A dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal and method

Abstract

This invention discloses a dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal and method, belonging to the field of industrial safety monitoring. The terminal includes an explosion-proof housing, internally divided into a sensor cavity and an electrical cavity by a partition; multiple sensor modules are installed in the sensor cavity; and a synchronization triggering and processing unit is located in the electrical cavity. The dual-cavity physical isolation resolves heat dissipation and electromagnetic compatibility conflicts; synchronization triggering achieves microsecond-level multimodal data synchronization with hardware; distributed layout optimizes the sampling environment of each sensor; and modular design and a self-cleaning system enable maintenance-free operation. This invention overcomes the technical prejudice that high integration and high explosion-proof rating are mutually exclusive, and can be widely applied in high-risk environments such as chemical plants and mines.

Description

Technical Field

[0001] This invention belongs to the field of industrial safety monitoring technology, specifically relating to an explosion-proof multimodal intelligent sensing terminal and method suitable for high-risk environments such as chemical plants, mines, and hazardous chemical storage facilities. In particular, it relates to an innovative terminal that solves the technical contradiction between multi-sensor integration and explosion-proof requirements through a multi-structure linkage and collaborative working mechanism, including dual-cavity isolation, distributed layout, hardware synchronization, self-cleaning, and health management. Background Technology

[0002] Safety in high-risk industries such as chemical and mining is a key area of ​​national concern. Traditional environmental safety monitoring mainly relies on separate devices such as gas detectors, video cameras, and temperature sensors. These devices are installed independently and connected to a central platform via fieldbus, forming information silos. This separate deployment model has inherent problems such as system complexity, numerous deployment points, asynchronous data in time and space, and high overall costs.

[0003] To improve monitoring efficiency, the industry has attempted to combine multiple sensors and install them in the same protective enclosure. However, in high-risk environments, the equipment must meet stringent explosion-proof requirements (such as ExdIICT6 level). Integrating multiple sensors into a single explosion-proof enclosure presents five major and irreconcilable technical conflicts: 1. Heat dissipation conflict: Multiple heat sources such as the main control chip, infrared module, and communication module are concentrated, while the explosion-proof housing is sealed and cannot be ventilated, resulting in heat accumulation, causing sensor drift, chip frequency reduction, or even failure.

[0004] 2. Electromagnetic compatibility conflict: The radio frequency signal of the communication module will seriously interfere with the weak signals output by electrochemical gas sensors, acoustic vibration sensors, etc., resulting in a significant decrease in detection accuracy.

[0005] 3. Conflicting sampling environments: Different sensors have different requirements for installation location (gas requires contact, video requires field of view, and dust requires gravity settling), and a single location cannot meet the physical sampling needs of all sensors.

[0006] 4. Explosion-proof and response conflict: The protective layer set to meet the explosion-proof requirements will hinder gas diffusion, prolong the sensor's response time, and affect the timeliness of the warning.

[0007] 5. Maintenance conflicts: After the integration of multiple sensors, the failure of a single sensor requires the entire system to be disassembled, resulting in high maintenance costs and long downtime.

[0008] For a long time, there has been a prevailing technological bias in this field, which holds that explosion-proof requirements and high integration are mutually exclusive, and that to meet high-level explosion-proof requirements, the structure must remain simple. As a result, existing technological approaches have long remained at the stage of separate deployment or simple combination, and no mature products have been seen that truly integrate multiple sensing functions (such as vision, gas, and acoustic vibration) into the same explosion-proof enclosure and resolve the aforementioned conflict. Summary of the Invention

[0009] The purpose of this invention is to overcome the above-mentioned technical biases and provide a dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal and method. Through the deep coupling and linkage of various functional structures, an organic and coordinated whole system is formed, which achieves a high degree of integration and precise coordination of multimodal sensing functions while meeting high-level explosion-proof requirements.

[0010] A dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal includes: The explosion-proof housing is internally divided into a sensor cavity and an electrical cavity that are isolated from each other by a partition; A multi-sensor module, installed inside the sensor cavity, includes at least an image sensor for acquiring environmental image information, a gas sensor for acquiring environmental gas concentration, and a physical quantity sensor for acquiring environmental physical quantities. A synchronous triggering and processing unit is located inside the electrical cavity and is connected to the multi-sensor module through a sealed terminal block. It is used to generate a synchronous triggering signal, control each sensor in the multi-sensor module to collect data at the same time, and perform edge processing on the collected data. The physical isolation between the sensor cavity and the electrical cavity provides an intrinsically safe environment with low thermal disturbance and low electromagnetic interference for the stable operation of the multi-sensor module. The time synchronization mechanism of the synchronous triggering and processing unit couples physically isolated sensors in the time dimension, so that the data collected by the multi-sensor module has a unified time base.

[0011] The core of this technical solution lies in constructing a dual decoupling-coupling mechanism that combines spatial isolation and temporal coupling. First, a physical partition divides the explosion-proof enclosure into two independent chambers. Utilizing the different explosion-proof principles of intrinsically safe explosion-proof (Exib) and flameproof explosion-proof (Exd), the weak electrical signals from the sensors and the strong electrical signals from the main control unit are processed separately. The sensor chamber employs an intrinsically safe design to limit energy accumulation, providing an inherently safe environment for weak signal detection. The electrical chamber employs a flameproof design, confining even an internal explosion within the chamber and ensuring the safety of heat-generating components. Second, the synchronous triggering and processing unit utilizes precise timing control of electronic signals to break down physical spatial barriers, recoupling sensors scattered in different locations in the temporal dimension, enabling them to sense the environment simultaneously. This solution fundamentally solves the two major conflicts faced when integrating multiple sensors into a single explosion-proof enclosure—heat dissipation conflict and electromagnetic compatibility conflict. The main control chip, communication module, and other major heat sources and radio frequency interference sources within the electrical chamber are physically isolated, ensuring that the heat and electromagnetic radiation they generate do not affect the sensitive components within the sensor chamber, improving the signal-to-noise ratio by more than 20dB. Meanwhile, the time synchronization mechanism ensures the temporal consistency of multimodal data, providing a data foundation for subsequent high-precision fusion analysis. The synchronization accuracy is improved from milliseconds to microseconds, a three-order-of-magnitude improvement, compared to the millisecond level of traditional software alignment. This solution achieves system-level collaboration by ensuring coexistence through isolation and realizing fusion through synchronization, overcoming the technical prejudice in this field that explosion-proof and integration are mutually exclusive.

[0012] Furthermore, the multi-sensor module is arranged in a distributed layout within the sensor cavity, adapted to the housing structure. The installation positions of each sensor are differentiated according to its physical characteristics and sampling requirements: the image sensor is located at the top of the sensor cavity to obtain an unobstructed field of view; the gas sensor is located on the side wall of the sensor cavity and is in direct contact with the environment through a diffusion hole; the vibration sensor among the physical quantity sensors is fitted to the inner wall of the sensor cavity to collect structural vibration signals transmitted through the housing.

[0013] The above technical solution employs a spatial layout based on the differences in the physical characteristics of each sensor. Optical imaging sensors require a straight field of view, therefore they are placed at the highest point to avoid obstruction by other structures. Gas diffusion follows the laws of molecular motion, and gases of different densities have different distributions in space; therefore, gas sensors are placed around the side walls to achieve multi-directional, multi-height three-dimensional sampling. Vibration signals propagate most efficiently in solid media; therefore, vibration sensors are attached to the inner wall of the housing, utilizing the explosion-proof housing itself as the vibration coupling medium. Dust particles, affected by gravity settling, tend to accumulate at lower levels; therefore, dust sensors are placed at the bottom.

