Methane recovery and purification system in intelligent cleaning equipment

Abstract

The invention provides a methane recovery and purification system in intelligent cleaning equipment, which relates to the technical field of environmental protection, and is characterized in that methane-containing gas is collected from the cleaning equipment, moisture and small particles in the gas are removed to obtain preliminarily purified gas, and then methane is separated and recovered by using a separation technology to obtain purified gas. And the purified methanol is obtained through multi-stage purification treatment. According to the methane recovery technology, a time sequence of gas components, a time sequence of physical properties and a time sequence of process parameters need to be obtained, and local time sequence coding and causal context time sequence aggregation are carried out, so that a separation technology recommendation result is recommended, and operation parameters are dynamically adjusted. Therefore, subjective experience can be abandoned, gas characteristics and process requirements are monitored in real time and accurately matched, the accuracy and objectivity of separation technology selection are improved, and it is ensured that the separation process is always in the optimal state.

Description

Technical Field

[0001] This application relates to the field of environmental protection technologies, and more particularly, in the embodiments of this application, to a methane recovery and purification system in an intelligent dredging device. Background Art

[0002] With the enhancement of environmental protection awareness and the progress of energy utilization technologies, the recovery and utilization of methane in waste have become increasingly important. Methane, as a gas widely produced in nature and human activities, is present in significant amounts in waste. It is not only a greenhouse gas with an extremely high global warming potential, with a warming effect approximately 28 - 36 times that of carbon dioxide, but also a highly efficient and clean energy source, producing relatively less carbon dioxide emissions when burned, and thus has great development and utilization value.

[0003] However, existing methane recovery devices usually have low recovery efficiency, unstable purity, and lack of intelligent control, making it difficult to meet the requirements of efficient and stable operation. These devices often rely on traditional technologies with fixed operating parameters and cannot be dynamically adjusted according to real-time working conditions, resulting in energy waste and poor treatment effects. In addition, existing technologies also face challenges in terms of operating costs and long-term maintenance, and cannot effectively meet the current environmental protection and energy reuse requirements.

[0004] Therefore, an optimized methane recovery and purification solution in an intelligent dredging device is desired. Summary of the Invention

[0005] To solve the above technical problems, this application is proposed. Embodiments of this application provide a methane recovery and purification system in an intelligent dredging device, which first collects methane-containing gas from the dredging device, removes moisture and fine particles in the gas to obtain a preliminarily purified gas, then uses separation technology to separate and recover methane, and obtains purified methanol through multi-stage purification treatment. Among them, the methane recovery technology needs to obtain time series of gas components, time series of physical properties, and time series of process parameters, and perform local time series encoding and causal context time series aggregation to recommend separation technology recommendation results and dynamically adjust operating parameters. This can discard subjective experience, monitor in real time and accurately match gas characteristics with process requirements, improve the accuracy and objectivity of separation technology selection, and ensure that the separation process is always in an optimal state.

[0006] According to one aspect of this application, there is provided a methane recovery and purification system in an intelligent dredging device, which includes: A methane collection module for collecting methane-containing gas from the dredging device; A pretreatment module for performing preliminary purification treatment on the collected methane-containing gas to remove moisture and fine particles in the methane-containing gas and obtain a preliminarily purified gas; A recovery module for separating the preliminarily purified gas through a separation technique to recover the separated methane, where the separation technique is one or a combination of two of membrane separation technique or pressure swing adsorption technique; A purification module for performing multi-stage purification on the separated methane to obtain purified methanol.

[0007] Compared with the prior art, the methane recovery and purification system in an intelligent cleaning device provided by the present application first collects methane-containing gas from the cleaning device, removes moisture and fine particles in the gas to obtain preliminarily purified gas, then separates and recovers methane using a separation technique, and obtains purified methanol through multi-stage purification. Among them, the methane recovery technology needs to obtain the time series of gas components, the time series of physical properties, and the time series of process parameters, and perform local time series coding and causal context time series aggregation to recommend the separation technique recommendation result and dynamically adjust the operation parameters. This can discard subjective experience, monitor in real time and accurately match the gas characteristics with the process requirements, improve the accuracy and objectivity of the separation technique selection, and ensure that the separation process is always in the optimal state. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 It is a system block diagram of the methane recovery and purification system in the intelligent cleaning device according to the embodiment of the present application.

