Explosion-proof electric gate and intelligent control method

Abstract

This invention provides an explosion-proof electric gate and an intelligent control method, belonging to the field of explosion-proof door technology. It includes a top seat, an alloy door, mounting shafts, a roller shutter door, a locking plate, and a lifting mechanism. A protective sleeve is fixedly connected to the top of the top seat. The alloy door is slidably connected within the protective sleeve, penetrating the bottom of the top seat and extending downwards. Two mounting shafts are provided, each rotatably connected to the inner wall of the top seat, with a winding roller fixedly connected to its circumferential surface. Two roller shutter doors are symmetrically arranged, one side of which is wound around the circumferential surface of the corresponding winding roller, and the other end of each door penetrates the top seat and extends downwards, located on both sides of the alloy door. Compared with existing single-layer steel plate structures, this effectively reduces the impact force during an explosion, greatly improving the overall explosion-proof strength. Furthermore, this solution uses a servo motor drive, allowing for remote control of the opening and closing, eliminating the need for manual closure and making it more convenient.

Description

Technical Field

[0001] This invention relates to the application of intelligent information technology in the field of explosion-proof door technology, specifically to explosion-proof electric gates and intelligent control methods. Background Technology

[0002] Explosion-proof doors are explosion-proof protective devices designed to resist accidental explosions of external equipment in industrial buildings, ensuring the safety of personnel and the integrity of internal equipment, protecting them from the shockwaves of explosions and effectively preventing the continuation of explosion hazards. They can be made of special industrial steel plates according to strictly defined mechanical data and equipped with high-performance hardware accessories to protect life and property.

[0003] The prior art discloses Chinese Patent Application No. CN201420419744.2, which discloses an explosion-proof door. The door body and door frame are connected by hinges. A locking mechanism is set inside the door body. A handwheel is set outside the door body. One end of the rotating shaft is sleeved on the handwheel, and the other end of the rotating shaft extends into the door body and is connected to the locking body. A large bevel gear is connected to the rotating shaft. A bevel gear A meshes with the large bevel gear. A lead screw A is set on the upper side of the bevel gear A, and a screw rod A is set on the upper part of the lead screw A. Locking frame A, bevel gear B meshes with large bevel gear, lead screw B is connected to the outer surface of bevel gear B, and locking frame B is connected to the other end of lead screw B. Bevel gear C meshes with large bevel gear, lead screw C is connected to the lower surface of bevel gear C, and locking frame C is connected to the lower part of lead screw C. Bevel gear D meshes with large bevel gear, lead screw D is connected to the outer surface of bevel gear D, and locking frame D is connected to the other end of lead screw D. It has the advantages of wide applicability, simple structure and comprehensive functions.

[0004] Most existing explosion-proof doors close using hinges, and the internal locking mechanism used in the closing mechanism is quite complex. This internal locking mechanism occupies the entire area of ​​the door. When subjected to impact, if any component in the locking mechanism is damaged, the entire locking mechanism will malfunction, resulting in the door being unable to be opened. This is not conducive to rescue and escape and has low safety. Therefore, we propose the explosion-proof electric gate. Summary of the Invention

[0005] The purpose of this invention is to provide an explosion-proof electric gate, aiming to solve the problems mentioned in the background art. Another objective of this invention is that, through an intelligent system driving a servo motor, the explosion-proof gate achieves automatic opening and closing, improving its automation level; by introducing real-time monitoring and machine learning models, the explosion-proof gate can self-monitor its physical state and operating environment, promptly detect and address potential safety issues, thereby improving its safety.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] Explosion-proof electric gates include:

[0008] The top of the base is fixedly connected to a protective sleeve.

[0009] The alloy door is slidably connected inside the protective sleeve, passing through the bottom of the top seat and extending downwards;

[0010] There are two mounting shafts, both of which are rotatably connected to the inner wall of the top seat, and both of which are fixedly connected to a take-up roller on their circumferential surface.

[0011] The roller shutter door has two symmetrically arranged sections. One side of each roller is wrapped around the circumference of the corresponding winding roller, and the other end of each roller passes through the top seat and extends downwards. They are located on both sides of the alloy door.

[0012] A locking plate, located on the underside of the top seat, is fixedly connected to the lower ends of the alloy door and the two roller shutter doors to the sliding sleeve; and

[0013] The lifting mechanism, located inside the top seat, is used to drive the two roller shutter doors to roll up and down synchronously, thereby driving the alloy door to rise and fall to achieve closure.

[0014] As a preferred embodiment of the present invention, the lifting mechanism includes a servo motor, a second gear, a third gear, and two first gears. The servo motor is fixedly connected to the lower inner wall of the top seat, the second gear is fixedly connected to the output end of the servo motor, the third gear is fixedly connected to the circumferential surface of one of the take-up rollers, the third gear meshes with the second gear, and the two first gears are respectively fixedly connected to the ends of different take-up rollers, and the two first gears mesh with each other.

[0015] In a preferred embodiment of the present invention, a self-locking mechanism is provided inside the top seat. The self-locking mechanism includes a ratchet, a first hinge pin, an electric push rod, a diagonal rod, a second hinge pin, and a pawl. The ratchet is fixedly connected to the circumferential surface of another take-up roller. The first hinge pin is fixedly connected to the side wall of the top seat. The bottom of the electric push rod is movably connected to the first hinge pin. The second hinge pin is fixedly connected to the lower inner wall of the top seat. One end of the diagonal rod is movably connected to the extended end of the electric push rod. One end of the pawl is rotatably connected to the second hinge pin. The other end of the pawl engages with the ratchet. The pawl is fixedly connected to the diagonal rod.

[0016] As a preferred embodiment of the present invention, a limiting plate is fixedly connected between the inner walls of the top seat. The limiting plate is located between the two winding rollers. The top of the limiting plate has two first sliding grooves and a second sliding groove. Each of the roller shutter doors passes through the adjacent first sliding groove and extends downward. The alloy door passes through the second sliding groove and extends downward.

[0017] As a preferred embodiment of the present invention, a base is provided on the lower side of the top seat, a protrusion is fixedly connected to the top of the base, and a groove is provided on the bottom of the locking foot plate, with the protrusion fitting into the groove.

[0018] As a preferred embodiment of the present invention, two side guards are fixedly connected between the top seat and the base, and the side guards are located near the edges of the top seat and the base.