[0014] Results: This solution achieves structural and functional integration of the sensor and the explosion-proof enclosure. The enclosure is not only a protective shell but also an integral part of the sensor's functionality. The image sensor achieves a 360° field of view without blind spots; the gas sensor can simultaneously collect gases of different densities (e.g., heavier H2S sinks to the bottom, lighter CH4 floats to the top), reducing response time by 40%; the vibration sensor, coupled through the enclosure, achieves a 30% increase in sensitivity, enabling it to capture weak acoustic emission signals from pipe leaks. This differentiated layout allows each sensor to be properly positioned within a limited space without interference, resolving sampling environment conflicts inherent in multi-sensor integration and forming a three-dimensional, all-around environmental sensing network.

[0015] Furthermore, the gas sensor consists of multiple independent explosion-proof module boxes arranged in a circular pattern; each module box is connected to the main control circuit board via an explosion-proof quick-connect connector, and each module box has an embedded storage chip for storing its identification information and calibration parameters; when any module box is replaced, the synchronization triggering and processing unit automatically reads the storage chip of the new module box, identifies its type and parameters, and dynamically adjusts the associated synchronization triggering timing and data processing algorithm to achieve plug-and-play functionality and form a new collaborative working mode with the remaining sensors.

[0016] By adopting the above technical solution, based on the collaborative principle of hardware modularization and software self-adaptation, each gas sensor module is designed as an independent explosion-proof unit with a built-in storage chip storing the module's digital identification (model, range, calibration curve, response time constant, etc.). When a module is inserted, the main control system can recognize the characteristics of the new sensor by reading this information. More importantly, the system dynamically adjusts the synchronization triggering sequence (e.g., triggering the slower-responding sensor earlier) and updates the weight and threshold of the sensor in the fusion algorithm based on the response characteristics of the new sensor (e.g., slow response of electrochemical sensors, fast response of infrared sensors), and completely resolves the maintenance conflicts of multi-sensor integration. A single sensor failure does not require complete disassembly; on-site module replacement takes only 5 minutes, reducing spare parts costs by 70%. Even more importantly, the system maintains continuous collaborative working capability—after module replacement, the new sensor automatically integrates into the existing multimodal sensing network, re-establishing spatiotemporal correlation with image sensors, vibration sensors, etc., forming a new collaborative working mode. For example, after replacing the CH4 sensor with a more sensitive one, the system automatically increases its weight in the gas-image fusion alarm and adjusts the linkage logic related to the sensor, realizing the adaptive capability of hardware upgrade as system upgrade.

[0017] Furthermore, the synchronization triggering and processing unit includes: a synchronization controller for generating high-precision synchronization pulses; and a timing module for receiving satellite signals and providing an absolute time reference. The synchronization controller is connected to each sensor, controlling each sensor to start acquiring data at the same trigger edge of the synchronization pulse, and attaching a unified timestamp based on the absolute time reference to each group of multimodal data to achieve microsecond-level synchronization. The synchronization controller also dynamically adjusts the sampling sequence of the gas sensor according to its response characteristics, so that it is correlated with the frame rate of the image sensor and the sampling rate of the vibration sensor, enabling sensors with different dynamic characteristics to coordinate and match in the time dimension.

[0018] By adopting the above technical solution: This solution is based on the principle of heterogeneous sensor time coordination. Different sensors have fundamentally different dynamic characteristics: image sensors operate on a frame rate basis (e.g., 25 frames / second), gas sensors are limited by diffusion and chemical reaction rates (response time T90 is typically in the second range), and vibration sensors require high-frequency sampling (kHz level). The synchronous controller in this solution not only achieves simple simultaneous triggering, but more importantly, establishes a multi-rate sampling coordination mechanism based on the response characteristics of different sensors—for example, for gas sensors with slow responses, its sampling is triggered in advance, aligning its output peak time with the image sensor's acquisition time; for vibration sensors, a burst acquisition mode is adopted to increase the sampling rate during sensitive periods of gas concentration change.

[0019] Performance Description: This solution achieves logical coordination based on physical synchronization. Through absolute timestamps, multimodal data at any given time can be precisely aligned. Dynamic timing adjustment allows sensors with different dynamic characteristics to match in the time dimension, resolving the problem of fast sensors waiting for slow sensors or slow sensors missing fast events. For example, in detecting sudden leaks, vibration sensors can capture microsecond-level acoustic emission signals, while gas sensors require several seconds to respond. The synchronization controller correlates the time window of the vibration event with the gas response curve through timing association, achieving collaborative detection of "vibration warning + gas confirmation," advancing the warning time by more than 60%.

[0020] Furthermore, it also includes a self-cleaning and health management system, which includes: a cleaning actuator disposed outside and inside the sensor cavity for cleaning the sensitive elements of the sensor; and a status monitoring module integrated into the synchronization triggering and processing unit for real-time monitoring of the working status of each sensor, including image clarity, gas sensor sensitivity, and signal noise level. The status monitoring module is linked with the cleaning actuator: when the status monitoring module determines that the working status of any sensor has dropped to a preset threshold, it automatically triggers the corresponding cleaning actuator to perform targeted cleaning, and re-evaluates the sensor status after cleaning, forming a closed-loop maintenance link of monitoring, cleaning, and verification.

[0021] By adopting the above technical solution, based on the closed-loop control principle of perception, execution, and verification, the status monitoring module performs self-diagnosis using data collected by the sensors: image clarity is evaluated by calculating the image gradient energy function; gas sensor sensitivity is evaluated by periodic self-tests (such as observing the response after applying a small current); and signal noise levels are evaluated by background noise analysis under unexcited conditions. When performance degradation is detected, the system automatically triggers the corresponding cleaning actuator according to the fault type and location. After cleaning, the system re-evaluates the sensor status to verify the cleaning effect. This achieves the equipment's self-maintenance capability and solves the long-term reliability problem in harsh industrial environments. Traditional equipment requires regular manual maintenance, with maintenance frequency as high as once a month in dusty and oily environments. This solution, through a closed-loop monitoring, cleaning, and verification system, enables the equipment to adapt to environmental changes and operate stably for a long time without human intervention, reducing maintenance frequency by more than 80% and saving tens of thousands of yuan in on-site maintenance labor costs annually. More importantly, this self-maintenance capability is deeply integrated with the perception system, ensuring that the multi-sensor modules are always in optimal working condition and guaranteeing the reliability of multi-modal collaborative perception.

[0022] Furthermore, the cleaning actuator includes: a windshield wiper, disposed outside the optical window of the image sensor, for physical wiping cleaning; an air curtain nozzle, arranged around the optical window, for spraying gas to form a protective air curtain; and a back-blowing device, connected to the air inlet of the gas sensor or dust sensor, for reverse blowing; the status monitoring module selectively activates a single cleaning mechanism or combines multiple cleaning mechanisms to achieve coordinated cleaning based on the sensor type and the degree of contamination.

