Belt conveyor chute blockage detection system and method based on AI agent

Abstract

This invention discloses a detection system and method for material blockage in a conveyor belt chute based on an AI agent. The system includes: a power and detection component, comprising a motor, a rotating shaft driven by the motor, and a rotating component connected to the rotating shaft, the rotating component being placed in the conveyor belt chute to be detected; a signal sensing component, comprising a rotation sensor for detecting the rotational speed of the rotating component; a torque overload protection component, comprising a magnetic coupler disposed between the motor and the rotating component, wherein if the rotation of the rotating component is obstructed by material blockage, causing the rotational torque of the magnetic coupler to exceed a preset mechanical protection torque threshold, the magnetic coupler disengages the transmission between the motor and the rotating component; and an intelligent control and decision-making component, comprising an AI agent communicatively connected to the rotation sensor, the AI ​​agent being configured to analyze and judge the rotational state of the rotating component within a preset time window based on the detection signal from the rotation sensor, in order to identify the material blockage status of the chute.

Description

Technical Field

[0001] This invention relates to the field of AI vision, and in particular to a detection system and method for detecting material blockage in a conveyor belt chute based on an AI intelligent agent. Background Technology

[0002] In the transportation process of conveyor belts for metallurgical smelting, various powder materials need to be transported and fed on time. However, chute blockage is a common fault in material conveying systems, which may lead to material accumulation, equipment overload, belt tearing, motor burnout, or even production stoppage. This affects the continuity and efficiency of the entire production line. Safe, accurate, and reliable detection of chute blockage and timely issuance of alarms or control signals are crucial for ensuring safe and stable production operation.

[0003] In the production practice of steel enterprises, some non-contact detection devices, such as ultrasonic or radio frequency admittance devices, are used. Although they do not require direct contact with materials, they are more sensitive to the cleanliness of the installation environment, dust concentration, and material accumulation. Under certain working conditions, their stability may not be as good as mechanical detection devices.

[0004] However, existing mechanical detection devices are generally unable to detect minor initial blockages in a timely manner; and due to installation positions that are too low or angles that are not suitable, they may impact the material, leading to malfunctions or premature damage to the detection mechanism itself.

[0005] The disclosure of the above background technical content is only for the purpose of assisting in understanding the concept and technical solution of this application, and does not necessarily provide technical instruction. Summary of the Invention

[0006] The purpose of this invention is to provide a detection system and method for material blockage in conveyor belt chutes based on AI intelligent agents, which uses non-transient detection signals from a contact detection device to achieve accurate detection of the material blockage status in the chutes.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A detection system for material blockage in a belt conveyor chute based on an AI intelligent agent includes: A power and detection assembly includes a motor, a shaft driven by the motor, and a rotating component connected to the shaft, the rotating component being configured to be disposed in a conveyor belt chute to be detected; A signal sensing component, comprising a rotation sensor for detecting the rotational speed of the rotating component; A torque overload protection assembly includes a magnetic coupler disposed between the motor and the rotating component. If the rotation of the rotating component is obstructed by a blockage, causing the rotational torque of the magnetic coupler to exceed a preset mechanical protection torque threshold, the magnetic coupler disengages the transmission between the motor and the rotating component. The intelligent control and decision-making component includes an AI agent that is communicatively connected to the rotation sensor. The AI ​​agent is configured to analyze and judge the rotation state of the rotating component within a preset time window based on the detection signal of the rotation sensor, so as to identify the blockage status of the chute.

[0008] Furthermore, following any one or a combination of the aforementioned technical solutions, the AI ​​agent integrates a temporal deep learning model, and the learning samples used for its training at least include learning samples obtained through the following methods: If, after a preset time t1 from the point t0 when the rotating component stops rotating, and without any intervention, the chute automatically unloads material, causing the rotating component to resume rotation within a preset recovery time t2, then the detection data of the rotation sensor corresponding to the time window is acquired as a learning sample. The starting point of the time window is t0-t2. pre1 , t pre1 This indicates the preset lead time, the end point of the time window is t0+t1+t2, and the label for this learning sample is set as no blockage. If the rotating component does not resume rotation within the preset recovery time t2, then the same time window [t0-t] is obtained. pre1 The detection data of the rotary sensor corresponding to [t0+t1+t2] is used as a learning sample, and the learning sample is labeled as blockage.