[0010] Figure 2 It is a flowchart of the methane collection module in the methane recovery and purification system in the intelligent cleaning device according to the embodiment of the present application.

[0011] Figure 3 It is a block diagram of the recovery module in the methane recovery and purification system in the intelligent cleaning device according to the embodiment of the present application.

[0012] Figure 4 It is a schematic diagram of the data flow of the recovery module in the methane recovery and purification system in the intelligent cleaning device according to the embodiment of the present application.

[0013] Figure 5 It is a block diagram of the multi-dimensional integration unit of separation parameters in the methane recovery and purification system in the intelligent cleaning device according to the embodiment of the present application.

[0014] Figure 6Block diagram of the sequential causal context aggregation subunit in the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application.

[0015] Figure 7 Block diagram of the gas composition topological feature construction secondary subunit in the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application.

[0016] Figure 8 Flow chart of the purification module in the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application. Detailed implementation manners

[0017] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0018] The special term "exemplary" herein means "serving as an example, embodiment or illustration". Any embodiment described as "exemplary" herein need not be construed as superior or better than other embodiments.

[0019] In addition, for a better description of the present application, numerous specific details are given in the following detailed implementation manners. Those skilled in the art should understand that the present application can be implemented without some specific details. In some instances, methods, means, elements and circuits well known to those skilled in the art are not described in detail so as to highlight the gist of the present application.

[0020] Furthermore, the terms "first" and "second" are used for descriptive purposes only and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality of" means two or more unless otherwise specifically defined.

[0021] With the enhancement of environmental protection awareness and the progress of energy utilization technology, the recovery and utilization of methane in waste have become increasingly important. However, existing methane recovery equipment usually has low recovery efficiency, unstable purity and lacks intelligent control, making it difficult to meet the requirements of efficient and stable operation. In addition, existing technologies also pose challenges in terms of operating costs and long-term maintenance, and cannot effectively meet the current requirements of environmental protection and energy reuse.

[0022] In view of the above technical problems, the present application proposes a methane recovery and purification system in an intelligent cleaning equipment. Figure 1 System block diagram of the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application. AsFigure 1 As shown, the methane recovery and purification system 100 in the intelligent dredging device according to an embodiment of the present application includes: a methane collection module 110 for collecting methane-containing gas from the dredging device; a pretreatment module 120 for preliminarily purifying the collected methane-containing gas to remove moisture and fine particles in the methane-containing gas, obtaining a preliminarily purified gas; a recovery module 130 for separating the preliminarily purified gas by separation technology to recover the separated methane, the separation technology being one or a combination of two of membrane separation technology or pressure swing adsorption technology; and a purification module 140 for performing multi-stage purification on the separated methane to obtain purified methanol.

[0023] Figure 2 It is a flowchart of the methane collection module in the methane recovery and purification system of the intelligent dredging device according to an embodiment of the present application. As Figure 2As shown, in the methane recovery and purification system 100 of the above intelligent dredging equipment, the methane collection module 110 is used to collect methane-containing gas from the dredging equipment. It should be understood that methane, as a greenhouse gas, has a warming effect about 28 to 36 times that of carbon dioxide, and at the same time it is also a highly efficient and clean energy source. Recycling methane through effective technologies and methods not only helps environmental protection but also provides a renewable clean energy source. Specifically, in the technical solution of this application, the methane collection module is mainly responsible for collecting methane-containing gas from the dredging equipment. First, an air suction device (such as an air suction pump) is used to suck the methane-containing gas in the dredging equipment into the collection chamber. In this way, it is ensured that the gas can be effectively guided from the source into the treatment system, avoiding direct emission into the atmosphere and causing pollution. In actual operation, the air suction pump will be automatically started according to a preset time interval or the detected gas concentration to ensure continuous and stable introduction of methane-containing gas into the collection chamber. To ensure the continuity and stability of the inhaled gas, a dedicated pipeline is usually set between the dredging equipment and the collection chamber. This pipeline not only needs to have good sealing performance but also sufficient corrosion resistance to cope with the possibly complex chemical composition of the exhaust gas environment. Then, a filtering device such as a metal mesh and an activated carbon layer provided in the collection chamber is used to remove large particle impurities in the gas. Considering that if large particle impurities are not effectively removed, they may affect subsequent treatment steps, such as clogging pipelines or damaging precision equipment. The metal mesh, as a primary filtering device, can intercept larger solid particles such as gravel and dust. The activated carbon layer has a stronger adsorption capacity. It can not only further remove fine particles but also initially adsorb some harmful gases, laying a foundation for subsequent more refined purification steps. In addition, the activated carbon also has a certain buffering effect, which can relieve the impact caused by too fast gas flow rate to a certain extent, making the whole gas flow more stable. In a specific embodiment of this application, the filtering device is a combination of a metal mesh and an activated carbon layer. Finally, the purified gas is sent through a pipeline to the pretreatment module for further treatment.