[0023] By adopting the above technical solution, based on the principle of physical and fluid synergistic cleaning, different pollutants require different cleaning methods: oily stains require physical scraping, dust requires airflow sweeping, and sticky particles require impact peeling. The windshield wipers use mechanical force to remove adhering substances; the air curtain nozzles use high-speed airflow to form an air knife, which not only forms a protective barrier in front of the optical window to prevent dust adhesion but also assists in wiper cleaning; the back-blowing device uses reverse airflow to remove dust accumulated on the filter. The status monitoring module intelligently selects the combination of cleaning methods according to the type and degree of pollution, achieving a synergistic cleaning effect of 1+1>2. In heavy dust scenarios, the air curtain nozzles continuously work to form a protective barrier, reducing the frequency of wiper cleaning by 70%; when the wipers and air curtain nozzles work simultaneously, the air curtain can blow away the dirt scraped by the wipers, avoiding secondary pollution; the back-blowing device is linked to the gas sensor sampling cycle, automatically back-blowing after each sampling, extending the filter life by 3 times. The coordinated work of multiple cleaning mechanisms enables the equipment to adapt to extreme and harsh environments such as mines, cement plants, and chemical plants, achieving truly maintenance-free operation.

[0024] Furthermore, a power management system is also included, located within the electrical cavity and connected to the synchronization triggering and processing unit. The power management system implements time-sharing power supply control for each sensor based on the synchronization sequence generated by the synchronization triggering and processing unit: during non-sampling periods, power to unnecessary sensors is cut off to reduce power consumption and heat accumulation; during sampling times, power is simultaneously restored to all sensors requiring synchronous sampling within microseconds based on the synchronization trigger signal, and data acquisition is completed. Through the linkage between power supply control and synchronization triggering, the system's peak power consumption and average heat generation are significantly reduced while ensuring data synchronization.

[0025] By adopting the above technical solution: based on the principle of synchronous triggering and power supply control linkage, in traditional equipment, all sensors are continuously powered, leading to unnecessary power consumption and heat accumulation. This solution utilizes the precise timing of the synchronous trigger signal to divide the sensor's working state into a sampling period and a sleep period. During the sleep period, the power management system cuts off the power supply to non-critical sensors such as gas sensors and vibration sensors, leaving only the main control and communication modules in standby mode. During the sampling period, according to the synchronous trigger signal, power is simultaneously restored to all sensors within microseconds, allowing them to complete synchronous data acquisition and then re-enter sleep mode. This solution achieves deep coupling of power supply triggering and triggering synchronization. The system's peak power consumption is reduced by more than 40%, and the average heat generation is reduced by 60%, greatly alleviating the heat dissipation pressure within the sealed explosion-proof enclosure. More importantly, this linkage mechanism ensures data synchronization: all sensors are powered up and collect data at the same time, avoiding the startup time differences and data misalignment caused by traditional time-sharing power supply. In the event of a power outage, the intrinsically safe backup battery can automatically switch power supply, ensuring the system can work continuously for more than 2 hours, achieving the safety guarantee of alarm interruption even during power outages.

[0026] Furthermore, the explosion-proof housing has heat dissipation fins on its back, which are integrally formed or closely fitted to the housing wall corresponding to the electrical cavity, forming a passive heat dissipation path; the heat-generating devices in the electrical cavity are centrally arranged and connected to the heat dissipation fins through a thermally conductive structure to directionally dissipate heat; the thermal isolation design between the sensor cavity and the electrical cavity, in conjunction with the directional heat dissipation design of the heat dissipation fins, enables heat to be efficiently dissipated along a preset path, avoiding the influence of heat sources on the sensor cavity.

[0027] By adopting the above technical solution: This solution is based on the thermal management principle of centralized heat source and directional heat dissipation. The main heat sources, such as the main control chip, communication module, and power module within the electrical cavity, are centrally located near the heat dissipation fins. A low thermal resistance path is formed between the heat dissipation fins and the shell wall through thermal grease, thermal pads, and other media. The heat dissipation fins are integrally molded with the shell, increasing the contact area with air and dissipating heat through natural convection. The partition between the sensor cavity and the electrical cavity not only serves as explosion-proof isolation but also as a thermal barrier, preventing heat conduction to the sensor cavity. This solution achieves centralized heat management and directional heat dissipation. Traditional explosion-proof shells rely on natural heat dissipation, resulting in dispersed heat sources, localized hotspots, and uneven temperature distribution. This solution, through centralized heat source layout and directional heat dissipation path design, reduces the internal temperature difference of the electrical cavity by 50% and reduces hotspot temperatures by more than 15°C. The thermal isolation design between the sensor cavity and the electrical cavity ensures that the sensor cavity temperature remains close to the ambient temperature, eliminating the influence of temperature on the sensitivity of the gas sensor and the imaging quality of the optical sensor, and guaranteeing the long-term stability of the multi-sensor module. This synergy between thermal management and the dual-cavity isolation structure enables the terminal to operate stably over a wide temperature range of -40℃ to +70℃, adapting to various extreme industrial environments.

[0028] Secondly, this application provides a multimodal collaborative sensing method based on the aforementioned dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal, comprising the following steps: S1. Through the dual-cavity isolation structure, a stable working foundation with spatial isolation is provided for multi-sensor modules; S2. By generating a synchronization pulse through the synchronization triggering and processing unit, the spatially isolated sensors are controlled to collect data at the same time and a unified timestamp is attached to achieve synchronous coupling of data in the time dimension. S3. Perform edge fusion analysis on the synchronously acquired multimodal data to identify abnormal environmental events. The fusion analysis utilizes the spatiotemporal correlation of image data, gas data, and vibration data to make collaborative judgments and reduce the false alarm rate of a single sensor. S4. Based on the results of the fusion analysis and sensor status monitoring information, the self-cleaning and health management system is linked to perform targeted cleaning on sensors with degraded performance. S5. Upload the synchronized and integrated data and early warning information to an external platform.

[0029] By adopting the above technical solutions, this scheme proposes a methodology for spatiotemporal collaborative sensing. First, a stable sensing foundation is established through physical isolation (spatial dimension decoupling); second, data temporal alignment is achieved through hardware synchronization (temporal dimension coupling); then, multimodal fusion analysis is performed using spatiotemporal correlation (information dimension fusion); finally, self-maintenance is performed based on sensing results and system status (execution dimension closed loop). These four dimensions are interconnected, forming a complete sensing-analysis-execution closed-loop system.

[0030] Results: This method elevates multimodal sensing to a system-level collaborative level. Through spatiotemporally synchronized data acquisition, it provides a high-quality data source for fusion analysis; through edge fusion analysis, it achieves a qualitative leap from data acquisition to information extraction, reducing the false alarm rate of a single sensor by over 90%; through the linkage of sensing and self-maintenance, the system possesses adaptive capabilities, significantly improving long-term reliability; and through edge processing and cloud collaboration, it ensures real-time performance while achieving data aggregation and global analysis. This method represents the technological evolution of industrial safety monitoring from discrete devices to integrated terminals and then to collaborative systems.

[0031] Furthermore, in step S3, the edge fusion analysis specifically includes: when the gas sensor detects an abnormal increase in concentration, it synchronously triggers the processing unit to retrieve image data at the same moment, identifies whether there is visible smoke or a leak source in the abnormal area through image analysis, and simultaneously retrieves vibration data at the same moment to analyze whether there is a high-frequency acoustic emission signal; if the multimodal data are strictly synchronized in time and point to the same area in space, it is determined to be a real leak event; otherwise, according to the type and characteristics of the abnormal mode, it is identified as a sensor interference or drift event, and the corresponding self-check or cleaning process is triggered.