[0009] Furthermore, following any one or a combination of the aforementioned technical solutions, the AI ​​agent integrates a temporal deep learning model, and the learning samples used for its training at least include learning samples obtained through the following methods: If, after a preset time t3 from the point t0 when the rotating component stops rotating, the component resumes rotation within a preset recovery time t4, and then stops rotating again during a preset time t5 after resuming rotation, then the time window [t0-t] is obtained. pre2 The detection data of the rotary sensor corresponding to [t0+t3+t4+t5] is used as a learning sample, and the label of this learning sample is set as blockage, where t pre2 Indicates the preset lead time; Alternatively, a preset time t3 is defined from the time point t0 when the rotating component stops rotating, and the rotating component resumes rotation within a preset recovery time t4. If the rotating component stops rotating again during the preset time t5 after resuming rotation, this is considered one cycle. If more than two cycles occur within the preset time window, the detection data of the rotation sensor corresponding to the time window is obtained as a learning sample, and the learning sample is labeled as "blockage".

[0010] Furthermore, following any one or a combination of the aforementioned technical solutions, the number of rotating components is multiple, and they have any one of the following forms: The rotating component is a blade, and each blade is connected to the rotating shaft via a spring torsion shaft, so that each blade can be independently deflected relative to the rotating shaft; Alternatively, the rotating component is a spring, with one end of each spring fixedly connected to the rotating shaft, and the free end of each spring independently deflectable relative to the rotating shaft.

[0011] Furthermore, based on any or a combination of the aforementioned technical solutions, the magnetic coupler includes an active rotor and a driven rotor, wherein the active rotor is fixedly connected to the output shaft of the motor, and the driven rotor is fixedly connected to the rotating shaft of the connecting rotating component; The active rotor and the driven rotor do not make physical contact. The active rotor generates a magnetic field as it rotates with the motor output shaft, thereby driving the driven rotor to rotate with the shaft and rotating components.

[0012] Furthermore, in accordance with any or a combination of the aforementioned technical solutions, a transmission sleeve is fixedly provided on the rotating shaft, and a friction plate is fixedly provided on the rotating component. The friction plate is pressed against the transmission sleeve by a spring, wherein the spring is subject to the preload of an adjusting nut. The mechanical protection torque threshold value is adjustable by the adjusting nut; When the rotation of the rotating component is obstructed by the blockage, relative slippage occurs between the friction plate and the transmission sleeve; after the blockage is cleared, the spring resets the friction plate and the transmission sleeve.

[0013] Furthermore, in accordance with any or a combination of the aforementioned technical solutions, the rotating component is disposed at the feed inlet of the conveyor belt chute, and the free end of the rotating component away from the rotating shaft extends into the interior of the chute; And / or, the shaft is also equipped with an oil seal structure; And / or, the detection system further includes a protective plate mounted above the rotating component and the rotating shaft.

[0014] Furthermore, based on any or a combination of the aforementioned technical solutions, the detection system provided by the present invention also includes an alarm device; If the AI ​​agent detects that the chute has changed from a non-blocked state to a blocked state, it will trigger the alarm device to issue an alarm signal. If the AI ​​agent detects that the chute has changed from a blocked state to a non-blocked state, it controls the alarm device to deactivate the alarm signal.

