjustment instruction to the actuator.

[0013] As a further solution of the present invention, the fire source bounding box is a spatial positioning coordinate, an outer circumscribed rectangle size, and a bounding box confidence level. The vertical airflow velocity includes a velocity direction component, a stratified velocity interval, and a velocity amplitude change rate. The mass flow rate of the fire extinguishing agent includes a time release amount, an airflow matching ratio, and a boundary matching ratio. The nozzle azimuth correction angle is a plane projection angle, a rotation direction determination value, and an azimuth angle change amount. The angle adjustment instruction refers to a correction angle value, a control signal number, and an actuator response time.

[0014] As a further solution of the present invention, the sensing and recognition module includes:

[0015] The signal extraction sub-module acquires the smoke concentration collected by a laser scattering type smoke sensor in the flammable section of the fireworks and firecrackers automated production line, and collects the infrared radiation intensity and ultraviolet radiation intensity through a dual-band flame sensor, arranges the three groups of data, and processes the repeated values and missing values to generate time series perception data;

[0016] The data fusion sub-module, based on the time series perception data, calculates the light intensity composite fluctuation index of the light radiation intensity through the SSD target detection algorithm, and simultaneously calculates the smoke concentration dynamic index. It annotates the data sequence based on the start and end positions of the change, and generates the fire source change interval.

[0017] The boundary generation sub-module calls the fire source change interval to frame the position of the suspicious area, filters the detection frames located in the concentrated area of the fire source change, and extracts the boundary positions according to the coordinate range to generate the fire source boundary box.

[0018] As a further solution of the present invention, the light intensity composite fluctuation index of the light radiation intensity is calculated through the SSD target detection algorithm.

[0019]

[0020] Where: x represents the sensor spatial distribution rate, and T0 represents the system sampling period reference value. is the spatial gradient modulus value of the light radiation intensity. represents the spatial change rate of the light radiation intensity in the x direction. represents the spatial change rate of the light radiation intensity in the y direction. is the upper limit of the light radiation intensity range, α is the spatio-temporal coupling coefficient, and L t―k represents the k-th order light intensity value, and β k is the data attenuation factor, k ∈ (1, m), and Δτ k is the data time interval.

[0021] As a further solution of the present invention, the airflow modeling module includes:

[0022] The height extraction sub-module obtains the z-axis height coordinate information in the fire source boundary box, identifies the vertical direction of the boundary box and performs hierarchical operations at fixed intervals, marks the vertical positions of each layer, and generates the height hierarchical coordinates.

[0023] The pressure difference sub-module, according to the height hierarchical coordinates, collects the corresponding pressure monitoring data of each layer through a piezoresistive pressure sensor, calculates the difference in pressure between adjacent height layers and statistically analyzes the pressure change trend of multiple layers to generate the vertical pressure gradient.

[0024] The airflow calculation sub-module, based on the vertical pressure gradient, calculates the airflow velocity in the vertical direction of each layer through an aerodynamic function, combines and processes the velocity values of multiple layers and extracts the velocity change interval in the central area to generate the vertical airflow velocity.

[0025] As a further solution of the present invention, the fire extinguishing evaluation module includes:

[0026] Liquid level acquisition sub-module, which acquires the monitoring data of the capacitive water level sensor, records the real-time liquid level height of ABC dry powder fire extinguishing agent and conducts time-continuous screening, identifies the fire extinguishing agent storage according to the stable liquid level section, and generates the fire extinguishing agent liquid level;

[0027] Flow calculation sub-module, based on the fire extinguishing agent liquid level, combines the vertical air flow velocity to calculate the overall circulation volume of the fire extinguishing agent, and converts it into mass flow rate with reference to the cross-sectional size of the discharge channel, and generates the fire extinguishing agent mass flow rate;

[0028] Status judgment sub-module, calls the fire extinguishing agent mass flow rate and compares it with the fire source bounding box, calculates the corresponding relationship between the flow rate and the bounding box, and triggers an alarm control signal when the fire extinguishing agent mass flow rate value is not sufficient to cover the space area, and generates an alarm signal.

[0029] As a further solution of the present invention, the nozzle correction module includes:

[0030] Flame recognition sub-module, based on the alarm signal, calculates the change amount of the horizontal direction of the fire source bounding box, compares the change rate of the horizontal projection area of the bounding box at adjacent time sequences and extracts the maximum slope change, obtains the lateral change trend of the fire in this area, and generates the flame propagation rate;

[0031] Vector construction sub-module, calls the flame propagation rate and the vertical air flow velocity, calculates the synthesis of the two-direction velocity components and constructs a velocity vector to calibrate the direction, and generates a flame motion vector;

[0032] Angle correction sub-module, according to the component ratio of the flame motion vector in the x-z plane, performs arctangent function conversion, obtains the included angle direction of the vector in the plane, combines with the original emission angle direction of the fire extinguishing nozzle to perform difference correction processing and synchronously refresh the nozzle control instruction parameters, and generates a nozzle azimuth correction angle.