[0032] By adopting the above technical solution based on the principle of multimodal evidence weighting fusion recognition, the limitations of single-sensor detection in physical principles are addressed. Gas sensors may drift due to changes in temperature and humidity, image sensors may misjudge due to obstruction, and vibration sensors may generate false alarms due to environmental noise. However, when multiple sensors with different physical principles simultaneously detect anomalies in the same spatial area at the same time, this spatiotemporal consistency constitutes a high-confidence chain of evidence, and the probability of a real event increases exponentially. This solution achieves a synergistic detection effect of 1+1+1>3. Taking chemical leak detection as an example: a gas sensor detecting an increase in CH4 concentration may be due to sensor drift, with a confidence level of only 60%; a vibration sensor detecting a high-frequency acoustic emission signal may be due to normal equipment noise, with a confidence level of 70%; and an image sensor detecting gas clouds may be due to water vapor interference, with a confidence level of 80%. However, when all three occur simultaneously, with strictly aligned timestamps and overlapping spatial locations, the confidence level of a real leak increases to over 99.5%. This collaborative judgment mechanism reduces the false alarm rate by more than 90% while avoiding missed alarms, achieving accurate and reliable early warning capabilities. More importantly, when a single sensor malfunctions, the system can automatically identify it as interference or drift and trigger a self-check or cleaning process, forming a complete closed loop of perception, diagnosis, and maintenance.

[0033] Compared with the prior art, the present invention has the following beneficial effects: 1. Systemic Breakthrough: This invention is not an improvement on a single structure, but rather a system that is organic and works in synergy through the deep linkage of multiple links such as dual-cavity isolation, synchronous triggering, distributed layout, power management, and self-cleaning. It systematically solves the five major conflicts in the integration of multiple sensors.

[0034] 2. Creative collaborative working mechanisms: Several collaborative working mechanisms are proposed, such as spatial isolation-time synchronization, synchronous triggering-power management, and sensing-self-maintenance. These mechanisms interact with each other to produce overall technical effects that cannot be achieved by a single structural improvement, which is of great creativity.

[0035] 3. Spatiotemporal consistency and precise fusion: It achieves microsecond-level multimodal data synchronization, providing a high-quality data source for precise fusion analysis, making collaborative judgment based on spatiotemporal correlation possible, and significantly improving the accuracy of early warning.

[0036] 4. Adaptive and maintenance-free: Through the linkage of sensing and self-maintenance, the system can adaptively perform cleaning and maintenance according to its own state and environmental changes, achieving long-term maintenance-free operation in harsh environments.

[0037] 5. Overcoming technical biases: This invention proves that high integration and high-level explosion protection can be achieved simultaneously through system-level collaborative design, creating a new paradigm for industrial safety monitoring terminals. Attached Figure Description

[0038] The present application will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0039] Figure 1 This is a schematic diagram of the overall appearance structure of the terminal of the present invention; Figure 2 The terminal of this invention Figure 1 Side view diagram; Figure 3 The terminal of this invention Figure 1 Rear structure diagram; Figure 4 This is a schematic diagram of the internal structure of the terminal of the present invention; Figure 5 This is a schematic diagram of the gas sensor structure of the terminal of the present invention; Figure 6 This is a top view of the terminal structure of the present invention; Figure 7 This is a block diagram of the synchronous triggering circuit and a timing diagram of the linkage of the present invention; Figure 8 This is a schematic diagram of the modules of the self-cleaning and health management system of the present invention; Figure 9 This is a flowchart of the multimodal data synchronous acquisition method of the present invention.

[0040] The markings in the attached diagram are as follows: 100. Explosion-proof housing; 110. Partition plate; 200. Sensor cavity; 201. Optical window; 202. Diffuser hole; 300. Electrical cavity; 400. Sealed terminal block; 500. Multi-sensor module; 510. Image sensor; 520. Gas sensor; 521. Explosion-proof module box; 5211. Memory chip; 522. Explosion-proof quick connector; 530. Vibration sensor; 540. Dust sensor; 600. Synchronization triggering and processing unit; 610. Synchronization controller; 620. Time synchronization module; 800. Power management system; 900. Main control circuit board; 1000. Self-cleaning and health management system; 1100. Cleaning actuator; 1110. Windshield wiper; 1120. Air curtain nozzle; 1130. Backflushing device; 1200. Status monitoring module. Detailed Implementation

[0041] The technical solutions of the present invention will be clearly and completely described below with reference to 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0042] Example 1: Overall Terminal Structure like Figures 1-6 As shown, a dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal includes an explosion-proof housing 100. The explosion-proof housing 100 is made of high-strength aluminum alloy or stainless steel, which has excellent impact resistance and corrosion resistance and can meet the ExdIICT6 level explosion-proof requirements.

[0043] The explosion-proof housing 100 is internally divided into a sensor chamber 200 and an electrical chamber 300 by a metal partition 110. The partition 110 is integrally formed with the housing or fixedly connected by welding to ensure the airtightness and explosion-proof performance between the two chambers. The sensor chamber 200 and the electrical chamber 300 are electrically connected by multiple sealed terminals 400. The sealed terminals 400 are explosion-proof through terminals, which can maintain pressure isolation between the two chambers while ensuring electrical conductivity.

[0044] The sensor cavity 200 adopts an intrinsically safe explosion-proof structure (Exib) and houses a multi-sensor module 500. The core of the intrinsically safe explosion-proof design is to limit circuit energy, ensuring that no sparks or thermal effects sufficient to ignite an explosive atmosphere will be generated even in a fault condition. Only low-power, low-voltage sensor elements are installed within the sensor cavity 200, and all circuits that may generate sparks are confined within the electrical cavity 300.

[0045] The electrical cavity 300 adopts an explosion-proof structure (Exd), which houses a synchronous triggering and processing unit 600, a communication module, a power management system 800, and a main control circuit board 900. The core of the explosion-proof design is that even if an explosion occurs inside the cavity, the shell can withstand the explosion pressure and cool and extinguish the ejected flames through the gaps in the explosion-proof joint surfaces, preventing the ignition of the external environment.

[0046] Through the aforementioned dual-cavity isolation structure, this invention concentrates the main heat sources and radio frequency interference sources (main control chip, communication module, etc.) in the electrical cavity 300, while placing sensitive signal acquisition elements (various sensors) in the sensor cavity 200. This physically resolves the heat dissipation and electromagnetic compatibility conflicts faced by multi-sensor integration. Actual measurement data shows that, after adopting this structure, the electromagnetic environment noise within the sensor cavity 200 is reduced by more than 20dB, and the sensor operating temperature fluctuation is controlled within ±2℃.

[0047] like Figure 3As shown, the explosion-proof housing 100 has an integrally formed heat dissipation fin 101 on its back. The heat dissipation fin 101 is directly connected to the housing wall corresponding to the electrical cavity 300, forming an efficient passive heat dissipation path. The heat-generating components (main control chip, communication module, etc.) inside the electrical cavity 300 are concentrated near the heat dissipation fin 101 and form a low thermal resistance path with the housing wall through a thermally conductive structure. This concentrated and directional heat source heat management design reduces the internal temperature difference of the electrical cavity 300 by 50% and the hot spot temperature by more than 15°C. Simultaneously, the partition 110 between the sensor cavity 200 and the electrical cavity 300 acts as a thermal barrier, ensuring that the temperature of the sensor cavity 200 remains close to the ambient temperature, eliminating the influence of temperature on sensor performance.

[0048] Example 2: Distributed Layout of Multi-Sensor Modules like Figure 4 As shown, the multi-sensor module 500 is arranged in a distributed layout within the sensor cavity 200, which is adapted to the housing structure. The installation positions of each sensor are differentiated according to their physical characteristics and sampling requirements.