[0024] In the methane recovery and purification system 100 of the above intelligent cleaning equipment, the pretreatment module 120 is used to perform preliminary purification on the collected methane-containing gas to remove moisture and fine particles in the methane-containing gas, and obtain the preliminarily purified gas. It should be understood that methane gas usually contains various impurities, such as moisture, carbon dioxide, nitrogen, and other trace gas components, and may also contain fine particles. The presence of these impurities will have an adverse impact on subsequent processing steps. Among them, moisture will cause pipeline corrosion, reduce the equipment life, and may condense into ice under low-temperature conditions, blocking pipelines or valves, seriously affecting the normal operation of the system. In addition, fine particles will also cause wear to the equipment, increase the maintenance cost, and may cause the adsorbent or membrane material to fail, thereby reducing the separation efficiency and purity. Through effective preliminary purification, these problems can be significantly reduced, providing favorable conditions for subsequent efficient recovery and purification. Specifically, during implementation, first, a desiccant (such as molecular sieve, silica gel, etc.) is used for dehydration treatment to remove moisture in the gas and prevent equipment corrosion and reduction of methane purity during subsequent processing. These desiccants have good moisture absorption performance and can effectively remove moisture in the gas. Among them, the molecular sieve is a porous material with a large number of micropores in its internal structure, which can adsorb a large number of water molecules. In practical applications, the molecular sieve is usually filled in a dedicated container. When the methane-containing gas passes through this container, the moisture will be adsorbed by the molecular sieve, thereby achieving the purpose of dehydration. To ensure the drying effect, the molecular sieve needs to be regenerated regularly, that is, the adsorbed moisture is released by heating or other methods to restore its adsorption capacity. In addition to the molecular sieve, silica gel is also a common desiccant. Compared with the molecular sieve, silica gel has a larger surface area and stronger moisture absorption capacity, and is suitable for different humidity environments. The advantage of silica gel is that its regeneration temperature is relatively low and the operation is relatively simple, but its moisture absorption capacity is relatively small. Therefore, it may need to be replaced or regenerated frequently in a high-humidity environment. To further improve the dehydration effect, sometimes the molecular sieve and silica gel are combined to form a composite drying system. In this system, the gas first passes through the silica gel layer for preliminary dehydration, and then passes through the molecular sieve layer for deep dehydration, thereby ensuring that the moisture content in the gas is reduced to the lowest level. Then, a high-efficiency filter (such as a HEPA filter) is used to further remove fine particles and dust in the gas for dust removal treatment. After being processed by the pretreatment module, the impurity content of the gas is greatly reduced, providing favorable conditions for subsequent efficient recovery and purification. Among them, the HEPA filter is a high-efficiency air filter. This filter is composed of multiple layers of filter materials, and each layer has a different filtering mechanism, which can effectively intercept various sizes of particulate matter. When the methane-containing gas passes through the HEPA filter, larger particulate matter will be directly intercepted on the outer layer of the filter material, while smaller particulate matter will be captured during the process of passing through multiple layers of filter materials, ultimately greatly reducing the particulate matter content in the gas.