[0033] As a further solution of the present invention, the control decision module includes:

[0034] Signal acquisition sub-module, based on the nozzle azimuth correction angle, obtains the infrared radiation intensity collected by the dual-band flame sensor in the current cycle and synchronizes the current time mark, calculates the difference from the infrared radiation intensity at the same position in the previous cycle, and generates an infrared radiation attenuation rate;

[0035] Attenuation calculation sub-module, for the infrared radiation attenuation rate, calculates the response deviation between the two items through a proportional-integral controller, and performs proportional processing and error accumulation processing on the response deviation, combines the two results and sets upper and lower limit constraints, and generates an angle adjustment correction value;

[0036] The angle adjustment sub-module adjusts the nozzle rotation angle command parameter in the actuator according to the offset difference between the angle adjustment correction value and the current nozzle basic emission angle, generates an optimized injection angle, and outputs an angle adjustment instruction to the actuator.

[0037] As a further solution of the present invention, the response deviation e(t) between the two items is calculated by a proportional-integral controller, and the formula is used:

[0038]

[0039] where C IR represents the current measured value of the infrared radiation attenuation rate, V FS represents the current measured value of the flame propagation rate, represents the upper limit value of the infrared radiation attenuation rate range, K p is the proportional coefficient adjustment factor, K i is the integral coefficient adjustment factor, T is the dynamic time integral threshold, τ is the integral process time variable, and d is the differential element of the integral process time variable τ.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] In the present invention, through the fusion analysis of multi-source information such as smoke concentration, infrared and ultraviolet radiation intensity, etc., the precise locking of the fire source boundary is realized, and the accuracy and response speed of the fire ignition point recognition are significantly improved. After converting the induction data into a three-dimensional coordinate frame, combined with the multi-layer height pressure difference collected by the piezoresistive pressure sensor, the vertical pressure gradient is calculated and the air flow velocity is further deduced, which can dynamically sense the vertical flow environment of the fire source area, and complete the precise matching before the fire extinguishing medium is delivered, ensuring that the medium release path highly coincides with the fire source state.

[0042] Through the dynamic coupling judgment between the fire extinguishing agent storage and the flow rate, an alarm is issued in time when the fire extinguishing agent flow rate is not enough to cover the fire source area, ensuring that the fire extinguishing response is not interrupted and avoiding the occurrence of spraying blind areas. Combining the horizontal diffusion trend of the fire source and the vertical air flow to generate a three-dimensional vector, and further adjusting the nozzle delivery angle, so that the fire extinguishing agent spraying trajectory forms an active response mode covering the flame trend in space. This coherent processing logic realizes the closed-loop collaboration from fire source identification, environment perception to action response, strengthens the accuracy, timeliness and stability of the medium release in high-risk sections. Especially in high-risk scenarios such as fireworks production, it improves the rapidity and full coverage ability of the initial fire situation handling, and greatly alleviates problems such as fire extinguishing direction deviation, incomplete medium coverage, and high fire extinguishing agent loss in traditional systems. Description of the Drawings

[0043] Figure 1 is the device flow chart of the present invention;

[0044] Figure 2 This is the block diagram of the device of the present invention. Specific embodiments

[0045] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0046] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0047] Please refer to Figure 1 , an intelligent fire extinguishing and buffering device for a fireworks and firecrackers automated production line includes:

[0048] A sensing and recognition module, which acquires the smoke concentration, infrared radiation intensity and ultraviolet radiation intensity obtained by sensors in the flammable section of the fireworks and firecrackers automated production line, performs multi-modal fusion on the three groups of data through the SSD target detection algorithm, generates a fire source bounding box and transmits it to the airflow modeling module;

[0049] An airflow modeling module, which acquires the z-axis height information in the fire source bounding box, obtains the stratified pressure data of the z-axis height through a piezoresistive pressure sensor, calculates the difference between the pressure data of adjacent height layers to construct a vertical pressure gradient and converts it through an aerodynamic function to generate a vertical airflow velocity and transmits it to the fire extinguishing evaluation module;

[0050] A fire extinguishing evaluation module, which acquires the liquid level data of ABC dry powder fire extinguishing agent through a capacitive water level sensor, combines and calculates it with the vertical airflow velocity to generate the mass flow rate of the fire extinguishing agent, and triggers an alarm signal and transmits it to the nozzle correction module when the mass flow rate of the fire extinguishing agent is lower than the fire source bounding box;

[0051] A nozzle correction module, based on the alarm signal, calculates the horizontal propagation rate of the flame through the fire source bounding box, constructs a three-dimensional motion vector by combining the vertical airflow velocity data, generates a nozzle azimuth correction angle based on the projection angle of the vector in the x-z plane and performs a fire extinguishing operation;

[0052] The control decision-making module, based on the nozzle azimuth correction angle, obtains the infrared radiation intensity in the new cycle through the dual-band flame sensor and calculates the attenuation rate. It processes the attenuation rate and the flame horizontal propagation rate through a proportional-integral controller to generate an optimized injection angle and outputs an angle adjustment instruction to the actuator.