[0019] As a preferred embodiment of the present invention, the two side edges of the alloy door and the two roller shutter doors are located within adjacent side guards.

[0020] In a preferred embodiment of the present invention, a straight rod is fixedly connected between the base and the upper inner wall of the two side guards, and a sliding sleeve is slidably connected to the surface of the two straight rods, and the locking foot plate is fixedly connected between the two sliding sleeves.

[0021] As a preferred embodiment of the present invention, a tempered glass window is fixedly connected to the side of the alloy door, and a transparent plate is fixedly connected to both roller shutter doors near the tempered glass window.

[0022] As a preferred embodiment of the present invention, the alloy door has a thickness of 4.00mm-6.00mm, and the surface of the alloy door is welded with multiple reinforcing ribs.

[0023] The control method for an explosion-proof electric gate of the present invention, wherein the explosion-proof electric gate further comprises a sensor group, a data collection module, a data processing module, and a control module connected in sequence, includes the following steps:

[0024] Step 1: The sensor group includes a temperature and humidity sensor, a pressure sensor, an accelerometer or a vibration sensor. The data collection module collects the temperature, humidity, pressure and structural integrity data of the explosion-proof electric gate and sends them to the data processing module periodically.

[0025] Step 2: The control module uses an improved decision tree model method to process and analyze the collected data and predict the door's movement requirements. If any abnormality is detected, the system can take immediate action and respond through built-in mechanisms such as smoke detectors or automatic fire suppression systems.

[0026] Step 3: The control module determines and controls the movement of the servo motor based on the output of the data processing module;

[0027] Step 4: The output of the servo motor (8) drives the second gear (9) to rotate, the second gear (9) drives the third gear (10) to rotate at a reduced speed, and the third gear (10) drives the mounting shaft (4) to rotate.

[0028] Step 5: Under the action of the two first gears (7) meshing with each other, the two winding rollers (5) rotate in opposite directions, causing the two roller shutter doors (6) to move downwards, and at the same time driving the alloy door (3) to move downwards. Finally, the locking foot plate (18) contacts the base (22), and the protrusion (23) is fitted into the groove. Through the three-layer superimposed design of the alloy door (3) and the two roller shutter doors (6), the impact force during the explosion is effectively reduced, which can greatly improve the overall explosion-proof strength.

[0029] The specific process of step 2 is as follows:

[0030] In step 2, the collected data is processed and analyzed using an improved decision tree model method to predict the door's movement requirements. The specific process is as follows:

[0031] Step 2.1, Introduce time series analysis: The operating status of explosion-proof doors and the external environment will change over time. Time can be added as a new feature to the decision tree model. Time series analysis can also be used to preprocess the data.

[0032] Step 2.2: Considering that the humidity and temperature of the external environment may jointly affect the physical state of the explosion-proof door, feature interaction terms are introduced into the decision tree model: one is the product of temperature and humidity; the other is to calculate the usage frequency of the door for each time period, and then use these frequencies as new features.

[0033] Step 2.3: Train multiple decision tree models, and then let these models make decisions together. To improve model performance, a random forest method is used. The specific process is as follows:

[0034] First, it is necessary to collect various characteristic data of the explosion-proof door, including: external environmental parameters such as temperature, humidity, pressure, and time; the usage status of the explosion-proof door such as the number of times the door is opened and closed, the frequency of use, and the duration of use; and the physical condition of the explosion-proof door such as the expansion or contraction of the door and the integrity of the door. Then, these data are preprocessed, including removing outliers and noise, standardization, and generating characteristic interaction terms such as the product of temperature and humidity, and the product of the frequency of use and the duration of use.

[0035] The processed data is divided into training and testing sets: typically 80% of the data is used as the training set and 20% as the testing set.

[0036] Multiple training sets are generated using bootstrap sampling, and a decision tree model is trained on each training set. During training, each decision tree randomly selects a subset of features as candidate split features. Key parameters for training the random forest model include: the number of trees n_estimators, which determines how many decision trees need to be trained; and the maximum number of features max_features, which determines the number of candidate features for each split.

[0037] For a training set D with P features containing N samples, a decision tree can be generated through the following steps:

[0038] a. Use the bootstrap sampling method to draw N samples from the training set D to obtain the training set D_i;

[0039] b. Train a decision tree on the training set D_i, and randomly select p features as candidate features at each split;

[0040] c. Repeat the above steps to generate M decision trees, forming a random forest;

[0041] Let each decision tree in the random forest be T_m, m=1,2,...,M. Then, the prediction results of the random forest can be calculated using the following formula:

[0042] Prediction result = mode(T_1(x), T_2(x), ..., T_M(x))

[0043] Where T_1(x), T_2(x), ..., T_M(x) represent the prediction results of M decision trees for input x, and the mode function is used to find the prediction that appears most frequently among these prediction results, that is, the pattern of these predictions; this formula means that the prediction result of random forest is the mode of all decision tree prediction results;

[0044] The importance of a feature is measured by calculating its Gini importance; for feature j, its Gini importance can be calculated using the following formula:

[0045] Importance(j) =∑_{m=1}^{M} ∑_{i ∈ Nodes_m} [p(i)(1-p(i)) - p_l(i)(1-p_l(i)) - p_r(i)(1-p_r(i))]

[0046]

[0047] Where Gini() is the Gini index, the smaller the index, the higher the purity of the node, D is the sample set in the node, c is the number of classes, and pi is the relative frequency of the i-th class in the node;

[0048] ∑_{m=1}^{M} represents the summation of all values ​​of m from 1 to M. This formula indicates that the Gini importance of feature j is the sum of the Gini impurity decreases of all nodes in the decision tree that use this feature for splitting. Here, Nodes_m represents all nodes in decision tree T_m, p(i) represents the proportion of the main class in node i, and p_l(i) and p_r(i) represent the proportions of the main class in the left and right child nodes after splitting node i, respectively. Through the above formula derivation, a random forest model can be trained based on the feature parameters of the explosion-proof door, and the importance of each feature can be calculated.