[0025] In the methane recovery and purification system 100 of the above intelligent cleaning equipment, the recovery module 130 is used to separate the preliminarily purified gas through separation technology to recover the separated methane, and the separation technology is one or a combination of two of membrane separation technology or pressure swing adsorption technology.

[0026] Specifically, in a specific embodiment of the present application, the recovery module includes a separation chamber and a recovery port. The separation chamber is used to separate methane from the mixed gas, and the recovery port is used to send the separated methane into the purification module. The membrane separation technology uses a membrane made of special material to selectively permeate the gas, so that methane molecules can pass through the membrane preferentially to achieve the separation purpose. The membrane separation technology has the advantages of low energy consumption and simple operation. The pressure swing adsorption technology is to periodically change the operating pressure to adsorb and desorb methane on the adsorbent (such as activated carbon, zeolite molecular sieve, etc.), so as to realize the separation and recovery of methane. The pressure swing adsorption technology has the characteristics of high recovery efficiency and flexible operation.

[0027] It should be understood that in the process of methane recovery, it is crucial to reasonably select the separation technology. The composition and physical properties of the methane-containing gas to be separated are complex and variable, and the impurity conditions vary greatly. It is possible to have high-concentration carbon dioxide, nitrogen, corrosive hydrogen sulfide, etc. Different separation technologies have their own advantages. Membrane separation relies on the differences in molecular size and properties, and pressure swing adsorption is based on the different adsorption capacities of gases on the adsorbent. At the same time, there are various process parameters for methane recovery, and the treatment capacity, recovery rate, and purity requirements are different. Therefore, by selecting the most suitable separation technology for the current working conditions, the overall efficiency of methane recovery can be maximized. However, the traditional manual selection method based on fixed experience and standards lacks a unified and objective standard, and it is difficult to monitor and evaluate the impact of these changes on the separation technology in real time, and it is impossible to adjust and optimize the separation technology in time, so it cannot meet the production requirements.

[0028] Accordingly, in the recovery module, the technical concept of the present application is to obtain the time series of the gas components (methane concentration, carbon dioxide ratio, and nitrogen ratio), the time series of physical properties (temperature, pressure, and humidity), and the time series of process parameters (treatment volume and recovery rate) of the preliminarily purified gas through sensors, and use artificial intelligence-based data analysis and feature extraction methods to perform local time series encoding on the time series of the gas components, the time series of the physical properties, and the time series of the process parameters respectively. Then, perform causal context time series aggregation on the sequences of the local time series features of each encoded parameter, so as to intelligently recommend the separation technology recommendation result based on the multi-dimensional representation between the dynamically time series context aggregation features of the aggregated gas components, physical properties, and process parameters, and dynamically adjust the operating parameters during the separation process based on this result. The present application abandons subjective experience judgment, can monitor the changes of gas components, physical properties, and process parameters in real time, and can intelligently recommend the most suitable separation technology for the current working conditions based on these changes, greatly improving the accuracy and objectivity of the selection, to more accurately match the gas characteristics with the process requirements. At the same time, dynamically adjusting the operating parameters ensures that the separation process is always in the optimal state.

[0029] Figure 3 It is a block diagram of the recovery module in the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application. Figure 4 It is a schematic diagram of data flow in the recovery module of the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application. As Figure 3 and Figure 4 shown, in the embodiment of the present application, the recovery module 130 includes: a preliminarily purified gas data acquisition unit 131, configured to obtain the time series of the gas components, the time series of physical properties, and the time series of process parameters of the preliminarily purified gas through sensors, where the gas components include methane concentration, carbon dioxide ratio, and nitrogen ratio, the physical properties include temperature, pressure, and humidity, and the process parameters include treatment volume and recovery rate; a separation parameter multi-dimensional integration unit 132, configured to perform dynamic separation parameter multi-dimensional integration based on time series features on the time series of the gas components, the time series of the physical properties, and the time series of the process parameters to obtain a separation technology parameter multi-dimensional integration vector; an operating parameter adjustment unit 133, configured to obtain a separation technology recommendation result based on the separation technology parameter multi-dimensional integration vector, and dynamically adjust the operating parameters during the separation process based on the separation technology recommendation result.