[0053] The fire source bounding box includes spatial positioning coordinates, the size of the circumscribed rectangle, and the bounding box confidence level. The vertical air velocity includes the velocity direction component, the stratified velocity interval, and the velocity amplitude change rate. The fire extinguishing agent mass flow rate includes the time release amount, the air flow mixing ratio, and the boundary matching ratio. The nozzle azimuth correction angle is the plane projection angle, the rotation direction determination value, and the azimuth angle change amount. The angle adjustment instruction refers to the correction angle value, the control signal number, and the actuator response time.

[0054] Please refer to Figure 2 , the sensing and identification module includes:

[0055] The signal extraction sub-module obtains the smoke concentration collected by the laser scattering type smoke sensor in the flammable section of the fireworks and firecrackers automated production line, and collects the infrared radiation intensity and ultraviolet radiation intensity through the dual-band flame sensor. It arranges the three groups of data and processes the repeated values and missing values to generate time series perception data;

[0056] In the flammable section of the fireworks and firecrackers production line, 3 groups of laser scattering type smoke sensors are deployed to collect data with a sampling period of 200 ms. At the same time, a dual-band flame sensor array is configured to collect the infrared radiation intensity L ir and L uv ultraviolet radiation intensity, L ir represents the 940 nm infrared radiation intensity, L uv represents the 260 nm ultraviolet radiation intensity, and the unit is μW / cm 2 ), when the charging machine completes n = 10 batches of operations, the original data set is obtained. Among them, the 325th group of data C 325 is missing and is filled by linear interpolation fill, C <src= 324 = 367 μg / m 3 ,C 326 = 401 μg / m 3 , the 728th group of data is repeatedly recorded as 1250 μW / cm 2 , retain the first valid value

[0057] Table 1 Example table of sensor data cleaning

[0058] Timestamp (ms) PM2.5 Infrared radiation Ultraviolet radiation 1632584200 352 1845 680 1632584300 367 1920 725 1632584400 384 2015 810 1632584500 401 1980 792

[0059] T s : Sampling period of the smoke sensor, Tf = 100 ms: Flame sensor sampling period, n: Job batch number.

[0060] The data fusion sub-module, based on time series perception data, calculates the light intensity composite fluctuation index of the light radiation intensity through the SSD target detection algorithm, and simultaneously calculates the smoke concentration dynamic index, annotates the data sequence based on the change start and end positions, and generates the fire source change interval;

[0061] When the infrared radiation intensity change rate > 15 μW / (cm 2 ·ms) is detected within m = 5 consecutive sampling periods, trigger the calculation of the light intensity composite fluctuation index:

[0062] 1. Spatial gradient modulus: Represents the radiation intensity change rate in the X direction, Represents the change rate in the y direction;

[0063] 2. Fluctuation term: T0 = 100 ms: T0 represents the reference sampling period, set according to the sampling rate of the flame sensor AD conversion chip, β k = 1 + 0.15(k―1): Decay factor, β k The decay factor is designed according to the data timeliness, Current upper limit of the light intensity range, Δτ k = 100 ms: Δτ k Represents the data interval, set synchronously with the sampling period;

[0064] 3. Composite fluctuation index: x = 0.85: Sensor spatial distribution rate, α = 0.62: α spatio-temporal coupling coefficient, determined by fitting 50 groups of experimental data. When Ψ L < 0.005, it is determined as the safe state, and the time period [t―200 ms, t + 300 ms is marked as the type II change interval, and the maximum radiation fluctuation within the interval is

[0065] The boundary generation sub-module calls the fire source change interval to frame the position of the suspicious area, filters and locates the detection frames in the concentrated area of the fire source change, and extracts the boundary positions according to the coordinate range to generate the fire source boundary box;

[0066] In the spatial coordinate system, filter the detection frame X: Horizontal coordinate, unit mm; Y: Vertical coordinate, unit mm. Set: Boundary condition: X ∈ [1200, 1800] mm, Y ∈ [800, 1200] mm, overlap rate threshold η min = 60%.

[0067] 3 Detection box: Box 1: (1250, 850) to (1700, 1150), area A1 = 550×300 = 165000 mm 2 , Box 2: (1300, 900) to (1750, 1100), area A2 = 450×200 = 90000 mm 2 , Box 3: (1400, 950) to (1800, 1200), area A3 = 400×250 = 100000 mm 2 , where, x min represents the minimum X coordinate, y max represents the maximum y coordinate, y min represents the minimum y coordinate, x max represents the maximum X coordinate;

[0068] x min = min(1250, 1300, 1400) = 1250 mm, y min = min(850, 900, 950) = 850 mm,

[0069] x max = max(1700, 1750, 1800) = 1800 mm, y max = max(1150, 1100, 1200) = 1200 mm,

[0070] The area of the bounding box A = (1800 - 1250)×(1200 - 850) = 192500 mm 2 , compared with the average area A of a single detection box avg , A avg = 118333 mm 2 , an increase of 62.7%.