[0049] Generation rules:

[0050] 1. Feature Importance Decision Rule:

[0051] Threshold setting: If the Gini index importance of a feature exceeds a certain multiple of the overall average, then the feature is considered to be of high importance;

[0052] 2. Anomaly Detection Decision Rules:

[0053] Temperature anomaly threshold: If the temperature data exceeds the normal range, an anomaly response is triggered;

[0054] Humidity anomaly threshold: If the humidity data exceeds the normal range, an anomaly response is triggered;

[0055] Pressure anomaly threshold: If the pressure data exceeds the normal range, an anomaly response is triggered;

[0056] Structural integrity anomaly threshold: If structural integrity data is abnormal, an anomaly response is triggered;

[0057] 3. Decision rules for time series analysis:

[0058] Time change rate threshold: If time series analysis shows that the operating status of the explosion-proof door and the external environment are affected by time changes, a time change rate threshold is set. It is normal if the time change rate does not exceed the time change rate threshold per hour.

[0059] 4. Feature interaction term decision rules:

[0060] Product feature weight threshold: If the product of temperature and humidity exceeds the normal range, the corresponding weights are adjusted.

[0061] Usage frequency and time product threshold: If the product of the gate's usage frequency and time is abnormal, the corresponding weight adjustment is triggered;

[0062] Abnormal trigger threshold: If the product of usage frequency and time exceeds a certain range, it is considered abnormal;

[0063] 5. Motion prediction decision rules:

[0064] Predicted motion demand threshold: If the predicted door motion demand exceeds the normal range of variation, the motion mode of the servo motor is adjusted in advance.

[0065] Model evaluation and optimization: The model's performance is evaluated using a test set, and the model's parameters are optimized based on the evaluation results. If any anomalies are detected, the system can take immediate action, responding through built-in mechanisms such as smoke detectors or automatic fire suppression systems.

[0066] Compared with the prior art, the beneficial effects of the present invention are:

[0067] 1. In this solution, the output of the servo motor drives the second gear to rotate. Under the action of the two first gears meshing with each other, the two winding rollers rotate in opposite directions, causing the two roller shutters to move downwards. At the same time, the alloy door moves downwards, and finally the locking plate contacts the base, and the protrusion is embedded in the groove. Through the three-layer superimposed design of the alloy door and the two roller shutters, compared with the existing single-layer steel plate structure, the impact force during an explosion is effectively reduced, and the overall explosion-proof strength can be greatly improved. In addition, this solution uses a servo motor drive, which can be remotely controlled to open and close the switch, making it more convenient without manual closing.

[0068] 2. In this design, when the pawl engages with the ratchet, the ratchet cannot move downwards. Normally, when the alloy door is open, it will not close due to the pawl's obstruction, thus preventing the alloy door from falling due to servo motor failure and improving structural safety. Both roller shutters and the alloy door pass through the limiting plate and extend downwards to further limit their movement, preventing them from tilting when moving downwards, reducing deformation caused by shaking, and extending their service life.

[0069] 3. This solution combines real-time monitoring and machine learning prediction technologies to monitor the physical state and operating environment of the explosion-proof door and predict its behavior based on this information. This innovative combination enhances the intelligence and safety of the explosion-proof door. The solution uses a control system to drive the opening and closing of the explosion-proof door, an innovation on traditional operation methods that significantly improves automation and efficiency. The decision tree and random forest models in this solution are specifically designed and trained for the characteristics of explosion-proof doors, enabling them to better adapt to the door's operating state and environment. Furthermore, it innovatively incorporates specific characteristic parameters of the explosion-proof door, such as temperature, pressure, and humidity, into the model, making predictions more accurate and practical. By employing ensemble learning, this solution improves the accuracy and stability of the explosion-proof door status prediction model. Finally, this solution uses an intelligent system to drive a servo motor, achieving automatic opening and closing of the explosion-proof door, greatly improving its automation level and reducing the burden of manual operation.

[0070] 4. This solution introduces real-time monitoring and machine learning models, enabling explosion-proof doors to self-monitor their physical state and operating environment, promptly detect and address potential safety issues, thereby significantly improving the safety of explosion-proof doors. Through the predictive function of the intelligent system, it can predict when the explosion-proof door needs to be opened or closed, thus improving its utilization efficiency. Furthermore, through the predictive functions of real-time monitoring and machine learning models, it can predict and prevent behavioral or environmental factors that may damage the explosion-proof door, thereby extending its service life. Attached Figure Description

[0071] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0072] In the attached diagram:

[0073] Figure 1 This is an overall structural diagram of the present invention;

[0074] Figure 2 This is a cross-sectional view of the present invention;

[0075] Figure 3 This is an exploded view of the top seat of the present invention;

[0076] Figure 4 This is an exploded view of the roller shutter door of the present invention;

[0077] Figure 5 For the present invention Figure 4 Exploded view;

[0078] Figure 6 This is a structural diagram of the limiting plate of the present invention;

[0079] Figure 7 This is a structural diagram of the roller shutter door after installation according to the present invention;

[0080] Figure 8 This is an exploded view of the side guard frame of the present invention;

[0081] Figure 9 This is a structural diagram of the alloy door of the present invention.

[0082] The following are the labeling instructions in the diagram: 1. Top mount; 2. Protective sleeve; 3. Alloy door; 301. Tempered glass window; 4. Mounting shaft; 5. Rewinding roller; 6. Roller shutter door; 601. Transparent panel; 7. First gear; 8. Servo motor; 9. Second gear; 10. Third gear; 11. Ratchet; 12. First hinge pin; 13. Electric push rod; 14. Diagonal rod; 15. Second hinge pin; 16. Pawl; 17. Limiting plate; 1701. First slide groove; 1702. Second slide groove; 18. Locking foot plate; 19. Sliding sleeve; 20. Side guard; 21. Straight rod; 22. Base; 23. Protrusion. Detailed Implementation

[0083] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0084] Example

[0085] Please see Figures 1-9 The technical solution provided in this embodiment is as follows:

[0086] An explosion-proof electric gate includes a top seat 1, an alloy door 3, a mounting shaft 4, a roller shutter door 6, a locking plate 18, and a lifting mechanism. A protective sleeve 2 is fixedly connected to the top of the top seat 1. The alloy door 3 is slidably connected inside the protective sleeve 2, passing through the bottom of the top seat 1 and extending downward. Two mounting shafts 4 are provided, each rotatably connected between the inner walls of the top seat 1, and a winding roller 5 is fixedly connected to the circumferential surface of each shaft. Two roller shutter doors 6 are symmetrically arranged, with one side of each door wound around the circumferential surface of the corresponding winding roller 5, and the other end of each door passing through the top seat 1 and extending downward, located on both sides of the alloy door 3. The locking plate 18 is located on the lower side of the top seat 1, and the lower ends of the alloy door 3 and the two roller shutter doors 6 are fixedly connected to a sliding sleeve 19. The lifting mechanism is located inside the top seat 1 and is used to drive the two roller shutter doors 6 to roll up and down synchronously, thereby driving the alloy door 3 to rise and fall to achieve closure.