[0015] According to another aspect of the present invention, a method for detecting material blockage in a conveyor belt chute based on an AI agent is provided, comprising the following steps: A rotating component installed in the chute is driven by a motor, and the rotation speed of the rotating component is detected in real time by a rotation sensor. Non-transient time-series rotational speed data is extracted in real time from the detection signal of the rotation sensor according to a preset time window. The captured time-series rotational speed data is input into the AI ​​agent of the integrated time-series deep learning model that has been trained; The AI ​​agent analyzes the time-series rotation speed data to identify the blockage status of the chute; The learning samples used for the pre-training of the AI ​​agent are obtained through one or more of the following methods: Method 1: If, after a preset time t1 from the point t0 when the rotating component stops rotating, and without any intervention, the chute automatically unloads material, causing the rotating component to resume rotation within a preset recovery time t2, then the detection data of the rotation sensor corresponding to the time window is acquired as a learning sample. The starting point of the time window is t0-t2. pre1 , t pre1 This indicates the preset lead time, the end point of the time window is t0+t1+t2, and the label for this learning sample is set as no blockage. Method 2: If, after a preset time t1 from the point t0 when the rotating component stops rotating, the component fails to resume rotation within a preset recovery time t2 without intervention, then the time window [t0-t1] is obtained. pre1 The detection data of the rotary sensor corresponding to [t0+t1+t2] is used as a learning sample, and the learning sample is labeled as "blockage". pre1 Indicates the preset lead time; Method 3: If, after a preset time t3 from the point t0 when the rotating component stops rotating, the component resumes rotation within a preset recovery time t4, and then stops rotating again during a preset time t5 after resuming rotation, then the time window [t0-t] is obtained. pre2 The detection data of the rotary sensor corresponding to [t0+t3+t4+t5] is used as a learning sample, and the label of this learning sample is set as blockage, where t pre2 Indicates the preset lead time; Method 4: Define a cycle in which the rotating component stops rotating at time point t0, and after a preset time t3, the rotating component resumes rotation within a preset recovery time t4. If the rotating component stops rotating again during a preset time t5 after resuming rotation, this constitutes one cycle. If more than two cycles occur within the preset time window, the detection data of the rotation sensor corresponding to that time window is obtained as a learning sample, and the learning sample is labeled as "material blockage".

[0016] Furthermore, following any one or a combination of the aforementioned technical solutions, the detection method provided by this invention further includes the following steps: A magnetic coupler is pre-installed between the motor and the rotating component. When the rotation of the rotating component is obstructed by a blockage, the magnetic coupler disengages the transmission between the motor and the rotating component, thereby allowing the motor to continue running while the rotating component stops rotating. And after the blockage is cleared, the magnetic coupler restores the driving relationship between the motor and the rotating component, thereby enabling the motor to drive the rotating component to rotate again.

[0017] Furthermore, following any one or a combination of the aforementioned technical solutions, the detection method for material blockage in the conveyor chute based on AI intelligent agents is executed based on the detection system described above.

[0018] The beneficial effects of the technical solution provided by this invention are as follows: a. By using an AI agent with an integrated temporal deep learning model to analyze the dynamic changes in the rotational speed of rotating parts within a time window, a decision can be made on whether the chute is blocked or not. This can accurately identify the state of false blockage (actually not blocked) and the state of false non-blockage (actual blockage). b. A magnetic coupler is installed between the motor and the rotating parts. This prevents damage to the motor in the event of a blockage and automatically restores the drive state after switching to a non-blockage state. This provides a reliable hardware condition for the analysis work of the AI ​​agent. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic block diagram of a belt conveyor chute blockage detection system based on an AI agent, provided as an exemplary embodiment of the present invention; Figure 2 for Figure 1 A schematic diagram of the power and detection components of the detection system; Figure 3 for Figure 1 A schematic diagram of the detection scenario of the detection system; Figure 4 A flowchart of a method for detecting material blockage in a belt conveyor chute based on an AI agent, provided as an exemplary embodiment of the present invention; Figure 5 A flowchart illustrating a method for obtaining AI training samples as an exemplary embodiment of the present invention; Figure 6 A flowchart illustrating a specific numerical embodiment of obtaining AI training samples provided by the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0023] In one embodiment of the present invention, a detection system for material blockage in a belt conveyor chute based on an AI agent is provided, such as... Figure 1 As shown, the detection system includes: The power and detection assembly includes a motor 1, a rotating shaft driven by the motor 1, and a rotating component 3 connected to the rotating shaft, the rotating component 3 being configured to be disposed in the conveyor belt chute 4 to be detected; A signal sensing component, comprising a rotation sensor for detecting the rotational speed of the rotating component; Torque overload protection assembly, comprising a magnetic coupler 2 disposed between the motor and the rotating component (e.g., Figure 2 As shown), if the rotation of the rotating component is obstructed by the blockage, causing the rotational torque of the magnetic coupler 2 to exceed a preset mechanical protection torque threshold, then the magnetic coupler 2 disengages the transmission between the motor 1 and the rotating component 3; and The intelligent control and decision-making component includes an AI agent that is communicatively connected to the rotation sensor. The AI ​​agent is configured to analyze and judge the rotation state of the rotating component within a preset time window based on the detection signal of the rotation sensor, so as to identify the material blockage status of the chute (including blockage and non-blockage).