[0030] Specifically, the preliminary purified gas data acquisition unit 131 is used to obtain the time series of the gas components, the time series of the physical properties, and the time series of the process parameters of the preliminarily purified gas through sensors. The gas components include the concentration of methane, the proportions of carbon dioxide and nitrogen. The physical properties include temperature, pressure, and humidity. The process parameters include throughput and recovery rate. It should be understood that during the methane recovery process, changes in gas components, physical properties, and process parameters will directly affect the separation effect and the quality of the final product. Among them, changes in gas components will affect the selection of membrane separation or pressure swing adsorption technology and their operating conditions; changes in physical properties will affect the adsorption capacity of adsorbents or the selective permeation performance of membrane materials; while changes in process parameters will affect the efficiency and economy of the overall system. Real-time monitoring and recording of the time series changes of these parameters can provide dynamic feedback for the system, helping to optimize the separation technology and operating parameters, thereby improving the recovery rate and purity of methane. Specifically, the monitoring of gas components usually uses gas analyzers or sensor arrays. These devices can measure the key components such as methane concentration, carbon dioxide proportion, and nitrogen proportion in the gas in real time. Among them, an infrared absorption spectrometer is a commonly used gas analyzer, which can use the absorption characteristics of different gas molecules for infrared light of specific wavelengths to determine the concentration of gas components. In addition to infrared absorption spectrometers, electrochemical sensors are also a common means of monitoring gas components. This type of sensor is based on the principle of electrochemical reactions and can sensitively detect the presence and concentration of specific gases. To ensure the accuracy and reliability of the data, multiple sensors are usually configured in the system to form a redundant design to prevent data loss caused by the failure of a single sensor. The monitoring of physical properties mainly relies on temperature sensors, pressure sensors, and humidity sensors. The monitoring of process parameters mainly relies on flow meters and level gauges. Among them, common flow meters include differential pressure flow meters, vortex street flow meters, and ultrasonic flow meters, and common level gauges include float type level gauges, magnetic flap level gauges, and radar level gauges.

[0031] Figure 5 Block diagram of the separation parameter multi-dimensional integration unit in the methane recovery and purification system of the intelligent cleaning equipment according to an embodiment of the present application. As Figure 5As shown, in the embodiment of the present application, the separation parameter multi-dimensional integration unit 132 includes: a data local time series encoding subunit 1321, configured to perform local time series encoding on the time series of the gas components, the time series of the physical properties, and the time series of the process parameters respectively to obtain a sequence of local time series correlation feature vectors of the gas components, a sequence of local time series correlation feature vectors of the physical properties, and a sequence of local time series correlation feature vectors of the process parameters; a time series causal context aggregation subunit 1322, configured to perform dynamic separation parameter time series causal context aggregation on the sequence of local time series correlation feature vectors of the gas components, the sequence of local time series correlation feature vectors of the physical properties, and the sequence of local time series correlation feature vectors of the process parameters respectively to obtain a gas component dynamic time series context aggregation feature vector, a physical property dynamic time series context aggregation feature vector, and a process parameter dynamic time series context aggregation feature vector; a separation technical parameter aggregation subunit 1323, configured to perform feature aggregation on the gas component dynamic time series context aggregation feature vector, the physical property dynamic time series context aggregation feature vector, and the process parameter dynamic time series context aggregation feature vector to obtain the separation technical parameter multi-dimensional integration vector.