[0087] In a specific embodiment of the present invention, the protective sleeve 2 is used to restrict the movement of the alloy door 3. The alloy door 3 is located between two roller shutter doors 6. The winding roller 5 is used to wind up the roller shutter doors 6. When the gate needs to be closed, the output end of the servo motor 8 drives the second gear 9 to rotate. The second gear 9 drives the third gear 10 to rotate at a reduced speed. The third gear 10 drives the mounting shaft 4 to rotate. Under the action of the two first gears 7 meshing with each other, the two winding rollers 5 rotate in opposite directions, causing the two roller shutter doors 6 to move downwards, and at the same time driving the alloy door 3 to move downwards. Finally, the locking foot plate 18 contacts the base 22, and the protrusion 23 is embedded in the groove. Through the three-layer stacked design of the alloy door 3 and the two roller shutter doors 6, compared with the existing single-layer steel plate structure, the impact force during the explosion is effectively reduced, and the overall explosion-proof strength can be greatly improved. In addition, this solution uses the servo motor 8 for driving, and the switch can be controlled remotely without manual closing, which is more convenient. Preferably, the thickness of the alloy door 3 is 4.00mm-6.00mm, and the surface of the alloy door 3 is welded with multiple reinforcing ribs, resulting in high structural strength and strong impact resistance.

[0088] Specifically, the lifting mechanism includes a servo motor 8, a second gear 9, a third gear 10, and two first gears 7. The servo motor 8 is fixedly connected to the lower inner wall of the top seat 1. The second gear 9 is fixedly connected to the output end of the servo motor 8. The third gear 10 is fixedly connected to the circumferential surface of one of the take-up rollers 5. The third gear 10 meshes with the second gear 9. The two first gears 7 are respectively fixedly connected to the ends of different take-up rollers 5. The two first gears 7 mesh with each other.

[0089] In a specific embodiment of the present invention, see [reference]. Figure 3 , Figure 4 and Figure 5 The servo motor 8 is located below the third gear 10. The output of the servo motor 8 drives the second gear 9 to rotate. The second gear 9 drives the third gear 10 to rotate in the opposite direction. The torque of the third gear 10 is greater than that of the second gear 9, so as to increase the torque and make it easier to drive the third gear 10 to rotate. The first gear 7 and the third gear 10 are located on different sides of the mounting shaft 4. When one of the first gears 7 rotates, it drives the other first gear 7 to rotate in the opposite direction synchronously. Finally, it drives the two winding rollers 5 to rotate in the opposite direction synchronously. The roller shutter door 6 unrolls or rerolls with the rotation of the winding rollers 5. When it needs to be closed, the roller shutter door 6 unrolls and moves downward, which in turn drives the locking foot plate 18 to move downward. The alloy door 3 moves downward with the locking foot plate 18, and the locking foot plate 18 provides the gravity for the downward movement. When the locking foot plate 18 contacts the base 22, the closure is completed. At this time, the three-layer protection structure composed of the alloy door 3 and the two roller shutter doors 6 can greatly enhance the explosion-proof performance and protect the safety of citizens' property and lives.

[0090] Specifically, a self-locking mechanism is provided inside the top seat 1. The self-locking mechanism includes a ratchet 11, a first hinge pin 12, an electric push rod 13, a diagonal bar 14, a second hinge pin 15, and a pawl 16. The ratchet 11 is fixedly connected to the circumferential surface of another take-up roller 5. The first hinge pin 12 is fixedly connected to the side wall of the top seat 1. The bottom of the electric push rod 13 is movably connected to the first hinge pin 12. The second hinge pin 15 is fixedly connected to the lower inner wall of the top seat 1. One end of the diagonal bar 14 is movably connected to the extended end of the electric push rod 13. One end of the pawl 16 is rotatably connected to the second hinge pin 15. The other end of the pawl 16 is engaged with the ratchet 11. The pawl 16 is fixedly connected to the diagonal bar 14.

[0091] In specific embodiments of the present invention, such as Figure 4 and Figure 5 As shown, ratchet 11 is used to follow the rotation of mounting shaft 4, and pawl 16 is used to prevent the roller shutter door 6 from moving downward. When the roller shutter door 6 needs to move downward, the extended end of the electric push rod 13 is controlled to push the inclined rod 14 upward. The inclined rod 14 drives the pawl 16 to move, so that the pawl 16 is separated from the ratchet 11. At this time, the mounting shaft 4 can rotate in both directions. When the pawl 16 is engaged with the ratchet 11, the ratchet 11 cannot move downward. Normally, when the alloy door 3 is in the open state, the alloy door 3 will not close due to the obstruction of the pawl 16, which avoids the situation where the alloy door 3 falls due to the failure of the servo motor 8 and improves the structural safety.

[0092] Specifically, a limiting plate 17 is fixedly connected between the inner walls of the top seat 1. The limiting plate 17 is located between the two winding rollers 5. The top of the limiting plate 17 has two first grooves 1701 and second grooves 1702. Each roller shutter door 6 passes through the adjacent first groove 1701 and extends downward. The alloy door 3 passes through the second groove 1702 and extends downward.

[0093] In specific embodiments of the present invention, such as Figure 2 and Figure 6 As shown, both roller shutter doors 6 and alloy door 3 pass through the limiting plate 17 and extend downwards to further limit the roller shutter doors 6 and alloy door 3, preventing them from tilting when moving downwards, reducing deformation caused by shaking, and extending their service life.

[0094] Specifically, a base 22 is provided on the lower side of the top seat 1, and a protrusion 23 is fixedly connected to the top of the base 22. A groove is provided at the bottom of the locking foot plate 18, and the protrusion 23 is fitted into the groove.

[0095] In specific embodiments of the present invention, such as Figure 2 and Figure 1As shown, the groove is located at the bottom of the locking plate 18. When the locking plate 18 contacts the base 22, the protrusion 23 is fitted into the groove. At this time, the bottom of the locking plate 18 is more stable under the obstruction of the protrusion 23. When installed, the base 22 is fitted into the ground, the top of the base 22 is flush with the ground, and the protrusion 23 is exposed on the ground and has an arc-shaped structure, so that pedestrians will not bump into it when walking.