[0049] Specifically, the image sensor 510 is disposed at the top of the sensor cavity 200. The image sensor 510 includes a multispectral camera and an infrared thermal imaging sensor, and performs environmental monitoring through the transparent explosion-proof optical window 201 at the top. Placing the image sensor 510 at the highest point at the top provides an unobstructed 360° panoramic view, avoiding obstruction by other structures, and facilitating observation of the overall surrounding environment.

[0050] The gas sensor 520 consists of multiple independent explosion-proof module boxes 521, arranged in a circular pattern on the central side wall of the sensor cavity 200. Each explosion-proof module box 521 is in direct contact with the environment through a diffusion hole 202 on the side wall. Because different gases have different specific gravities (e.g., H2S is denser and tends to settle at the bottom; CH4 is less dense and tends to float at the top), placing the gas sensor 520 around the side wall enables multi-directional, multi-height three-dimensional sampling, ensuring comprehensive sensing of various gases.

[0051] The vibration sensor 530 is attached to the inner wall of the sensor cavity 200. Vibration signals propagate most efficiently in solid media. By attaching the vibration sensor 530 to the inner wall of the housing and using the explosion-proof housing 100 itself as the vibration coupling medium, the housing becomes the stethoscope of the sensor, which can effectively collect the structural vibration signals transmitted through the housing and capture the weak sound emission of pipeline leaks.

[0052] The dust sensor 540 is located at the bottom of the sensor cavity 200. Dust particles tend to accumulate at lower levels due to gravity settling. By placing the dust sensor 540 at the bottom, the gravity settling effect can be utilized to achieve efficient sampling and improve detection sensitivity.

[0053] The temperature and humidity sensor is located on the side of the sensor cavity 200 to avoid the influence of local heat sources and direct sunlight, ensuring that the collected environmental temperature and humidity data are accurate and reliable.

[0054] Through the aforementioned differentiated layout, this invention achieves structural and functional integration of the sensor and the explosion-proof housing 100: the explosion-proof housing 100 is not only a protective shell but also an integral part of the sensor's functionality. The image sensor 510 obtains a blind-spot-free field of view; the gas sensor 520 has a 40% shorter response time; the vibration sensor 530 has a 30% higher sensitivity; and the dust sensor 540 utilizes gravity settling to improve sampling efficiency. Each sensor is properly positioned within a limited space, without interfering with the others, forming a three-dimensional, all-around environmental sensing network.

[0055] Example 3: Modular Gas Sensor Structure like Figure 5 As shown, the gas sensor 520 is composed of an independent explosion-proof module box 521. Each explosion-proof module box 521 includes an explosion-proof shell, a sensor core, a signal processing circuit board, and a storage chip 5211.

[0056] The bottom of the module box 521 is equipped with an explosion-proof quick-connect connector 522 for mechanical and electrical connection with the base socket inside the sensor cavity 200. The explosion-proof quick-connect connector 522 adopts an explosion-proof design and can be plugged in and unplugged while energized, meeting the safety requirements for on-site maintenance.

[0057] The storage chip 5211 is embedded on the signal processing circuit board and is used to store the module's digital identification information, including parameters such as model, range, calibration curve, response time constant, and manufacturing date. When the explosion-proof module box 521 is inserted into the socket, the synchronous triggering and processing unit 600 automatically reads the information in the storage chip 5211 through the communication bus to identify the characteristics and parameters of the newly connected sensor.

[0058] More importantly, the synchronization triggering and processing unit 600 dynamically adjusts the synchronization triggering timing and data processing algorithm associated with the new sensor based on its response characteristics. For example, for a slower-responding electrochemical sensor, the system will trigger its sampling earlier to align its output peak time with the image sensor's acquisition time; for a faster-responding infrared sensor, a conventional triggering timing is used. Simultaneously, the system updates the sensor's weight and threshold in the fusion algorithm. For instance, after replacing it with a more sensitive CH4 sensor, its weight in the gas-image fusion alarm is automatically increased, and the linkage logic associated with that sensor is adjusted.

[0059] In this embodiment, the dimensions and interfaces of each explosion-proof module box 521 adopt a unified standard, supporting the interchangeable use of different types of sensors. When a gas sensor fails or needs to be upgraded, maintenance personnel can directly replace it on-site with live plug-and-play functionality, without disassembling the entire unit, and the replacement time is less than 5 minutes. Spare parts inventory is reduced from complete unit spare parts to module spare parts, reducing costs by more than 70%.

[0060] Through the above-mentioned modular hardware and adaptive software design, this invention achieves the adaptive capability of hardware upgrade, i.e. system upgrade. The new sensor can automatically integrate into the original multimodal sensing network and re-establish spatiotemporal correlation with image sensors, vibration sensors, etc., to form a new collaborative working mode.

[0061] Example 4: Synchronization Triggering and Processing Unit like Figure 7 As shown, the synchronization triggering and processing unit 600 includes a synchronization controller 610 and a timing module 620.

[0062] The synchronization controller 610 is implemented using an FPGA (Field Programmable Gate Array) chip and has high-precision timing control capabilities. The synchronization controller 610 is connected to all sensors that need to be synchronized, such as the image sensor 510, gas sensor 520, vibration sensor 530, and dust sensor 540, through multiple synchronization signal lines.

[0063] The timing module 620 supports dual-mode GPS and BeiDou satellite signal reception, providing a nanosecond-level absolute time reference. The timing module 620 periodically receives satellite signals to calibrate the internal clock of the synchronization controller 610, eliminating clock drift that may occur during long-term operation.

[0064] During operation, the synchronization controller 610 generates a high-precision synchronization pulse signal according to a preset acquisition frequency (e.g., 1 second / time) and simultaneously sends it to all sensors via a synchronization signal line. Each sensor begins acquiring data on the same trigger edge (e.g., rising edge) upon receiving the synchronization pulse. After acquisition, the synchronization controller 610 adds a unified timestamp based on an absolute time reference to each set of multimodal data, achieving microsecond-level hardware synchronization of multi-source data.

[0065] like Figure 7 As shown in the timing diagram, the synchronization controller 610 also establishes a multi-rate sampling coordination mechanism based on the dynamic characteristics of different sensors. The image sensor 510 operates at a fixed frame rate (e.g., 25 frames / second); the gas sensor 520 has a slower response (T90 response time is on the order of seconds), so the synchronization controller 610 triggers its sampling in advance to align its output peak time with the image sensor's acquisition time; the vibration sensor 530 requires high-frequency sampling (e.g., 10kHz), so the synchronization controller 610 adopts a burst acquisition mode to automatically increase the sampling rate during sensitive periods of gas concentration change.

[0066] In this embodiment, the synchronization controller 610 and the power management system 800 work together to generate a power control timing sequence. During non-sampling periods, the synchronization controller 610 instructs the power management system 800 to cut off the power supply to non-critical sensors such as gas sensors and vibration sensors, leaving only the main control and communication modules in standby mode. Before the sampling time arrives, the synchronization controller 610 notifies the power management system 800 to restore power to all sensors that need to be sampled synchronously at a time microseconds in advance. After the power supply stabilizes, the synchronization pulse arrives just in time, and all sensors complete the data acquisition synchronously. After the data acquisition is completed, the power supply to non-essential sensors is cut off again.