[0024] like Figure 2 As shown, the magnetic coupler 2 includes an active rotor 21 and a driven rotor 22. The active rotor 21 is fixedly connected to the output shaft of the motor 1, and the driven rotor 22 is fixedly connected to the rotating shaft (covered by the driven rotor and not shown) that connects to the rotating component 3. The active rotor 21 and the driven rotor 22 are adjacent to each other but do not make physical contact. The active rotor 21 generates a magnetic field as it rotates with the output shaft of the motor 1, thereby driving the driven rotor 22 to rotate with the rotating shaft and the rotating component 3.

[0025] In one embodiment of the present invention, the mechanical protection torque threshold corresponding to the magnetic coupler 2 is adjustable: a transmission sleeve is fixedly mounted on the rotating shaft, and a friction plate is fixedly mounted on the rotating component. The friction plate is pressed against the transmission sleeve by a spring. When the rotation of the rotating component is hindered by blockage material, relative slippage occurs between the friction plate and the transmission sleeve. After the blockage material is cleared, the spring resets the friction plate and the transmission sleeve. The spring is preloaded by an adjusting nut; the greater the preload of the adjusting nut, the greater the mechanical protection torque threshold, and vice versa.

[0026] like Figure 3 As shown, the rotating component 3 is located at the inlet of the conveyor belt chute 4, and the free end of the rotating component 3 away from the rotating shaft extends into the interior of the chute 4. The longer the extension length, the earlier the warning effect. In one embodiment of the invention, the rotating shaft is also equipped with an oil seal structure to prevent dust from seeping in along the shaft. The detection system also includes a protective plate installed above the rotating component and the rotating shaft to prevent materials from hitting the shaft during use. The protective plate is preferably made of 304 stainless steel. When performing multiple measurements, the distance between the power and detection components should be appropriately increased to avoid the flexible shafts from becoming entangled. The invention does not limit the type of chute. Figure 3 The shape, size, and orientation shown.

[0027] like Figure 2As shown, the rotating component is a spring, preferably multiple, such as two. One end of each spring is fixedly connected to the rotating shaft, and the free end of each spring is independently deflectable relative to the rotating shaft. In different embodiments, the rotating component can be one or more blades, each blade being connected to the rotating shaft via a spring torsion shaft, such that each blade is independently deflectable relative to the rotating shaft.

[0028] The detection system provided in this embodiment of the invention also includes an alarm device; If the AI ​​agent detects that the chute has changed from a non-blocked state to a blocked state, it will trigger the alarm device to issue an alarm signal. If the AI ​​agent detects that the chute has changed from a blocked state to a non-blocked state, it controls the alarm device to deactivate the alarm signal.

[0029] The core inventive point of this invention, namely the AI ​​intelligent agent, will be described in detail below: Unlike existing technologies that detect the transient rotational speed of rotating parts to determine chute blockage, this invention utilizes an AI agent integrating a temporal deep learning model. This agent learns the correlation between the rotational speed frequency (fluctuations) of rotating parts and the chute blockage status within a preset time window through pre-training. The training samples used include at least those obtained through the following methods: like Figure 5 As shown, if, after a preset time t1 from the point t0 when the rotating component stops rotating, and without any intervention, the chute automatically unloads material, causing the rotating component to resume rotation within a preset recovery time t2, then the detection data of the rotation sensor corresponding to the time window is acquired as a learning sample. The starting point of the time window is t0-t2. pre1 , t pre1 This indicates the preset lead time, the end point of the time window is t0+t1+t2, and the label for this learning sample is set as no blockage. If the rotating component does not resume rotation within the preset recovery time t2, then the same time window [t0-t] is obtained. pre1 The detection data of the rotary sensor corresponding to [t0+t1+t2] is used as a learning sample, and the learning sample is labeled as blockage.