[0096] Specifically, two side guards 20 are fixedly connected between the top seat 1 and the base 22. The side guards 20 are located near the edges of the top seat 1 and the base 22. The two sides of the alloy door 3 and the two roller shutter doors 6 are located within the adjacent side guards 20.

[0097] In a specific embodiment of the present invention, see [reference]. Figure 1 The side guard 20 is used to support the top seat 1. When the roller shutter door 6 and the alloy door 3 move downward, the edges on both sides are located inside the side guard 20, reducing the gaps on both sides of the alloy door 3.

[0098] Specifically, a straight rod 21 is fixedly connected between the base 22 and the upper inner wall of the two side guards 20, and a sliding sleeve 19 is slidably connected to the surface of the two straight rods 21. The locking foot plate 18 is fixedly connected between the two sliding sleeves 19.

[0099] In specific embodiments of the present invention, such as Figure 7 and Figure 8 As shown, the sliding sleeve 19 slides on the surface of the straight rod 21, causing the locking foot plate 18 to move up and down, and then move linearly under the limiting action of the sliding sleeve 19.

[0100] Specifically, a tempered glass window 301 is fixedly connected to the side of the alloy door 3, and a transparent panel 601 is fixedly connected to both roller shutter doors 6 near the tempered glass window 301.

[0101] In a specific embodiment of the present invention, the tempered glass window 301 and the transparent plate 601 are of the same height and are made of uniformly transparent material, which facilitates observation of the internal situation when the gate is closed.

[0102] In a specific embodiment of the present invention, the control method for the explosion-proof electric gate of the present invention, wherein the explosion-proof electric gate further includes a sensor group, a data collection module, a data processing module, and a control module connected in sequence, includes the following steps:

[0103] Step 1: The sensor group includes a temperature and humidity sensor, a pressure sensor, an accelerometer or a vibration sensor. The data collection module collects the temperature, humidity, pressure and structural integrity data of the explosion-proof electric gate and sends them to the data processing module periodically.

[0104] Step 2: The control module uses an improved decision tree model method to process and analyze the collected data and predict the door's movement requirements. If any abnormality is detected, the system can take immediate action and respond through built-in mechanisms such as smoke detectors or automatic fire suppression systems.

[0105] Step 3: The control module determines and controls the movement of the servo motor based on the output of the data processing module;

[0106] Step 4: The output of the servo motor (8) drives the second gear (9) to rotate, the second gear (9) drives the third gear (10) to rotate at a reduced speed, and the third gear (10) drives the mounting shaft (4) to rotate.

[0107] Step 5: Under the action of the two first gears (7) meshing with each other, the two winding rollers (5) rotate in opposite directions, causing the two roller shutter doors (6) to move downwards, and at the same time driving the alloy door (3) to move downwards. Finally, the locking foot plate (18) contacts the base (22), and the protrusion (23) is fitted into the groove. Through the three-layer superimposed design of the alloy door (3) and the two roller shutter doors (6), the impact force during the explosion is effectively reduced, which can greatly improve the overall explosion-proof strength.

[0108] In step 2, the collected data is processed and analyzed using an improved decision tree model method to predict the door's movement requirements. The specific process is as follows:

[0109] Step 2.1, Introducing Time Series Analysis: The operating status of explosion-proof doors and the external environment change over time. For example, the frequency of door use and the external environment may differ at different times of day or in different seasons. Therefore, time factors can be incorporated into the decision tree model. Specifically, time (such as the number of hours in a day, the number of months in a year, etc.) can be added as a new feature to the decision tree model, or time series analysis methods (such as autocorrelation, trend analysis, etc.) can be used to preprocess the data.

[0110] The specific implementation process is as follows: First, define the following characteristic parameters of the explosion-proof door, which will change over time: the frequency of use of the door (number of times used per hour or per day), the external ambient temperature, the external ambient humidity, the season (spring, summer, autumn, winter), and the time of day (morning, noon, afternoon, evening).

[0111] These feature data are recorded in time series format, and then the feature data is used to train a decision tree model.

[0112] 1. Generation of Time Series Features

[0113] When generating time series features, the raw data first needs to be preprocessed. This may include: converting the season and time periods of day into a set of dummy variables (also known as one-hot encoding); standardizing continuous variables such as frequency, temperature, and humidity so that their mean is 0 and their standard deviation is 1; and then using a sliding window method to generate time series features. Specifically, for each predicted time t, data from the previous n times can be considered, and the statistical properties of these data, such as mean, variance, trend, and seasonality, can be calculated as input features for the decision tree model.

[0114] 2. Training and Application of Decision Tree Models

[0115] In training the decision tree model, the generated time-series features and the state of the explosion-proof door (e.g., whether it needs to be opened or closed) are used as input data. Then, methods such as information gain are used to select the optimal features and split points. In applying the model, the current time-series features are first calculated based on the current time and past feature data. These features are then input into the decision tree model to obtain the predicted state of the explosion-proof door. To improve the model's accuracy and stability, methods such as cross-validation and grid search can be used to adjust the model's parameters, and new data information can be fed back in a timely manner through model updates and retraining.

[0116] Step 2.2, consider the interaction effect of features: In reality, different features may interact with each other. For example, the humidity and temperature of the external environment may jointly affect the physical state of a blast door. To capture this interaction effect, feature interaction terms can be introduced into the decision tree model. Specifically, in addition to the original features, we can also add the product of some features as new features.

[0117] The specific implementation process is as follows:

[0118] To better capture the interaction effects between the characteristic parameters of explosion-proof doors, feature interaction terms will be introduced into the decision tree model. Specifically, the interaction effects of the following features will be considered:

[0119] External environmental temperature and humidity: These can jointly affect the physical state of the explosion-proof door, such as its expansion or contraction. Therefore, we can introduce a new feature: the product of temperature and humidity.

[0120] Door usage frequency and time of day: Door usage frequency may vary at different times of the day. Therefore, the door usage frequency can be calculated separately for each time period, and then these frequencies can be used as new features.