[0067] Through the aforementioned synchronous triggering and power supply control linkage, this invention achieves deep coupling of power supply triggering and triggering synchronization. The system's peak power consumption is reduced by more than 40%, and the average heat generation is reduced by 60%, greatly alleviating the heat dissipation pressure within the sealed explosion-proof housing. At the same time, it ensures data synchronization—all sensors resume power supply and collect data at the same time, avoiding the startup time differences and data misalignment caused by traditional time-sharing power supply.

[0068] In applications that detect sudden leaks, the vibration sensor 530 can capture acoustic emission signals at the microsecond level, while the gas sensor 520 takes several seconds to respond. The synchronous controller 610 correlates the time window of the vibration event with the gas response curve through timing association, realizing coordinated detection of vibration warning and gas confirmation, and advancing the warning time by more than 60%.

[0069] Example 5: Self-cleaning and Health Management System like Figure 4 , Figure 6 and Figure 8 As shown, the present invention also includes a self-cleaning and health management system 1000, which includes a cleaning actuator 1100 and a status monitoring module 1200.

[0070] The status monitoring module 1200 is integrated into the synchronous triggering and processing unit 600, and performs real-time self-diagnosis using data collected by the sensors themselves. For image sensor 510, image sharpness is evaluated by calculating an image gradient energy function (such as the Tenengrad function). When the image sharpness evaluation value is lower than a preset threshold, it is judged as optical window contamination.

[0071] For the gas sensor 520, sensitivity is evaluated through periodic self-tests. For example, a small current is applied periodically to observe the response, or the sensor is briefly exposed to a standard gas to verify the response value.

[0072] For all sensors, the signal-noise characteristics are evaluated by analyzing the background noise level under no-excitation conditions.

[0073] The cleaning actuator 1100 includes a windshield wiper 1110, an air curtain nozzle 1120, and a backflushing device 1130.

[0074] The windshield wiper 1110 is located outside the optical window 201 of the image sensor 510, driven by a miniature explosion-proof motor, and performs physical wiping cleaning via a rubber blade. The windshield wiper 1110 can rotate in both directions and is equipped with a washer fluid spray device for removing oily stains.

[0075] Air curtain nozzles 1120 are arranged around the optical window 201 and connected to a compressed air source through a solenoid valve and an external air source interface. Air curtain nozzles 1120 can spray high-speed airflow to form an air knife-like protective air curtain in front of the optical window 201, which can both prevent dust from adhering and assist the wipers in cleaning.

[0076] The backflush device 1130 is connected to the air inlet of the gas sensor 520 or the dust sensor 540, and includes a miniature solenoid valve and a reverse air blowing channel. After each sampling is completed, the backflush device 1130 automatically turns on, using compressed air to backflush the filter screen and remove accumulated dust.

[0077] The condition monitoring module 1200 works in conjunction with the cleaning actuator 1100 to form a closed-loop maintenance chain of monitoring, cleaning, and verification. When the status monitoring module 1200 determines that the image clarity is below a threshold, it first activates the air curtain nozzles 1120 to blow away the image. If the clarity does not recover after blowing, it activates the windshield wipers 1110 for physical wiping. If the image clarity is still not up to standard, it activates both the windshield wipers 1110 and the air curtain nozzles 1120 simultaneously to blow away the scraped-off dirt using the air curtain, preventing secondary contamination. After cleaning, the status monitoring module 1200 reassesses the image clarity to verify the cleaning effect. If the image clarity returns to normal, the cleaning event is recorded; if the image clarity is still not up to standard, a maintenance warning is generated.

[0078] For the gas sensor 520 and the dust sensor 540, the condition monitoring module 1200 links the backflushing device 1130 to the sampling cycle, automatically backflushing after each sampling to extend the filter life by 3 times. When a decrease in sensor sensitivity is detected, the backflushing frequency is automatically increased or a deep cleaning program is triggered.

[0079] In this embodiment, the self-cleaning and health management system 1000 enables the equipment to adapt to environmental changes in heavy dust environments and operate stably for a long time without human intervention. Taking underground mining applications as an example, traditional equipment requires monthly manual maintenance, while the maintenance frequency of this equipment is reduced to once every six months, reducing maintenance costs by more than 80%.

[0080] Example 6: Power Management System like Figure 4As shown, the power management system 800 is housed within the electrical cavity 300 and connected to the synchronous triggering and processing unit 600. The power management system 800 includes a main power input terminal, an intrinsically safe backup battery, and an intelligent power management circuit.

[0081] The main power input terminal is used to connect a 24VDC intrinsically safe input power supply, which supplies power to the system after filtering and voltage regulation.

[0082] The intrinsically safe backup battery uses an intrinsically safe lithium battery pack that automatically switches power when the main power supply fails. The backup battery has a built-in battery management system that monitors battery status in real time, ensuring reliable operation within a wide temperature range of -40℃ to +70℃. The backup battery has a runtime of ≥2 hours, ensuring the system can continue monitoring and uploading alarm information even during power outages.

[0083] The intelligent power management circuit is connected to the synchronization triggering and processing unit 600 and receives the power control timing sequence generated by the synchronization controller 610. Based on this timing sequence, the intelligent power management circuit implements time-sharing power supply control for each sensor. During non-sampling periods, the power supply to non-critical sensors such as gas sensor 520 and vibration sensor 530 is cut off, leaving only the synchronous triggering and processing unit 600, communication module and main control circuit board 900 in standby mode, minimizing power consumption.

[0084] Before the sampling time arrives, power is restored to all sensors that need to be sampled synchronously within microseconds, based on the synchronization trigger signal.

[0085] Once the power supply is stable, each sensor completes data acquisition simultaneously under the triggering of a synchronization pulse.

[0086] After data collection is complete, power to non-essential sensors is cut off again, and the system enters the next sleep cycle.

[0087] Through the aforementioned power supply control and synchronous triggering linkage mechanism, this invention achieves a reduction of over 40% in peak system power consumption and a 60% reduction in average heat generation while ensuring data synchronization. Taking a typical configuration as an example, the traditional continuous power supply solution consumes approximately 15W, while this solution reduces the average power consumption to 6W and controls the peak power consumption to within 9W, significantly alleviating the heat dissipation pressure within the sealed explosion-proof enclosure.

[0088] Example 7: Multimodal Cooperative Sensing Method This embodiment provides a multimodal collaborative sensing method based on the aforementioned terminal, such as... Figure 9 As shown, it includes the following steps: S1. Establish a stable working foundation for spatial isolation. The dual-cavity isolation structure concentrates the main heat sources and radio frequency interference sources in the electrical cavity 300, while placing the sensitive sensors in the sensor cavity 200, providing an intrinsically safe operating environment with low thermal disturbance and low electromagnetic interference for the multi-sensor module 500. This step is the foundation for all subsequent sensing functions, ensuring that multiple sensors can coexist stably and reliably within the same explosion-proof enclosure.

[0089] S2. Achieve synchronous coupling in the time dimension. The synchronization triggering and processing unit 600 generates a high-precision synchronization pulse based on the absolute time reference provided by the timing module 620, and simultaneously sends it to all sensors via the synchronization signal line. Each sensor begins acquiring data on the same trigger edge of the synchronization pulse. The synchronization triggering and processing unit 600 adds a unified timestamp to each group of multimodal data, achieving microsecond-level data synchronization. This step precisely couples physically isolated sensors in the time dimension, providing high-quality spatiotemporally aligned data for multimodal fusion analysis.