[0071] Please refer to Figure 2 , the airflow modeling module includes:

[0072] A height extraction sub-module that obtains the z-axis height coordinate information in the fire source bounding box, identifies the vertical direction of the bounding box and performs layering operations at fixed intervals, marks the vertical positions of each layer, and generates height layering coordinates;

[0073] Receives the vertical coordinate z of the fire source bounding box min = 850 mm to z max = 1200 mm, the layering interval Δh = 50 mm, starting from 850 mm for marking, detecting the height H of the bounding box, when H = 350 mm, the number of layers Coordinates 850, 900, 950, 1000, 1050, 1100, 1150, 1200 mm. When currently deployed in the gunpowder filling workshop, laser ranging sensors are installed at each floor height, and the measured values of multiple floor heights are collected at a cycle of 100 ms. When the measured value of the third floor exceeds the range of 950 ± 2 mm for three consecutive times, the system automatically calibrates the sensor offset δ = 1.2 mm, and updates the floor coordinate to 951.2 mm, as shown in Table 2:

[0074] Table 2 Vertical Stratification Coordinate Calibration Record Table

[0075] Layer number Theoretical height Measured mean value Height after calibration 1 850 849.8 850.0 2 900 901.1 900.0 3 950 953.2 951.2 4 1000 999.5 1000.0

[0076] As shown in Table 2, the measured height of the third floor is 953.2 mm, with a deviation of 3.2 mm from the theoretical value. After calibration, it is adjusted to 951.2 mm to ensure that the stratification accuracy is controlled within the range of ±1.5 mm.

[0077] The pressure difference sub-module collects the corresponding pressure monitoring data of each layer through a piezoresistive pressure sensor according to the height stratification coordinates, calculates the pressure difference between adjacent height layers, and statistically analyzes the pressure change trend of multiple layers to generate a vertical pressure gradient;

[0078] Piezoresistive sensors are arranged at 7 height layers, and the sampling period T p , T p = 50 ms. When it is detected that the pressure difference ΔP 54 between the pressure value P5 = 102.3 kPa of the fifth layer and P4 = 98.7 kPa of the fourth layer th exceeds the threshold P = 2.5 kPa, calculate the vertical pressure gradient 54 Take the pressure difference data of adjacent three layers: ΔP 65 = 3.6 kPa, ΔP 76 = 2.8 kPa, ΔP When the gradient value for five consecutive sampling periods, it is marked as an abnormal pressure change area.

[0079] The air flow calculation sub-module calculates the air flow velocity in the vertical direction of each layer through an aerodynamic function based on the vertical pressure gradient, combines and processes the multiple layer velocity values, and extracts the velocity change interval in the central area to generate the vertical air flow velocity;

[0080] According to the pressure gradient Adopt the aerodynamic equation to calculate the air flow velocity. Take the air density ρ = 1.225 kg / m 3 , the pressure difference between layers ΔP = 0.055 × 50 = 2.75 kPa, and calculate the air flow velocity = 0.8 × 67.2 = 53.8 m / s. Three velocity values, 53.8 m / s, 49.2 m / s, and 51.6 m / s, were continuously measured in the central region where z = 1000 mm to 1100 mm. The average value was taken as 51.5 m / s, and the velocity change interval was defined as [49.2, 53.8] m / s. When the upper limit of the interval exceeds the safety threshold v max = 50 m / s, a wind speed alarm is triggered.

[0081] Please refer to Figure 2 , the fire extinguishing evaluation module includes:

[0082] A liquid level acquisition sub-module that acquires the monitoring data of the capacitive water level sensor, records the real-time liquid level height of the ABC dry powder fire extinguishing agent and performs time continuity screening, identifies the fire extinguishing agent storage based on the stable liquid level section, and generates the fire extinguishing agent liquid level;

[0083] By deploying a capacitive liquid level sensor (range 0 - 3 m, accuracy ±0.5%) in the dry powder fire extinguishing agent storage tank, liquid level data is collected at a cycle of 200 ms. When it is detected that the liquid level fluctuates by more than ±2 cm within 10 consecutive cycles, data screening is started, and the data in the time period from t = 1632585000 ms to 1632585600 ms is taken;

[0084] {1.52, 1.53, 1.51, 1.67, 1.50, 1.49, 1.48, 1.47, 1.46, 1.45} m. The above is the original sequence. After removing the 4th outlier 1.67 m, the average value is calculated as 1.496 m. Combining with the storage tank diameter D = 1.2 m, the volume of the fire extinguishing agent V = π × (0.6) 2 × 1.496 = 1.696 m 3 , with a density ρ = 1100 kg / m 3 , the storage M = 1.696 × 1100 = 1865.6 kg, as shown in Table 3:

[0085] Table 3 Liquid level data screening record form

[0086] Timestamp (ms) Original liquid level (m) Effective liquid level (m) 1632585000 1.52 1.52 1632585200 1.67 - 1632585400 1.50 1.50

[0087] As shown in Table 3, the outlier 1.67 m is removed, and the valid data participates in the calculation of the storage.