[0121] The specific solutions are as follows:

[0122] 1. Generation of Feature Interaction Items

[0123] When generating feature interaction terms, we first need to preprocess the raw data. For example, we might need to convert time periods of the day into a set of dummy variables (also known as one-hot encoding) and standardize continuous variables such as temperature and humidity. Then, we can directly calculate the product of two features to obtain the feature interaction terms. For example, we can calculate the product of ambient temperature and humidity as a new feature.

[0124] 2. Training and Application of Decision Tree Models

[0125] In training the decision tree model, we use the original features, time-series features, and feature interaction terms as input data, and then use methods such as information gain to select the optimal features and split points. In applying the model, we first calculate the feature interaction terms based on the current feature data, and then input these features (including the original features, time-series features, and feature interaction terms) into the decision tree model to obtain the predicted state of the explosion-proof door. To improve the model's accuracy and stability, we can use methods such as cross-validation and grid search to adjust the model's parameters, and update and retrain the model to promptly provide feedback on new data information.

[0126] Step 2.3, using ensemble learning: Ensemble learning is an effective method to improve model performance. In this method, we can train multiple decision tree models and then let these models make decisions together. Specifically, we can use Bagging or Boosting methods to train a set of decision trees, and then merge the predictions of these decision trees through voting or weighted averaging. This can reduce the model's variance and improve its stability and accuracy.

[0127] The specific implementation process is as follows:

[0128] To improve model performance, we will use ensemble learning methods to train and apply the decision tree model. Specifically, we will use the Random Forest method, which is a decision tree ensemble model based on the Bagging method.

[0129] 1. Data Collection and Preprocessing

[0130] First, we need to collect various characteristic data of the explosion-proof door, including: external environmental parameters such as temperature, humidity, pressure, time, etc.

[0131] Information on the use of explosion-proof doors: such as the number of times the door is opened and closed, the frequency of use, and the duration of use.

[0132] The physical condition of the explosion-proof door: such as the expansion or contraction of the door, the integrity of the door, etc.

[0133] Next, we need to preprocess this data, including data cleaning (removing outliers and noise), standardization (converting the data into a form with a mean of 0 and a standard deviation of 1), and generation of feature interaction terms (e.g., the product of temperature and humidity, the product of door usage frequency and time, etc.).

[0134] 2. Division of training and test sets

[0135] Next, we will divide the processed data into training and testing sets. Typically, 80% of the data will be used as the training set, and 20% as the testing set. The training set will be used to train the model, and the testing set will be used to evaluate its performance.

[0136] 3. Training the Random Forest Model

[0137] Next, we begin training the random forest model. During training, multiple training sets are generated using bootstrap sampling, and a decision tree model is trained on each training set. During training, each decision tree randomly selects a subset of features as candidate splitting features.

[0138] Key parameters for training a random forest model include: Number of trees (n_estimators): This determines how many decision trees we need to train. Generally, more trees result in better model performance, but also increase computation time and resource consumption. Maximum number of features (max_features): This determines the number of candidate features for each split. Generally, a higher maximum number of features increases model complexity and the risk of overfitting. These parameters can be optimized using methods such as grid search or random search.

[0139] 1. Random Forest consists of multiple decision trees. Each decision tree is trained independently, and bootstrap sampling and random selection of feature subsets are used during training.

[0140] For a training set D with P features containing N samples, we can generate a decision tree using the following steps:

[0141] a. Use the bootstrap sampling method to draw N samples from the training set D to obtain the training set D_i.

[0142] b. Train a decision tree on the training set D_i, and randomly select p features as candidate features at each split.

[0143] c. Repeat the above steps to generate M decision trees, forming a random forest.

[0144] Let T_m be the decision tree in the random forest, where m = 1, 2, ..., M. Then, the prediction results of the random forest can be calculated using the following formula:

[0145] Prediction result = mode(T_1(x), T_2(x), ..., T_M(x))

[0146] This formula indicates that the prediction result of a random forest is the mode of all decision tree prediction results.

[0147] 2. Feature Importance

[0148] In random forests, we can measure the importance of a feature by calculating its Gini importance. For feature j, its Gini importance can be calculated using the following formula:

[0149] Importance(j) =∑_{m=1}^{M} ∑_{i ∈ Nodes_m} [p(i)(1-p(i)) - p_l(i)(1-p_l(i)) - p_r(i)(1-p_r(i))]

[0150]

[0151] Where Gini() is the Gini index, the smaller the index, the higher the purity of the node; D is the sample set in the node; c is the number of classes; and pi is the relative frequency of the i-th class in the node. ∑_{m=1}^{M} represents the summation of all values ​​of m from 1 to M. This formula indicates that the Gini importance of feature j is the sum of the reductions in Gini impurity of all nodes in the decision tree that use this feature for splitting. Nodes_m represents all nodes in the decision tree T_m; p(i) represents the proportion of the main class in node i; and p_l(i) and p_r(i) represent the proportions of the main class in the left and right child nodes after splitting node i, respectively. Through the above formula derivation, we can train a random forest model based on the feature parameters of the explosion-proof door and calculate the importance of each feature.

[0152] Generation rules:

[0153] 1. Feature Importance Decision Rule:

[0154] Threshold setting: If the Gini importance of a feature exceeds 1.5 times the overall average, the feature is considered to have high importance. For example, in a random forest model, the average Gini importance of all features is 0.1. Therefore, if the Gini importance of a feature exceeds 1.5 times the overall average, its threshold can be set to 0.1 × 1.5 = 0.15.

[0155] 2. Anomaly Detection Decision Rules:

[0156] Temperature anomaly threshold: If the temperature data exceeds the normal range (e.g., the range is 20 to 80 degrees Celsius), an anomaly response is triggered.

[0157] Humidity anomaly threshold: If the humidity data exceeds the normal range (e.g., the humidity range is 20% to 80%), an anomaly response is triggered.

[0158] Pressure anomaly threshold: If the pressure data exceeds the normal range (e.g., 0 to 100 kPa), an anomaly response is triggered.

[0159] Structural integrity anomaly threshold: If structural integrity data is abnormal, such as the expansion or contraction of a door exceeding the normal range (0-5% is considered normal), an abnormal response is triggered.

[0160] 3. Decision rules for time series analysis:

[0161] Time change rate threshold: If time series analysis shows that the operating status of the explosion-proof door and the external environment are affected by time changes, set a time change rate threshold, for example, no more than 5% per hour.