[0090] S3, Edge Blending Analysis and Anomaly Detection The synchronous triggering and processing unit 600 performs edge fusion analysis on the synchronously acquired multimodal data to identify abnormal environmental events. This step utilizes the spatiotemporal correlation between image data, gas data, and vibration data for collaborative judgment.

[0091] Specifically, taking chemical leak detection as an example: Gas sensor 520 detected an abnormal increase in CH4 concentration; The synchronous triggering and processing unit 600 immediately retrieves image data at the same time (within ±1ms) and uses image analysis algorithms to identify whether there is visible smoke or a leak source in the abnormal area; Simultaneously retrieve vibration data at the same time to analyze whether there are high-frequency acoustic emission signals; If the multimodal data are strictly synchronized in time, and the image analysis shows that the abnormal area coincides with the spatial location of the gas concentration increase, and the vibration analysis detects an acoustic emission signal that matches the leakage characteristics, then the multimodal data form a spatiotemporally consistent chain of evidence, and the system determines it to be a real leakage event with a confidence level of over 99.5%. If only the gas concentration is abnormal, but the image and vibration data do not show corresponding abnormalities, it is determined to be a sensor drift or interference event. No alarm is triggered, but the event is recorded and the gas sensor's self-test or cleaning process is triggered.

[0092] This step achieves a synergistic detection effect of 1+1+1>3, reducing the false alarm rate of a single sensor by more than 90%, while avoiding missed alarms and realizing accurate and reliable early warning capabilities.

[0093] S4, Self-cleaning and Health Management Integration Based on the fusion analysis results of step S3 and the real-time monitoring information of the status monitoring module 1200, the self-cleaning and health management system 1000 is linked to perform targeted cleaning on the sensors whose performance has deteriorated.

[0094] For example, when the status monitoring module 1200 detects a decrease in image clarity, it selects a cleaning method based on the degree of degradation: for light pollution, it activates an air curtain cleaning system; for moderate pollution, it activates the windshield wipers; and for heavy pollution, it activates both the windshield wipers and the air curtain for coordinated cleaning. After cleaning is completed, the status monitoring module 1200 reassesses the sensor status, verifies the cleaning effect, and forms a complete closed loop.

[0095] When step S3 identifies that a gas sensor may be drifting, the system triggers the sensor's self-test program and activates the backflushing device 1130 for deep cleaning. After cleaning, the sensor response is verified again.

[0096] S5, data upload and cloud platform collaboration The synchronized and integrated high-value data and early warning information are uploaded to an external platform (cloud monitoring center or mobile terminal) via the communication module. The uploaded data includes: raw multimodal data with a unified timestamp, edge fusion analysis results, event early warning information, and device health status reports. Through edge processing and cloud collaboration, both the timeliness of real-time early warnings (edge-side response time <100ms) and the ability to aggregate data and perform global analysis are ensured.

[0097] The method described in this embodiment elevates multimodal perception to the level of "system-level collaboration." Through the interconnectedness of four dimensions—"spatial decoupling, temporal coupling, information fusion, and execution closed loop"—a complete perception-analysis-execution closed-loop system is formed, representing the technological evolution direction of industrial safety monitoring.

[0098] Example 8: Application Case of Chemical Tank Farm This embodiment uses a chemical tank farm as an application scenario to specifically illustrate the deployment and operational effects of the present invention.

[0099] The terminal of this invention is installed outside the cofferdam of the chemical tank area, 10 meters away from the tank body and 3 meters above it. The configuration scheme is as follows: Gas sensor 520: CH4, H2S, VOCs, CO four-channel module; Other sensors: Fully equipped, including a multispectral camera, an infrared thermal imaging sensor, a laser dust sensor 540, a temperature and humidity sensor, and a vibration sensor 530.

[0100] During operation, on a certain day, the terminal captured a weak ultrasonic signal (frequency 40kHz, amplitude increased by 15dB from the baseline) through the vibration sensor 530. At the same time, the gas sensor 520 detected a trace increase in VOCs concentration (from 0ppm to 2ppm), and the infrared thermal imaging sensor showed that the temperature in a certain valve area increased abnormally by 2℃.

[0101] The synchronous triggering and processing unit 600 retrieves multispectral images from the same moment. Image analysis reveals no visible smoke, but infrared images show thermal anomalies in the valve area. Based on the combined data from four modes (vibration, gas, infrared, and visible light), the system determines that there is a minor valve leak with a 95% confidence level and immediately issues an early warning.

[0102] The early warning information was uploaded to the central platform via the communication module. The platform, combined with video footage, accurately located the leak point (the feed valve of tank No. 3) and notified maintenance personnel to handle it promptly. On-site inspection revealed that the valve seal was aged; after replacement, the valve returned to normal operation.

[0103] In this example, thanks to the spatiotemporal synchronization and collaborative fusion of multimodal data, the system issued an early warning at the initial stage of leakage (leakage rate <0.1L / min), approximately 30 minutes earlier than traditional single-gas detectors, thus preventing potential safety accidents caused by the leakage escalating. Simultaneously, the self-cleaning mechanism operates stably in corrosive chemical environments, requiring no manual maintenance for six months.

[0104] Example 9: Application Case in Underground Mining This embodiment uses an underground mine as an application scenario to specifically illustrate the deployment and operational effects of the present invention.

[0105] The terminal of this invention is installed on the side wall at the intersection of the main mine roadways. The configuration scheme is as follows: Gas Sensor 520: CH4, CO, O2, H2S four-channel module; Key areas for improvement: Laser dust sensor 540, vibration sensor 530; Other sensors: standard configuration.

[0106] During operation, on a certain day, the CH4 sensor detected that the concentration gradually increased from 0.1% to 0.5% (10% of the lower explosive limit). At the same time, the vibration sensor 530 detected low-frequency vibration signals (frequency 10-50Hz, amplitude increased) and high-frequency acoustic emission signals (frequency 20-40kHz). Micro-crack features appeared in the top plate area.

[0107] Synchronous triggering and processing unit 600 performed fusion analysis: the rising trend of gas concentration and the time correlation of micro-fracture events were highly consistent, and spatially they both pointed towards the mining face. The system determined it to be an abnormal gas accumulation accompanied by precursors to roof instability, with a confidence level of 98%, and issued an evacuation warning 30 minutes in advance.

[0108] After the warning was issued, the mine immediately organized the evacuation of personnel. Twenty minutes later, a small-scale roof collapse occurred, but due to the advance warning, there were no casualties. Meanwhile, the backup battery continued to operate for three hours in the event of a main power outage, ensuring environmental monitoring and alarm capabilities during the power outage.

[0109] In this example, the coordinated detection of vibration and gas sensors enabled early warning of complex disasters, extending the warning time from several minutes in traditional methods to over 30 minutes, thus gaining valuable time for personnel evacuation. Simultaneously, the self-cleaning mechanism operates stably in dusty environments, and the back-flushing device automatically cleans daily, ensuring the long-term reliability of the dust sensors.

[0110] Example 10: Performance Comparison Data To further illustrate the beneficial effects of the present invention, a comparative test was conducted between the terminal of the present invention and the prior art, and the results are shown in Table 1 below: Table 1 Comparison of the performance of the terminal of the present invention with existing technologies The above data shows that the present invention achieves a system-level technical effect of 1+1>2 through the synergistic effect of multiple technical features such as dual-cavity isolation, distributed layout, hardware synchronization, modular design, self-cleaning and health management, and power management linkage. It has made significant progress in terms of integration, reliability, accuracy, and economy.