[0032] In an embodiment of the present application, the data local temporal encoding subunit 1321 is configured to: respectively input the time series of the gas components, the time series of the physical properties, and the time series of the process parameters into a local temporal encoder based on one-dimensional convolution to obtain a sequence of local temporal correlation feature vectors of the gas components, a sequence of local temporal correlation feature vectors of the physical properties, and a sequence of local temporal correlation feature vectors of the process parameters. It should be understood that considering that there are local patterns and short-term dependencies in gas components, physical properties, and process parameters at different local times, such as short-term fluctuations and periodic continuous changes, etc. Therefore, in order to learn the correlation between data at different time points to analyze whether it is increasing, decreasing, or there is a certain specific change pattern, etc., in the technical solution of the present application, the time series of the gas components, the time series of the physical properties, and the time series of the process parameters are respectively input into a local temporal encoder based on one-dimensional convolution to obtain a sequence of local temporal correlation feature vectors of the gas components, a sequence of local temporal correlation feature vectors of the physical properties, and a sequence of local temporal correlation feature vectors of the process parameters. Among them, one-dimensional convolution is particularly suitable for processing data with a sequence structure, such as time series data. For the time series of gas components, physical properties, and process parameters, the local temporal encoder based on one-dimensional convolution can capture the local correlation between adjacent elements of the data on the time axis by sliding the convolution kernel in the time dimension. For example, in the time series of gas components, for the change situations of methane concentration, carbon dioxide ratio, and nitrogen ratio at adjacent time points, one-dimensional convolution can effectively extract such local change patterns and correlation information, and convert these local fluctuation characteristics into feature vectors, so as to better represent the local characteristics of the data.

[0033] Specifically, the temporal causal context aggregation subunit 1322 is configured to perform dynamic separated parameter temporal causal context aggregation on the sequence of the local temporal correlation feature vectors of the gas components, the sequence of the local temporal correlation feature vectors of the physical properties, and the sequence of the local temporal correlation feature vectors of the process parameters respectively to obtain the dynamic temporal context aggregation feature vectors of the gas components, the dynamic temporal context aggregation feature vectors of the physical properties, and the dynamic temporal context aggregation feature vectors of the process parameters. It should be understood that considering the states of the gas components, physical properties, and process parameters at different time points do not exist in isolation. Although one-dimensional convolution can effectively extract local patterns, it is difficult to capture long-term dependencies in time series. That is, there are complex causal relationships between adjacent time points and different gas components (such as methane concentration, carbon dioxide ratio, nitrogen ratio, etc.). For example, the change in methane concentration at a certain moment may be due to the temperature change at the previous moment affecting the relevant chemical reactions, thereby changing the gas component ratio. Moreover, time series data is dynamically changing, and the parameter values at each time point are not only related to adjacent time points but also affected by the context environment in which they are located. Based on this, the present application performs dynamic separated parameter temporal causal context aggregation on the sequence of the local temporal correlation feature vectors of the gas components, the sequence of the local temporal correlation feature vectors of the physical properties, and the sequence of the local temporal correlation feature vectors of the process parameters respectively to obtain the dynamic temporal context aggregation feature vectors of the gas components, the dynamic temporal context aggregation feature vectors of the physical properties, and the dynamic temporal context aggregation feature vectors of the process parameters. In this way, the causal connections hidden in the time series data can be deeply mined to more comprehensively understand the temporal variation mechanism of the gas components, physical properties, and process parameters, thereby providing a more accurate and comprehensive basis for intelligent recommendation of separation technologies and dynamic adjustment of operation parameters.

[0034] Figure 6 It is a block diagram of the temporal causal context aggregation subunit in the methane recovery and purification system of the intelligent cleaning device according to an embodiment of the present application. As Figure 6As shown, in the embodiment of the present application, the timing causal context aggregation subunit 1322 includes: a gas component implicit feature mining secondary subunit 1322-1, configured to perform implicit feature mining on each gas component local timing correlation feature vector in the sequence of gas component local timing correlation feature vectors to obtain a sequence of gas component local timing correlation deep implicit feature encoding vectors; a gas component topological feature construction secondary subunit 1322-2, configured to construct semantic causal association topological features of the sequence of gas component local timing correlation deep implicit feature encoding vectors to obtain a gas component local timing semantic causal association topological feature matrix; a gas component dynamic timing fusion secondary subunit 1322-3, configured to perform feature sequence dynamic fusion of graph convolution on the sequence of gas component local timing correlation deep implicit feature encoding vectors and the sequence of gas component local timing correlation deep implicit feature encoding vectors based on the gas component local timing semantic causal association topological feature matrix to obtain the gas component dynamic timing context aggregation feature vector.

[0035] Specifically, here, the processing process of the sequence of gas component local timing correlation feature vectors is taken as an example for specific illustration.