[0088] A flow rate calculation sub-module that calculates the overall flow rate of the fire extinguishing agent based on the fire extinguishing agent liquid level, combines the vertical air flow velocity, and converts it into a mass flow rate with reference to the cross-sectional size of the discharge channel, generating the fire extinguishing agent mass flow rate;

[0089] Receiving the liquid level average value data 1.496 m, combining with the vertical air flow velocity v avg = 50.4 m / s, the pipe cross-sectional area A = 0.25 × π × 0.32 = 0.0707 m 2 , calculate the volumetric flow rate Q v = A × v = 0.0707 × 50.4 = 3.564 m 3 / s, mass flow rate Q m = 3.564 × 1100 = 3920.4 kg / s. When it is detected that the liquid level in the storage tank drops by the pipeline height Δh = 0.15 m within the time Δt = 10 s, verify the flow consistency: the theoretical consumption ΔV = π × 0.6 2 × 0.15 = 0.1696 m 3 , the current output Q v × Δt = 3.564 × 10 = 35.64 m 3 , error rate Trigger the sensor calibration instruction.

[0090] The status judgment sub-module calls the mass flow rate of the fire extinguishing agent and compares it with the fire source bounding box, calculates the corresponding relationship between the flow rate and the bounding box, and triggers an alarm control signal and generates an alarm signal when the mass flow rate value of the fire extinguishing agent is not sufficient to cover the space area;

[0091] Obtain the mass flow rate Q of the fire extinguishing agent m = 3920.4 kg / s, the volume V of the fire source bounding box box = (1.8 ― 1.25) × (1.2 ― 0.85) × (1.2 ― 0.85) = 0.55 × 0.35 × 0.35 = 0.0674 m 3 , according to the NFPA17 standard fire extinguishing concentration C min = 0.65 kg / m 3 , calculate the mass flow rate Q of the fire extinguishing agent req = C min × V box × 60 = 0.65 × 0.0674 × 60 = 2.628 kg / s, compare the current flow rate with the required flow rate: when Q m = 3920.4 kg / s > Q req = 2.628 kg / s, it is determined that the coverage ability is sufficient, Q req is the required flow rate. When it is detected that the mass flow rate of the fire extinguishing agent is lower than 1.2 × Q req = 3.154 kg / s for three consecutive cycles, trigger a level 3 alarm signal and start the standby pressure supply system.

[0092] Please refer to Figure 2 , the nozzle correction module includes:

[0093] The flame recognition sub-module calculates the change in the horizontal direction of the fire source bounding box based on the alarm signal, compares the change rate of the horizontal projection area of the bounding box under adjacent time series, extracts the maximum slope change, obtains the lateral change trend of the fire in this area, and generates the flame propagation rate.

[0094] The system extracts the horizontal projection coordinates from the fire source bounding box data. At the moment of t1 = 1632587000 ms, the coordinates of the lower left corner of the bounding box (1.25 m, 0.85 m) and the upper right corner (1.8 m, 1.2 m) are recorded, and the horizontal projection area A1 = (1.8 - 1.25) × (1.2 - 0.85) = 0.1925 m 2 , at the moment of t2 = 1632587500 ms, it is detected that the bounding box expands to (1.20 m, 0.80 m) to (1.85 m, 1.25 m), and the new area A2 = (1.85 - 1.20) × (1.25 - 0.80) = 0.2925 m 2 , and calculates the area change rate k within the 500 ms time window A = (0.2925 - 0.1925) / 0.5 = 0.2 m 2 / s. The data for 5 consecutive monitoring cycles are shown in Table 4:

[0095] Table 4 Flame Propagation Rate Monitoring Table

[0096] Time window (ms) <![CDATA[Bounding box area (m 2 )]]> <![CDATA[Rate of change (m 2 / s)]]> 1632587000-7500 0.1925 0.20 1632587500-8000 0.2420 0.25 1632588000-8500 0.3100 0.34

[0097] As shown in Table 4, the maximum change rate of 0.34 m is detected in the time window 1632588000 - 8500 2 / s, exceeding the preset threshold of 0.30 m 2 / s (set according to the gunpowder combustion experiment data). The system marks the time period as the stage of accelerated fire spread and extracts the rate value as the output of the flame propagation rate.

[0098] The vector construction sub-module calls the flame propagation rate and the vertical air flow velocity, calculates the synthesis of the velocity components in two directions, constructs a velocity vector to calibrate the direction, and generates a flame motion vector.