[0030] See also Figure 5 Another process path: If, after a preset time t3 from the point t0 when the rotating component stops rotating, the component resumes rotation within a preset recovery time t4, and then stops rotating again during a preset time t5 after resuming rotation, then the time window [t0-t] is obtained. pre2The detection data of the rotary sensor corresponding to [t0+t3+t4+t5] is used as a learning sample, and the label of this learning sample is set as blockage, where t pre2 Indicates the preset lead time; Alternatively, a preset time t3 is defined from the time point t0 when the rotating component stops rotating, and the rotating component resumes rotation within a preset recovery time t4. If the rotating component stops rotating again during the preset time t5 after resuming rotation, this is considered one cycle. If more than two cycles occur within the preset time window, the detection data of the rotation sensor corresponding to the time window is obtained as a learning sample, and the learning sample is labeled as "blockage".

[0031] Specific numerical implementation examples Figure 6 As shown, if the rotating part continues to rotate, the label corresponding to the learning sample obtained in this case is "not blocked". If the rotating part stops rotating at a certain moment (t0) and remains in a stopped state for 4s (t3), and then resumes rotation in the next 2s (t4), that is, the duration of stopping rotation does not meet the requirement of 6s, then we continue to observe whether the rotating part stops rotating again very quickly after resuming rotation (for example, within 3s, i.e., t5). If so, the current round count is incremented by 1 (the initial value of the round count can be 0). If the time window size is set to 27 seconds, then the system monitors whether there is a next round within this time window until the end of the time window is reached. Then, the latest value of the current round number is counted. If the number N is reached, the detection data of the rotary sensor corresponding to the time window [t0-2s, t0+25s] is obtained as a learning sample, and the learning sample is labeled as blockage. Alternatively, N can be set to 1, and the time window can be reduced to 11 seconds. If the number of cycles is incremented by 1, the detection data of the rotation sensor corresponding to the time window [t0-2s, t0+(4s+2s+3s)] can be directly obtained as a learning sample, and the learning sample can be labeled as "blocked". In this case, although the rotating part can resume rotation after a brief stop, it stops rotating again after a brief resumption of rotation. Especially when this happens more than once within a time window, although it does not meet the rule that the rotating part stops rotating for a long time (e.g., 25 seconds), it is actually a false non-blocked state that requires manual intervention.

[0032] If the number of rounds does not meet the condition of incrementing by 1, or if N rounds are not reached within the time window, the label of the corresponding learning sample is "not blocked".

[0033] If the rotating part remains stationary for no more than 4 seconds, the label of the corresponding learning sample is uncertain. The label of this part of the learning sample needs to be associated with the sample label determined in the subsequent process.

[0034] If the rotation remains stopped for 4 seconds (t3 duration) and fails to resume rotation within the next 2 seconds (t4 duration), meaning the stoppage time reaches 6 seconds, then the timing continues. If the total stoppage duration (starting from t0) reaches 20 seconds, the label of the acquired learning sample cannot be determined as blocked or unblocked. If, after a continuous stoppage of 20 seconds, the label of the learning sample is determined based on whether rotation resumes within the next 5 seconds: if rotation resumes within the next 5 seconds, it indicates that the rotating part stopped for 2 seconds. If 0s represents a false blockage state, the detection data of the rotary sensor corresponding to the time window [t0-2s, t0+(20s+5s)] is acquired as a learning sample, and the learning sample is labeled as non-blockage. Experience shows that if rotation does not resume within the next 5s, the probability of self-recovery and material discharge without manual intervention is extremely low. In this case, the detection data of the rotary sensor corresponding to the time window [t0-2s, t0+(20s+5s)] is acquired as a learning sample, and the learning sample is labeled as blockage.

[0035] In response to the false material blockage and false non-blockage states found in actual production, this invention proposes an AI agent based on a temporal deep learning model (such as the CNN-LSTM model). Using a time window as a unit, the rotation frequency of the rotating body is analyzed within a certain time period, thereby accurately identifying the material blockage state, false material blockage state, non-blockage state, and false non-blockage state.

[0036] The current conventional logic of determining the chute blockage status by detecting the transient rotation speed of rotating parts (e.g., if the speed is zero, it is determined to be blocked; if the speed is not zero, it is determined to be not blocked) is prone to false alarms and cannot accurately identify false blockage status.