[0036] Specifically, the gas component implicit feature mining secondary subunit 1322-1 is configured to perform implicit feature mining on each gas component local timing correlation feature vector in the sequence of gas component local timing correlation feature vectors to obtain a sequence of gas component local timing correlation deep implicit feature encoding vectors, which is expressed by the gas component implicit feature mining formula as: ; Where is the sequence of gas component local timing correlation feature vectors, and are respectively the 1st, 2nd, th, and th gas component local timing correlation feature vectors in the sequence of gas component local timing correlation feature vectors, is point convolution encoding, is the convolution encoding activation function, and are respectively the 1st, 2nd, th, th, and th gas component local timing correlation deep implicit feature encoding vectors in the sequence of gas component local timing correlation deep implicit feature encoding vectors, It is a sequence of local temporal correlation depth implicit feature encoding vectors of the gas components. It should be understood that although the gas component time series after preliminary processing already contains certain information, this information is often explicit and surface features, which cannot fully reflect the complex relationships and potential patterns among gas components. Through implicit feature mining, more abstract and deep-level feature representations can be extracted from the original data, thereby revealing these complex patterns and relationships hidden behind the data. Considering that explicit features are usually directly extracted from the original data and are easily affected by noise and outliers, resulting in a decline in model performance, implicit modeling can alleviate the noise problems and obvious limitations of explicit features. By mapping the sequence of local temporal correlation feature vectors of gas components to the latent space, a more extensive semantic representation is generated, where important information is efficiently compressed and redundant information is eliminated.

[0037] Figure 7 It is a block diagram of constructing a secondary subunit according to the topological features of gas components in the methane recovery and purification system of the intelligent dredging device according to an embodiment of the present application. As Figure 7 shown, in the embodiment of the present application, the secondary subunit 1322-2 for constructing the topological features of gas components includes: a tertiary subunit 1322-21 for calculating semantic causal correlation factors between any two local temporal correlation depth implicit feature encoding vectors in the sequence of local temporal correlation depth implicit feature encoding vectors of gas components to obtain a local temporal semantic causal correlation topological matrix composed of multiple local temporal semantic causal correlation factors of gas components; a tertiary subunit 1322-22 for gating causal trigger activation on the local temporal semantic causal correlation topological matrix to obtain the local temporal semantic causal correlation topological feature matrix.

[0038] In an embodiment of the present application, the gas component topological semantic causal factor calculation three - level subunit 1322 - 21 is configured to: calculate the correlation matrix between any two gas component local temporal correlation depth implicit feature encoding vectors in the sequence of gas component local temporal correlation depth implicit feature encoding vectors to obtain a sequence of gas component local temporal correlation matrices; calculate the semantic causal correlation factors of each gas component local temporal correlation matrix in the sequence of gas component local temporal correlation matrices to obtain the gas component local temporal semantic causal correlation topological matrix composed of multiple gas component local temporal semantic causal correlation factors, where the gas component local temporal semantic causal correlation factor is related to the mean, variance, maximum value of its corresponding gas component local temporal correlation matrix, and the gas component causal correlation bias value; wherein, in response to the variance of the gas component local temporal correlation matrix being greater than or equal to a predetermined threshold, the weighted average of the distances between any two gas component local temporal correlation depth implicit feature encoding vectors in the sequence of gas component local temporal correlation depth implicit feature encoding vectors is used as the gas component causal correlation bias value; in response to the variance of the gas component local temporal correlation matrix being less than the predetermined threshold, the weighted mean of the gas component local temporal correlation matrix is used as the gas component causal correlation bias value.