[0099] Receives the flame propagation rate v h = 0.34 m / s and the vertical air flow velocity v v = 50.4 m / s. First, the horizontal velocity component is normalized, and v h is mapped to the range of 0 - 1 (the maximum design value is 1.0 m / s), and the normalized value v h ' = 0.34 / 1.0 = 0.34. The vertical component v v is normalized according to the range of 0 - 60 m / s to v v' = 50.4 / 60 = 0.84, calculate the magnitude of the resultant vector (normalized unit), de-normalize to obtain the current velocity v real = 0.91×60 = 54.6 m / s. The direction angle calculation uses piecewise processing. When v v ' > 0.8, it is directly determined as the vertical dominant direction (angle > 85°). Otherwise, calculate θ = arctan(v v ' / v h '), since v v ' = 0.84 > 0.8, directly output the direction angle θ = 87°, and the generated vector direction is close to vertically upward.

[0100] Angle correction sub-module. According to the component ratio of the flame motion vector in the x-z plane, perform arctangent function conversion to obtain the included angle direction of the vector in the plane. Combine it with the original emission angle direction of the fire extinguishing nozzle for difference correction processing and synchronously refresh the nozzle control instruction parameters to generate the nozzle azimuth correction angle;

[0101] Obtain the current nozzle emission angle θ0 = 80°, which is fed back by the servo motor encoder, and calculate the target correction angle θ target = 87°, the difference Δθ = 87° - 80° = 7°. According to the step angle of the stepper motor of 0.72° (200 steps / revolution), calculate the required number of steps n = 7 / 0.72 ≈ 9.7, round up to 10 steps, with a delay of 10 ms for each step, and the total adjustment time is 100 ms. Generate a PWM control signal with a duty cycle set as D = 10 / 200 = 5%. After driving the motor to rotate 10 steps, calibrate the nozzle angle to 80° + 10×0.72 = 87.2°, and update the control parameter to θ new = 87.2°, and synchronously record the error value of +0.2° to the calibration database.

[0102] Please refer to Figure 2 , the control decision module includes:

[0103] Signal acquisition sub-module. Based on the nozzle azimuth correction angle, obtain the infrared radiation intensity collected by the dual-band flame sensor in the current cycle and synchronize the current time mark, calculate the difference in the infrared radiation intensity at the same position in the previous cycle, and generate the infrared radiation attenuation rate;

[0104] At time t, at t = 1632589000 ms, obtain the nozzle azimuth correction angle θ = 87.2°, call the infrared radiation intensity data of the dual-band flame sensor at the position of x = 1.5 m and y = 1.0 m, the measured value in the current cycle The value at the same position in the previous cycle t - 100 ms Calculate the attenuation rate When three consecutive cycles CIR > 0.3 μW / (cm 2 ·ms), it is marked as an effective attenuation signal, as shown in Table 5:

[0105] Table 5 Infrared Radiation Attenuation Rate Record Table

[0106] Timestamp Infrared radiation intensity Attenuation rate 1632588900 2015 - 1632589000 1980 0.35 1632589100 1950 0.30

[0107] As shown in Table 5, the attenuation rate of timestamp 1632589000 is 0.35 μW / (cm 2 ·ms), which exceeds the threshold of 0.3, triggering subsequent calculations.

[0108] Attenuation calculation sub-module, infrared radiation attenuation rate, calculates the response deviation between two items through a proportional-integral controller, performs proportional processing and error accumulation processing on the response deviation, combines the two results and sets upper and lower limit constraints to generate an angle adjustment correction value;

[0109] Receive C IR = 0.35 μW / (cm 2 ·ms) and the flame propagation rate V FS = 0.34 m / s, set the proportional coefficient K p = 0.8, optimized through 10 step response tests, the integral coefficient K i = 0.2, calculated according to the system delay time T d = 50 ms, range upper limit Calculate the proportional term For the integral term, select T = 3 cycle data {0.35, 0.30, 0.28}. Among them, the proportional term coefficient K p = 0.8 is obtained by adjusting through 10 step response experiments (the overshoot of experimental data is all < 5%), and the integral term coefficient K i = 0.2 is calculated according to the system delay time T d = 50 ms (K i = 1 / (5T d )), range upper limit

[0110] Set according to the technical manual of sensor MP-L3-1000, the dynamic time integral threshold T = 3 is determined according to the flame spread characteristic experiment, covering the typical fire trend change cycle. The integral process time variable τ represents the discrete position of the data in the time series, and the differential element dτ is discretized into a fixed sampling interval of 100 ms. Calculate the integral part: Total deviation After restricting the output range to [0, 0.05], generate the angle adjustment correction value Δθ adj = 0.0101 × 50 = 0.505°

[0111] An angle adjustment sub-module, based on the offset difference between the angle adjustment correction value and the current nozzle basic exit angle, re-sets the nozzle rotation angle command parameter in the actuator, generates an optimized injection angle, and outputs an angle adjustment instruction to the actuator;