[0037] One embodiment of the present invention provides a method for detecting material blockage in a conveyor belt chute based on an AI agent, see [link to relevant documentation]. Figure 4 Includes the following steps: S100: A rotating component installed in the chute is driven to rotate by a motor, and the rotation speed of the rotating component is detected in real time by a rotation sensor; S200: Extract non-transient time-series rotational speed data from the detection signal of the rotation sensor in real time according to a preset time window; S300: Inputs the captured time-series speed data into the AI ​​agent that has completed training of the integrated time-series deep learning model; S400: The AI ​​agent analyzes the time-series rotation speed data to identify the blockage status of the chute; The learning samples used for the pre-training of the AI ​​agent are obtained through one or more of the following methods: Method 1: If, after a preset time t1 from the point t0 when the rotating component stops rotating, and without any intervention, the chute automatically unloads material, causing the rotating component to resume rotation within a preset recovery time t2, then the detection data of the rotation sensor corresponding to the time window is acquired as a learning sample. The starting point of the time window is t0-t2. pre1 , t pre1 This indicates the preset lead time, the end point of the time window is t0+t1+t2, and the label for this learning sample is set as no blockage. Method 2: If, after a preset time t1 from the point t0 when the rotating component stops rotating, the component fails to resume rotation within a preset recovery time t2 without intervention, then the time window [t0-t1] is obtained. pre1 The detection data of the rotary sensor corresponding to [t0+t1+t2] is used as a learning sample, and the learning sample is labeled as "blockage". pre1 Indicates the preset lead time; Method 3: If, after a preset time t3 from the point t0 when the rotating component stops rotating, the component resumes rotation within a preset recovery time t4, and then stops rotating again during a preset time t5 after resuming rotation, then the time window [t0-t] is obtained. pre2 The detection data of the rotary sensor corresponding to [t0+t3+t4+t5] is used as a learning sample, and the label of this learning sample is set as blockage, where t pre2 Indicates the preset lead time; Method 4: Define a cycle in which the rotating component stops rotating at time point t0, and after a preset time t3, the rotating component resumes rotation within a preset recovery time t4. If the rotating component stops rotating again during a preset time t5 after resuming rotation, this constitutes one cycle. If more than two cycles occur within the preset time window, the detection data of the rotation sensor corresponding to that time window is obtained as a learning sample, and the learning sample is labeled as "material blockage".

[0038] Furthermore, following any one or a combination of the aforementioned technical solutions, the detection method provided by this invention further includes the following steps: S000: A magnetic coupler is pre-installed between the motor and the rotating component, so that when the rotation of the rotating component is obstructed by the blockage, the magnetic coupler disengages the transmission between the motor and the rotating component, thereby allowing the motor to continue running while the rotating component stops rotating; And so that after the blockage is removed, the magnetic coupler restores the driving relationship between the motor and the rotating component, thereby enabling the motor to drive the rotating component to rotate again.

[0039] A magnetic coupler is used to connect the blades to the motor, driving the blades to rotate. When the flow channel being monitored becomes blocked, causing the material level to rise and hindering the blade rotation, the AI ​​agent detects the abnormal blade rotation through a high-precision rotation sensor. The system then sends a set of signals to indicate the blockage. During blockage, the blades are stuck and do not move, but the motor rotates normally. The magnetic coupler between them protects the motor from damage. When the obstruction is removed, the motor can drive the blades to rotate normally. This avoids the inaccuracies, malfunctions, or premature damage to the detection mechanism itself that occur with the mechanical methods currently used in practice for detecting blockages.

[0040] It should be noted that the detection method for material blockage in a conveyor chute based on an AI agent provided in this embodiment and the detection system for material blockage in a conveyor chute based on an AI agent provided in the above embodiment belong to the same inventive concept. Here, all the contents of the above embodiment of the detection system for material blockage in a conveyor chute based on an AI agent are incorporated into this embodiment of the detection method for material blockage in a conveyor chute based on an AI agent by way of reference.