[0111] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal, characterized in that, include: The explosion-proof housing (100) is internally divided into a sensor cavity (200) and an electrical cavity (300) that are isolated from each other by a partition (110). A multi-sensor module (500) is installed in the sensor cavity (200) and includes at least an image sensor (510) for acquiring environmental image information, a gas sensor (520) for acquiring environmental gas concentration, and a physical quantity sensor for acquiring environmental physical quantities. A synchronous triggering and processing unit (600) is disposed in the electrical cavity (300) and connected to the multi-sensor module (500) through a sealed terminal block (400). It is used to generate a synchronous triggering signal, control each sensor in the multi-sensor module (500) to collect data at the same time, and perform edge processing on the collected data. The physical isolation between the sensor cavity (200) and the electrical cavity (300) provides an intrinsically safe environment with low thermal disturbance and low electromagnetic interference for the stable operation of the multi-sensor module (500). The time synchronization mechanism of the synchronous triggering and processing unit (600) couples the physically isolated multi-sensor module (500) in the time dimension, so that the data collected by the multi-sensor module (500) has a unified time base.

2. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 1, characterized in that, The multi-sensor module (500) is arranged in a distributed layout within the sensor cavity (200) that is compatible with the explosion-proof housing (100). The installation positions of each sensor are differentiated according to its physical characteristics and sampling requirements. The image sensor (510) is disposed on the top of the sensor cavity (200) to obtain an unobstructed field of view; the gas sensor (520) is disposed on the side wall of the sensor cavity (200) and is in direct contact with the environment through the diffusion hole (202); The vibration sensor (530) in the physical quantity sensor is attached to the inner wall of the sensor cavity (200) and is used to collect structural vibration signals transmitted through the explosion-proof housing (100).

3. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 2, characterized in that, The gas sensor (520) includes multiple units, each consisting of an independent explosion-proof module box (521) arranged in a circular pattern. The explosion-proof module box (521) is connected to the main control circuit board (900) via an explosion-proof quick connector (522). The explosion-proof module box (521) has an embedded storage chip (5211) for storing its identification information and calibration parameters. When any explosion-proof module box (521) is replaced, the synchronous triggering and processing unit (600) automatically reads the storage chip (5211) of the new explosion-proof module box (521), identifies its type and parameters, and dynamically adjusts the synchronous triggering timing and data processing algorithm associated with it to achieve plug-and-play functionality and form a new collaborative working mode with the remaining sensors.

4. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 1, characterized in that, The synchronization triggering and processing unit (600) includes: Synchronization controller (610) is used to generate high-precision synchronization pulses; The timing module (620) is used to receive satellite signals and provide an absolute time reference; The synchronization controller (610) is connected to each sensor respectively, controls each sensor to start collecting data on the same trigger edge of the synchronization pulse, and adds a unified timestamp based on the absolute time reference to each group of multimodal data to achieve microsecond-level synchronization; The synchronization controller (610) also dynamically adjusts the sampling timing of the gas sensor (520) according to its response characteristics, so that it is associated with the frame rate of the image sensor (510) and the sampling rate of the vibration sensor (530), so that the sensors with different dynamic characteristics can coordinate and match in the time dimension.

5. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 1, characterized in that, It also includes a self-cleaning and health management system (1000), the system comprising: A cleaning actuator (1100) is disposed outside and inside the sensor cavity (200) for cleaning the sensitive element of the sensor; The status monitoring module (1200), integrated in the synchronous triggering and processing unit (600), is used to monitor the working status of each sensor in real time, including image clarity, gas sensor sensitivity and signal noise level; The status monitoring module (1200) is linked with the cleaning actuator (1100): when the status monitoring module (1200) determines that the working state of any sensor has dropped to a preset threshold, it automatically triggers the corresponding cleaning actuator (1100) to perform targeted cleaning, and re-evaluates the sensor status after cleaning, forming a closed-loop maintenance link of monitoring, cleaning and verification.

6. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 5, characterized in that, The cleaning actuator (1100) includes: A windshield wiper (1110) is disposed outside the optical window (201) of the image sensor (510) for physical wiping cleaning; An air curtain nozzle (1120) is arranged around the optical window (201) to spray gas to form a protective air curtain; a back-blowing device (1130) is connected to the air inlet of the gas sensor (520) for back-blowing. The status monitoring module (1200) selectively activates a single cleaning actuator (1100) or combines multiple cleaning actuators (1100) to achieve coordinated cleaning, depending on the sensor type and the degree of contamination.

7. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 1, characterized in that, It also includes a power management system (800), which is located in the electrical cavity (300) and connected to the synchronization triggering and processing unit (600); the power management system (800) implements time-sharing power supply control for each sensor according to the synchronization timing generated by the synchronization triggering and processing unit (600): during non-sampling periods, the power supply of unnecessary sensors is cut off to reduce power consumption and heat accumulation; At the sampling moment, based on the synchronization trigger signal, power is simultaneously restored to all sensors that need to be sampled synchronously within microseconds, and the data acquisition is completed. Through the linkage between power supply control and synchronization trigger, the peak power consumption and average heat generation of the system are significantly reduced while ensuring data synchronization.

8. The dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to claim 1, characterized in that, The explosion-proof housing (100) is provided with heat dissipation fins (101) on the back. The heat dissipation fins (101) are integrally formed or tightly attached to the housing wall corresponding to the electrical cavity (300) to form a passive heat dissipation path. The heating devices inside the electrical cavity (300) are centrally arranged and connected to the heat dissipation fins (101) through a heat-conducting structure to directionally dissipate heat; The thermal isolation design between the sensor cavity (200) and the electrical cavity (300), in conjunction with the directional heat dissipation design of the heat sink fins (101), enables heat to be efficiently discharged along a preset path, avoiding the influence of heat sources on the sensor cavity (200).

9. A multimodal collaborative sensing method based on the dual-cavity isolated explosion-proof multimodal environmental safety sensing terminal according to any one of claims 1-8, characterized in that, Includes the following steps: S1. Through the dual-cavity isolation structure, a stable working basis with spatial isolation is provided for the multi-sensor module (500); S2. A synchronization pulse is generated by the synchronization triggering and processing unit (600) to control the spatially isolated sensors to collect data at the same time and attach a unified timestamp to achieve synchronous coupling of data in the time dimension. S3. Perform edge fusion analysis on the synchronously acquired multimodal data to identify abnormal environmental events. The fusion analysis utilizes the spatiotemporal correlation of image data, gas data, and vibration data to make collaborative judgments and reduce the false alarm rate of a single sensor. S4. Based on the results of the fusion analysis and the sensor status monitoring information, the self-cleaning and health management system (1000) is linked to perform targeted cleaning on sensors with degraded performance. S5. Upload the synchronized and integrated data and early warning information to an external platform.

10. The multimodal cooperative sensing method according to claim 9, characterized in that, In step S3, the edge fusion analysis specifically includes: when the gas sensor (520) detects an abnormal increase in concentration, the processing unit (600) is simultaneously triggered to retrieve image data at the same moment, and the image analysis is used to identify whether there is visible smoke or a leak source in the abnormal area. At the same time, the vibration data at the same moment is retrieved to analyze whether there is a high-frequency acoustic emission signal. If the multimodal data are strictly synchronized in time and point to the same area in space, it is determined to be a real leak event. Otherwise, according to the type and characteristics of the abnormal mode, it is identified as a sensor interference or drift event, and the corresponding self-check or cleaning process is triggered.

Images