[0112] Obtain the current nozzle angle θ current = 87.2°, calculate the target angle θ target = 87.2 + 0.505 = 87.705°, according to the stepper motor resolution of 0.1°, round it to 87.7°, and generate the duty cycle of the PWM control signal Drive the motor to rotate 5 steps (each step is 0.1°), which takes 5 × 10 ms = 50 ms, update the nozzle angle to 87.7°, synchronously write it to the control register address 0x3A5. Based on the analysis of the flammable section of the fireworks and firecrackers automated production line, based on the multi-module collaborative analysis results, the flame propagation rate v h = 0.34 m / s, the vertical air flow velocity v v = 50.4 m / s, the nozzle azimuth correction angle θ new = 87.7°, the mass flow rate of the fire extinguishing agent Q m = 3920.4 kg / s, generate a control instruction: adjust the nozzle to 87.7°, the stepper motor completes it in 50 ms, send the instruction to the fire extinguishing execution structure, and perform the fire extinguishing operation.

[0113] The above is only a preferred embodiment of the present invention, and it does not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical content of the technical solution of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent fire extinguishing and buffering device for an automated production line of fireworks and firecrackers, characterized in that, The device includes: A sensing and recognition module that acquires the smoke concentration, infrared radiation intensity, and ultraviolet radiation intensity obtained by sensors in the flammable section of the automated fireworks production line, performs multi-modal fusion on the three groups of data through the SSD target detection algorithm, generates a fire source bounding box, and transmits it to the airflow modeling module; An airflow modeling module that acquires the z-axis height information in the fire source bounding box, obtains the stratified pressure data of the z-axis height through a piezoresistive pressure sensor, calculates the difference between the pressure data of adjacent height layers to construct a vertical pressure gradient, and generates a vertical airflow velocity through pneumatic function conversion, and transmits it to the fire extinguishing evaluation module; A fire extinguishing evaluation module that acquires the liquid level data of ABC dry powder fire extinguishing agent through a capacitive water level sensor, combines and calculates it with the vertical airflow velocity to generate the mass flow rate of the fire extinguishing agent, and triggers an alarm signal and transmits it to the nozzle correction module when the mass flow rate of the fire extinguishing agent is lower than the fire source bounding box; A nozzle correction module that, based on the alarm signal, calculates the horizontal propagation rate of the flame through the fire source bounding box, constructs a three-dimensional motion vector in combination with the vertical airflow velocity data, generates a nozzle azimuth correction angle based on the projection angle of the vector in the x-z plane, and performs a fire extinguishing operation.

2. The intelligent fire extinguishing and buffering device for the automated production line of fireworks and firecrackers according to claim 1, wherein The fire source bounding box is the spatial positioning coordinates, the size of the circumscribed rectangle, and the confidence level of the bounding box. The vertical airflow velocity includes the velocity direction component, the stratified velocity interval, and the rate of change of the velocity amplitude. The mass flow rate of the fire extinguishing agent includes the time release amount, the airflow matching ratio, and the boundary matching ratio. The nozzle azimuth correction angle is the plane projection angle, the rotation direction determination value, and the azimuth angle change amount.

3. The intelligent fire extinguishing and buffering device for the automated production line of fireworks and firecrackers according to claim 1, characterized in that, The sensing and recognition module includes: A signal extraction sub-module that acquires the smoke concentration collected by a laser scattering type smoke sensor in the flammable section of the automated fireworks production line, and collects the infrared radiation intensity and ultraviolet radiation intensity through a dual-band flame sensor, arranges the three groups of data, processes the repeated values and missing values, and generates time series perception data; A data fusion sub-module that, based on the time series perception data, calculates the light intensity composite fluctuation index of the light radiation intensity through the SSD target detection algorithm, calculates the smoke concentration dynamic index at the same time, combines the start and end positions of the change to label the data sequence, and generates the fire source change interval; A boundary generation sub-module that calls the fire source change interval to frame the position of the suspicious area, screens the detection frames located in the concentrated area of the fire source change, and extracts the boundary positions according to the coordinate range to generate a fire source bounding box.

4. The intelligent fire extinguishing and buffering device for the automated production line of fireworks and firecrackers according to claim 3, characterized in that, The light intensity composite fluctuation index Ψ for calculating the light radiation intensity through the SSD target detection algorithm L where: x represents the sensor spatial distribution rate, and T0 represents the system sampling period reference value, is the modulus of the spatial gradient of the light radiation intensity, represents the spatial change rate of the light radiation intensity in the x direction, represents the spatial change rate of the light radiation intensity in the y direction, is the upper limit of the light radiation intensity range, α is the spatio-temporal coupling coefficient, L t―k represents the k-th order light intensity value, β k is the data attenuation factor, k ∈ (1, m), Δτ k is the data time interval.