[0041] The AI-based intelligent agent-based conveyor chute blockage detection system provided in this invention is suitable for use in mines, coal mines, and other industrial plants. Utilizing a high-precision rotary sensor detection method combined with AI computing power, the AI ​​intelligent agent analyzes the detection signals from the rotary sensor to predict the blockage status of the chute in advance. On the one hand, before the blades are jammed by material, an early warning is issued to the staff to promptly clear the material from the chute, preventing blockage. On the other hand, if the blades are jammed due to insufficient material clearing, a magnetic coupler disengages the motor from the blade transmission to protect the motor from damage and outputs a blockage alarm signal to remind the staff to clean the chute as soon as possible, improving the production efficiency of the industrial plant.

[0042] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0043] The above description is only a specific embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A detection system for material blockage in a conveyor chute based on an AI intelligent agent, characterized in that, include: A power and detection assembly includes a motor, a shaft driven by the motor, and a rotating component connected to the shaft, the rotating component being configured to be disposed in a conveyor belt chute to be detected; A signal sensing component, comprising a rotation sensor for detecting the rotational speed of the rotating component; A torque overload protection component includes a magnetic coupler disposed between the motor and the rotating component. If the rotation of the rotating component is obstructed by a blockage, causing the rotational torque of the magnetic coupler to exceed a preset mechanical protection torque threshold, the magnetic coupler will disengage the transmission between the motor and the rotating component. as well as The intelligent control and decision-making component includes an AI agent that is communicatively connected to the rotation sensor. The AI ​​agent is configured to analyze and judge the rotation state of the rotating component within a preset time window based on the detection signal of the rotation sensor, so as to identify the blockage status of the chute.

2. The detection system according to claim 1, characterized in that, The AI ​​agent integrates a temporal deep learning model, and the learning samples used for its training include at least those obtained through the following methods: If, after a preset time t1 from the point t0 when the rotating component stops rotating, and without any intervention, the chute automatically unloads material, causing the rotating component to resume rotation within a preset recovery time t2, then the detection data of the rotation sensor corresponding to the time window is acquired as a learning sample. The starting point of the time window is t0-t2. pre1 , t pre1 This indicates the preset lead time, the end point of the time window is t0+t1+t2, and the label for this learning sample is set as no blockage. If the rotating component does not resume rotation within the preset recovery time t2, then the same time window [t0-t] is obtained. pre1 The detection data of the rotary sensor corresponding to [t0+t1+t2] is used as a learning sample, and the learning sample is labeled as blockage.

3. The detection system according to claim 1, characterized in that, The AI ​​agent integrates a temporal deep learning model, and the learning samples used for its training include at least those obtained through the following methods: If, after a preset time t3 from the point t0 when the rotating component stops rotating, the component resumes rotation within a preset recovery time t4, and then stops rotating again during a preset time t5 after resuming rotation, then the time window [t0-t] is obtained. pre2 The detection data of the rotary sensor corresponding to [t0+t3+t4+t5] is used as a learning sample, and the label of this learning sample is set as blockage, where t pre2 Indicates the preset lead time; Alternatively, a preset time t3 is defined from the time point t0 when the rotating component stops rotating, and the rotating component resumes rotation within a preset recovery time t4. If the rotating component stops rotating again during the preset time t5 after resuming rotation, this is considered one cycle. If more than two cycles occur within the preset time window, the detection data of the rotation sensor corresponding to the time window is obtained as a learning sample, and the learning sample is labeled as "blockage".

4. The detection system according to claim 1, characterized in that, The number of rotating components is multiple, and they have any of the following forms: The rotating component is a blade, and each blade is connected to the rotating shaft via a spring torsion shaft, so that each blade can be independently deflected relative to the rotating shaft; Alternatively, the rotating component is a spring, with one end of each spring fixedly connected to the rotating shaft, and the free end of each spring independently deflectable relative to the rotating shaft.

5. The detection system according to claim 1, characterized in that, The magnetic coupler includes an active rotor and a driven rotor, wherein the active rotor is fixedly connected to the output shaft of the motor, and the driven rotor is fixedly connected to the rotating shaft of the connecting rotating component; The active rotor and the driven rotor do not make physical contact. The active rotor generates a magnetic field as it rotates with the motor output shaft, thereby driving the driven rotor to rotate with the shaft and rotating components.