[0162] 4. Feature interaction term decision rules:

[0163] Product feature weight threshold: If the product of temperature and humidity exceeds the normal range, the corresponding weight is adjusted. If the normal range is temperature (humidity) between 20 (0.2) and 80 (0.8) respectively, then the normal range of the product can be set to 20×0.2=4 to 80×0.8=64.

[0164] Usage frequency and time product threshold: If the product of the door's usage frequency and time is abnormal, a corresponding weight adjustment is triggered. Normal range: Usage frequency is 1 to 10 times per hour, and time is 8 hours of operation per day, i.e., 1×8=8 to 10×8=80.

[0165] Abnormal trigger threshold: If the product of usage frequency and time is less than 4 or greater than 100, it is considered abnormal.

[0166] 5. Motion prediction decision rules:

[0167] Predicted motion demand threshold: If the predicted door motion demand exceeds the normal range of variation, the motion mode of the servo motor is adjusted in advance, for example, the motion time is advanced by 10%.

[0168] 4. Model Evaluation and Optimization

[0169] Finally, the model's performance is evaluated using a test set, and the model's parameters are optimized based on the evaluation results. We primarily focus on the model's accuracy and recall, because we want the model to accurately predict the state of the explosion-proof door while also detecting as many anomalies as possible.

[0170] The working principle or process of the explosion-proof electric gate provided by the present invention is as follows: the output end of the servo motor 8 drives the second gear 9 to rotate, the second gear 9 drives the third gear 10 to rotate at a reduced speed, the third gear 10 drives the mounting shaft 4 to rotate, and under the action of the two first gears 7 meshing with each other, the two winding rollers 5 rotate in opposite directions, causing the two roller shutter doors 6 to move downward, and at the same time driving the alloy door 3 to move downward, finally causing the locking foot plate 18 to contact the base 22, and the protrusion 23 to be embedded in the groove. Through the three-layer superimposed design of the alloy door 3 and the two roller shutter doors 6, compared with the existing single-layer steel plate structure, the impact force during the explosion is effectively reduced, and the overall explosion-proof strength can be greatly improved.

[0171] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An explosion-proof electric gate, characterized in that, include: Top seat (1), with a protective sleeve (2) fixedly connected to its top; Alloy door (3), which is slidably connected inside protective sleeve (2), penetrates the bottom of top seat (1) and extends downward; There are two mounting shafts (4), both of which are rotatably connected between the inner walls of the top seat (1), and both of which are fixedly connected to a take-up roller (5) on their circumferential surfaces. There are two roller shutter doors (6) symmetrically arranged. One side of each door is wrapped around the circumferential surface of the corresponding winding roller (5), and the other end of each door passes through the top seat (1) and extends downward. They are located on both sides of the alloy door (3). A locking foot plate (18) is located on the lower side of the top seat (1), and the lower ends of the alloy door (3) and the two roller shutter doors (6) are fixedly connected to the sliding sleeve (19). And a lifting mechanism, which is located in the top seat (1), is used to drive the two roller shutter doors (6) to roll up and down synchronously, and then drive the alloy door (3) to lift up and down to achieve closure; The lifting mechanism includes a servo motor (8), a second gear (9), a third gear (10), and two first gears (7). The servo motor (8) is fixedly connected to the lower inner wall of the top seat (1). The second gear (9) is fixedly connected to the output end of the servo motor (8). The third gear (10) is fixedly connected to the circumferential surface of one of the take-up rollers (5). The third gear (10) meshes with the second gear (9). The two first gears (7) are respectively fixedly connected to the ends of different take-up rollers (5). The two first gears (7) mesh with each other. The top seat (1) is provided with a self-locking mechanism, which includes a ratchet (11), a first hinge (12), an electric push rod (13), a diagonal rod (14), a second hinge (15), and a pawl (16). The ratchet (11) is fixedly connected to the circumferential surface of another take-up roller (5). The first hinge (12) is fixedly connected to the side wall of the top seat (1). The bottom of the electric push rod (13) is movably connected to the first hinge (12). The second hinge (15) is fixedly connected to the lower inner wall of the top seat (1). One end of the diagonal rod (14) is movably connected to the extended end of the electric push rod (13). One end of the pawl (16) is rotatably connected to the second hinge (15). The other end of the pawl (16) is engaged with the ratchet (11). The pawl (16) is fixedly connected to the diagonal rod (14).

2. The explosion-proof electric gate according to claim 1, characterized in that, A limiting plate (17) is fixedly connected between the inner walls of the top seat (1). The limiting plate (17) is located between the two take-up rollers (5). The top of the limiting plate (17) has two first grooves (1701) and a second groove (1702). Each of the roller shutter doors (6) passes through the adjacent first groove (1701) and extends downward. The alloy door (3) passes through the second groove (1702) and extends downward.

3. The explosion-proof electric gate according to claim 2, characterized in that, A base (22) is provided on the lower side of the top seat (1), and a protrusion (23) is fixedly connected to the top of the base (22). A groove is provided at the bottom of the locking foot plate (18), and the protrusion (23) is fitted into the groove.

4. The explosion-proof electric gate according to claim 3, characterized in that, Two side guards (20) are fixedly connected between the top seat (1) and the base (22), and the side guards (20) are located near the edges of the top seat (1) and the base (22).

5. The explosion-proof electric gate according to claim 4, characterized in that, The two sides of the alloy door (3) and the two roller shutter doors (6) are located within the adjacent side guards (20).

6. The explosion-proof electric gate according to claim 5, characterized in that, Straight rods (21) are fixedly connected between the base (22) and the upper inner walls of the two side guards (20). Sliding sleeves (19) are slidably connected to the surfaces of the two straight rods (21). The locking foot plate (18) is fixedly connected between the two sliding sleeves (19). The alloy door (3) is fixedly connected to a tempered glass window (301) on its side, and the two roller shutter doors (6) are fixedly connected to a transparent plate (601) near the tempered glass window (301); The alloy door (3) has a thickness of 4.00mm-6.00mm, and multiple reinforcing ribs are welded to the surface of the alloy door (3).