5. The intelligent fire extinguishing and buffering device for the automatic production line of fireworks and firecrackers according to claim 1, characterized in that, The airflow modeling module includes: A height extraction sub-module that acquires the z-axis height coordinate information in the fire source bounding box, identifies the vertical direction of the bounding box and performs a layering operation at a fixed interval, marks the vertical position of each layer, and generates height stratified coordinates; A pressure difference sub-module that, according to the height stratified coordinates, collects the corresponding pressure monitoring data of each layer through a piezoresistive pressure sensor, calculates the difference between the pressures of adjacent height layers, and statistically analyzes the multi-layer pressure change trend to generate a vertical pressure gradient; An air flow calculation sub-module, based on the vertical pressure gradient, calculates the air flow velocity in the vertical direction of each layer through an aerodynamic function, combines and processes the multi-layer velocity values, extracts the velocity change range in the central region, and generates the vertical air flow velocity.

6. The intelligent fire extinguishing and buffering device for the automated production line of fireworks and firecrackers according to claim 1, characterized in that, The fire extinguishing evaluation module includes: A liquid level acquisition sub-module, which acquires the monitoring data of the capacitive water level sensor, records the real-time liquid level height of the ABC dry powder fire extinguishing agent, performs time continuity screening, identifies the fire extinguishing agent storage according to the stable liquid level section, and generates the fire extinguishing agent liquid level; A flow rate calculation sub-module, based on the fire extinguishing agent liquid level, combines with the vertical air flow velocity to calculate the overall circulation amount of the fire extinguishing agent, and converts it into a mass flow rate with reference to the cross-sectional size of the discharge channel, and generates the fire extinguishing agent mass flow rate; A state judgment sub-module, which calls the fire extinguishing agent mass flow rate and compares it with the fire source bounding box, calculates the corresponding relationship between the flow rate and the bounding box, and triggers an alarm control signal when the fire extinguishing agent mass flow rate value is not sufficient to cover the space area, and generates an alarm signal.

7. The intelligent fire extinguishing and buffering device for the automatic production line of fireworks and firecrackers according to claim 1, characterized in that, The nozzle correction module includes: A flame recognition sub-module, based on the alarm signal, calculates the change amount of the fire source bounding box in the horizontal direction, compares the change rate of the horizontal projection area of the bounding box in adjacent time series and extracts the maximum slope change, obtains the lateral change trend of the fire in this area, and generates the flame propagation rate; A vector construction sub-module, which calls the flame propagation rate and the vertical air flow velocity, calculates the synthesis of the two-direction velocity components, constructs a velocity vector to calibrate the direction, and generates a flame motion vector; An angle correction sub-module, performs arctangent function conversion according to the component ratio of the flame motion vector in the x-z plane, obtains the included angle direction of the vector in the plane, combines with the original emission angle direction of the fire extinguishing nozzle for difference correction processing, and synchronously refreshes the nozzle control instruction parameters, and generates a nozzle azimuth correction angle.

8. The intelligent fire extinguishing and buffering device for the automatic production line of fireworks and firecrackers according to claim 1, wherein, The device further includes: A control decision module, based on the nozzle azimuth correction angle, obtains the new cycle infrared radiation intensity through a dual-band flame sensor and calculates the attenuation rate, processes the attenuation rate and the flame horizontal propagation rate through a proportional-integral controller, generates an optimized injection angle, and outputs an angle adjustment instruction to the actuator; The angle adjustment instruction refers to the correction angle value, the control signal number, and the actuator response time.

9. The intelligent fire extinguishing and buffering device for the automatic production line of fireworks and firecrackers according to claim 8, characterized in that, The control decision module includes: A signal acquisition sub-module, based on the nozzle azimuth correction angle, obtains the infrared radiation intensity collected by the dual-band flame sensor in the current cycle and synchronizes the current time mark, calculates the difference from the infrared radiation intensity at the same position in the previous cycle, and generates the infrared radiation attenuation rate; An attenuation calculation sub-module, which calls the infrared radiation attenuation rate and the flame propagation rate, calculates the response deviation between the two through a proportional-integral controller, performs proportional processing and error accumulation processing on the response deviation, combines the two results and sets upper and lower limit constraints, and generates an angle adjustment correction value; An angle adjustment sub-module, according to the offset difference between the angle adjustment correction value and the current nozzle basic emission angle, re-sets the nozzle rotation angle command parameter in the actuator, generates an optimized injection angle, and outputs an angle adjustment instruction to the actuator.

10. The intelligent fire extinguishing and buffering device for the automated production line of fireworks and firecrackers according to claim 9, characterized in that, The response deviation e(t) between two items is calculated by a proportional-integral controller, using the formula: Among them, C IR represents the current measured value of the infrared radiation attenuation rate, V FS represents the current measured value of the flame propagation rate, represents the upper limit value of the infrared radiation attenuation rate range, K p is the proportionality coefficient adjustment factor, K i is the integral coefficient adjustment factor, T is the dynamic time integration threshold, τ is the integral process time variable, and d is the differential element of the integral process time variable τ.

Images