6. The detection system according to claim 1, characterized in that, A transmission sleeve is fixedly mounted on the rotating shaft, and a friction plate is fixedly mounted on the rotating component. The friction plate is pressed onto the transmission sleeve by a spring, wherein the spring is subject to the preload of an adjusting nut. The mechanical protection torque threshold value is adjustable by the adjusting nut; When the rotation of the rotating component is obstructed by the blockage, relative slippage occurs between the friction plate and the transmission sleeve; after the blockage is cleared, the spring resets the friction plate and the transmission sleeve.

7. The detection system according to claim 1, characterized in that, The rotating component is located at the feed inlet of the conveyor belt chute, and the free end of the rotating component away from the rotating shaft extends into the interior of the chute. And / or, the shaft is also equipped with an oil seal structure; And / or, the detection system further includes a protective plate mounted above the rotating component and the rotating shaft.

8. The detection system according to any one of claims 1 to 7, characterized in that, It also includes alarm devices; If the AI ​​agent detects that the chute has changed from a non-blocked state to a blocked state, it will trigger the alarm device to issue an alarm signal. If the AI ​​agent detects that the chute has changed from a blocked state to a non-blocked state, it controls the alarm device to deactivate the alarm signal.

9. A method for detecting material blockage in a belt conveyor chute based on an AI intelligent agent, characterized in that, Includes the following steps: A rotating component installed in the chute is driven by a motor, and the rotation speed of the rotating component is detected in real time by a rotation sensor. Non-transient time-series rotational speed data is extracted in real time from the detection signal of the rotation sensor according to a preset time window. The captured time-series rotational speed data is input into the AI ​​agent of the integrated time-series deep learning model that has been trained; The AI ​​agent analyzes the time-series rotation speed data to identify the blockage status of the chute; The learning samples used for the pre-training of the AI ​​agent are obtained through one or more of the following methods: Method 1: If, after a preset time t1 from the point t0 when the rotating component stops rotating, and without any intervention, the chute automatically unloads material, causing the rotating component to resume rotation within a preset recovery time t2, then the detection data of the rotation sensor corresponding to the time window is acquired as a learning sample. The starting point of the time window is t0-t2. pre1 , t pre1 This indicates the preset lead time, the end point of the time window is t0+t1+t2, and the label for this learning sample is set as no blockage. Method 2: If, after a preset time t1 from the point t0 when the rotating component stops rotating, the component fails to resume rotation within a preset recovery time t2 without intervention, then the time window [t0-t1] is obtained. pre1 The detection data of the rotary sensor corresponding to [t0+t1+t2] is used as a learning sample, and the learning sample is labeled as "blockage". pre1 Indicates the preset lead time; Method 3: If, after a preset time t3 from the point t0 when the rotating component stops rotating, the component resumes rotation within a preset recovery time t4, and then stops rotating again during a preset time t5 after resuming rotation, then the time window [t0-t] is obtained. pre2 The detection data of the rotary sensor corresponding to [t0+t3+t4+t5] is used as a learning sample, and the label of this learning sample is set as blockage, where t pre2 Indicates the preset lead time; Method 4: Define a cycle in which the rotating component stops rotating at time point t0, and after a preset time t3, the rotating component resumes rotation within a preset recovery time t4. If the rotating component stops rotating again during a preset time t5 after resuming rotation, this constitutes one cycle. If more than two cycles occur within the preset time window, the detection data of the rotation sensor corresponding to that time window is obtained as a learning sample, and the learning sample is labeled as "material blockage".

10. The method for detecting material blockage in a belt conveyor chute based on an AI intelligent agent according to claim 9, characterized in that, It also includes the following steps: A magnetic coupler is pre-installed between the motor and the rotating component. When the rotation of the rotating component is obstructed by a blockage, the magnetic coupler disengages the transmission between the motor and the rotating component, thereby allowing the motor to continue running while the rotating component stops rotating. And after the blockage is cleared, the magnetic coupler restores the driving relationship between the motor and the rotating component, thereby enabling the motor to drive the rotating component to rotate again.

11. The method for detecting material blockage in a belt conveyor chute based on an AI intelligent agent according to claim 9, characterized in that, The detection method is performed based on the detection system as described in any one of claims 1 to 8.

Images