7. The control method for the explosion-proof electric gate according to claim 1, characterized in that, The explosion-proof electric gate also includes a sensor group, a data collection module, a data processing module, and a control module connected in sequence, and includes the following steps: Step 1: The sensor group includes a temperature and humidity sensor, a pressure sensor, an accelerometer or a vibration sensor. The data collection module collects the temperature, humidity, pressure and structural integrity data of the explosion-proof electric gate and sends them to the data processing module periodically. Step 2: The control module uses an improved decision tree model method to process and analyze the collected data and predict the door's movement requirements. If any abnormality is detected, the system takes immediate action and responds through built-in mechanisms such as smoke detectors or automatic fire suppression systems. Step 3: The control module determines and controls the movement of the servo motor based on the output of the data processing module; Step 4: The output of the servo motor (8) drives the second gear (9) to rotate, the second gear (9) drives the third gear (10) to rotate at a reduced speed, and the third gear (10) drives the mounting shaft (4) to rotate. Step 5: Under the action of the two first gears (7) meshing with each other, the two winding rollers (5) rotate in opposite directions, causing the two roller shutter doors (6) to move downwards, and at the same time driving the alloy door (3) to move downwards. Finally, the locking foot plate (18) contacts the base (22), and the protrusion (23) is fitted into the groove. Through the three-layer overlapping design of the alloy door (3) and the two roller shutter doors (6), the impact force during the explosion is effectively reduced, which can greatly improve the overall explosion-proof strength.

8. The control method for the explosion-proof electric gate according to claim 7, characterized in that, The specific process of step 2 is as follows: In step 2, the collected data is processed and analyzed using an improved decision tree model method to predict the door's movement requirements. The specific process is as follows: Step 2.1, Introduce time series analysis: The operating status of explosion-proof doors and the external environment will change over time. Time is added as a new feature to the decision tree model, and time series analysis is also used to preprocess the data. Step 2.2: Considering that the humidity and temperature of the external environment may jointly affect the physical state of the explosion-proof door, feature interaction terms are introduced into the decision tree model: one is the product of temperature and humidity; the other is to calculate the usage frequency of the door for each time period, and then use these frequencies as new features. Step 2.3: Train multiple decision tree models, and then let these models make decisions together. To improve model performance, a random forest method is used. The specific process is as follows: First, it is necessary to collect various characteristic data of the explosion-proof door, including: external environmental parameters such as temperature, humidity, pressure, and time; the usage status of the explosion-proof door, such as the number of times the door is opened and closed, the frequency of use, and the duration of use; and the physical condition of the explosion-proof door, including the expansion or contraction of the door and the integrity of the door. Then, these data are preprocessed, including removing outliers and noise, standardization, and generating characteristic interaction terms: the product of temperature and humidity, and the product of the frequency of use and the duration of use. The processed data is divided into training and testing sets: typically 80% of the data is used as the training set and 20% as the testing set. Multiple training sets are generated using bootstrap sampling, and a decision tree model is trained on each training set. During training, each decision tree randomly selects a subset of features as candidate split features. Key parameters for training the random forest model include: the number of trees n_estimators, which determines how many decision trees need to be trained; and the maximum number of features max_features, which determines the number of candidate features for each split. For a training set D with P features containing N samples, a decision tree is generated through the following steps: a. Use the bootstrap sampling method to draw N samples from the training set D to obtain the training set D_i; b. Train a decision tree on the training set D_i, and randomly select p features as candidate features at each split; c. Repeat the above steps to generate M decision trees, forming a random forest; Let each decision tree in the random forest be T_m, where m = 1, 2, ..., M. Then, the prediction results of the random forest are calculated using the following formula: Prediction result = mode(T_1(x),T_2(x),...,T_M(x)) Where T_1(x), T_2(x), ..., T_M(x) represent the prediction results of M decision trees for input x, and the mode function is used to find the prediction that appears most frequently among these prediction results, that is, the pattern of these predictions; this formula means that the prediction result of random forest is the mode of all decision tree prediction results. The importance of a feature is measured by calculating its Gini importance; for feature j, its Gini importance is calculated using the following formula: Importance(j)=∑_{m=1}^{M}∑_{i∈Nodes_m}[p(i)(1-p(i))-p_l(i)(1-p_l(i))-p_r(i)(1-p_r(i))] Where Gini() is the Gini index, the smaller the index, the higher the purity of the node, D is the sample set in the node, c is the number of classes, and pi is the relative frequency of the i-th class in the node; ∑_{m=1}^{M} represents the summation of all values ​​of m from 1 to M. This formula indicates that the Gini importance of feature j is the sum of the Gini impurity decreases of all nodes in the decision tree that use this feature for splitting. Here, Nodes_m represents all nodes in decision tree T_m, p(i) represents the proportion of the main class in node i, and p_l(i) and p_r(i) represent the proportions of the main class in the left and right child nodes after splitting node i, respectively. Based on the above formula derivation, a random forest model is trained according to the feature parameters of the explosion-proof door, and the importance of each feature is calculated. Generation rules:

1. Feature Importance Decision Rule: Threshold setting: If the Gini index importance of a feature exceeds a certain multiple of the overall average, then the feature is considered to be of high importance; 2. Anomaly Detection Decision Rules: Temperature anomaly threshold: If the temperature data exceeds the normal range, an anomaly response is triggered; Humidity anomaly threshold: If the humidity data exceeds the normal range, an anomaly response is triggered; Pressure anomaly threshold: If the pressure data exceeds the normal range, an anomaly response is triggered; Structural integrity anomaly threshold: If structural integrity data is abnormal, an anomaly response is triggered; 3. Decision rules for time series analysis: Time change rate threshold: If time series analysis shows that the operating status of the explosion-proof door and the external environment are affected by time changes, a time change rate threshold is set. It is normal if the time change rate does not exceed the time change rate threshold per hour.

4. Feature interaction term decision rules: Product feature weight threshold: If the product of temperature and humidity exceeds the normal range, the corresponding weights are adjusted. Usage frequency and time product threshold: If the product of the gate's usage frequency and time is abnormal, the corresponding weight adjustment is triggered; Abnormal trigger threshold: If the product of usage frequency and time exceeds a certain range, it is considered abnormal; 5. Motion prediction decision rules: Predicted motion demand threshold: If the predicted door motion demand exceeds the normal range of variation, the motion mode of the servo motor is adjusted in advance. Model evaluation and optimization: The model's performance is evaluated using a test set, and the model's parameters are optimized based on the evaluation results. If any anomalies are detected, the system takes immediate action, responding through built-in mechanisms such as smoke detectors or automatic fire suppression systems.

Images