[0039] Specifically, the gas component topological semantic causal factor calculation three - level subunit 1322 - 21 is configured to calculate the semantic causal correlation factors between any two gas component local temporal correlation depth implicit feature encoding vectors in the sequence of gas component local temporal correlation depth implicit feature encoding vectors to obtain the gas component local temporal semantic causal correlation topological matrix composed of multiple gas component local temporal semantic causal correlation factors, which is expressed by the gas component topological semantic causal factor calculation formula as: ; Wherein, is matrix multiplication, is the transposed vector of, is and the gas component local temporal correlation matrix between, is the variance of, is to take the maximum value in, is the mean of, is the gas component causal correlation bias value, is the corresponding gas component local temporal semantic causal correlation factor, is and the distance between, is the number of vectors in, is a predetermined threshold, and are weighted hyperparameters, and are respectively the gas component local temporal semantic causal association factors at each position in the gas component local temporal semantic causal association topology matrix, is the gas component local temporal semantic causal association topology matrix. It should be understood that after the implicit feature mining is completed, the model enters the semantic causal association calculation stage. The goal of this step is to deduce the causal association between pairwise features from the sequence of gas component local temporal association deep implicit feature encoding vectors. In a specific implementation, this process can be implemented in combination with causal modeling techniques. In particular, in the technical solution of the present application, it constructs the gas component local temporal association matrix between any gas component local temporal association deep implicit feature encoding vectors in the sequence of the gas component local temporal association deep implicit feature encoding vectors, and uses the causal association energy metric function to explicitly quantify and encode the causal association between any two gas component local temporal association deep implicit feature encoding vectors in the sequence of the gas component local temporal association deep implicit feature encoding vectors to obtain the gas component local temporal semantic causal association factors.

[0040] In particular, by regarding the low-level causal associations in a complex system as molecular-level relationships inferred based on statistical correlations, causal association energy intervention prediction can be further performed on the basis of the global fine-grained statistical association representation, so as to study the causal association fine-grained structure and its dynamic regulation based on the high-dimensional and heterogeneous representation of causal relationship omics. Among them, when the aggregative distribution representation of the causal graph is greater than the predetermined threshold, there is a bias in the source data integration based on the matrix node effect representation of the gas component local temporal semantic causal association factor, and when the aggregative distribution representation of the causal graph is less than the predetermined threshold, the condensed structure modeling can be directly performed through the feature pattern integration and compression. In this way, not only can the causal association energy in the system be encoded and described, but also the implicit causal intervention prediction results can be condensed, so as to obtain a more efficient revelation of the key causal associations.

[0041] Specifically, the gas component gating causal trigger activation three-level subunit 1322-22 is used to perform gating causal trigger activation on the gas component local temporal semantic causal association topology matrix to obtain the gas component local temporal semantic causal association topology feature matrix, which is represented by the gas component gating causal trigger activation formula as: ; where is a non-linear activation function, is the normalization threshold, For to perform gated activation processing, is the topological feature matrix of local temporal semantic causal associations of gas components. It should be understood that after arranging the local temporal semantic causal association factors of gas components into the local temporal semantic causal association topological matrix of gas components, a causal trigger network is used to perform dynamic modeling on the local temporal semantic causal association topological matrix of gas components. The role of the causal trigger network is to perform more detailed feature extraction and modeling on the causal topological relationship. Its core is the dynamic gating mechanism and the non-linear activation function to perform dynamic causal triggering on the representation of the local temporal semantic causal association topological matrix of gas components. Specifically, in practical applications, causal relationships are often not static but dynamically changing. The gating mechanism allows the model to identify key causal paths in a dynamic context, thereby strengthening important associations and weakening noise interference. At the same time, the activation function in the network enhances the model's expressive power by introducing non-linearity and captures higher-order regularities hidden in complex causal structures. Through gated causal trigger activation, the weights of causal paths can be dynamically adjusted to ensure that the model always focuses on the most important causal relationships, thereby improving the robustness and generalization ability of the system.

[0042] In the embodiment of the present application, the gas component dynamic temporal fusion secondary subunit 1322-3 is configured to: perform dynamic random walks of feature sequences of graph convolution on the sequences of the local temporal semantic causal association topological feature matrix of gas components and the local temporal association feature vectors of gas components to obtain local temporal surface context dynamic random walk semantic encoding vectors of gas components; perform dynamic random walks of feature sequences of graph convolution on the sequences of the local temporal semantic causal association topological feature matrix of gas components and the local temporal association depth implicit feature encoding vectors of gas components to obtain local temporal hidden layer context dynamic random walk semantic encoding vectors of gas components; fuse the local temporal surface context dynamic random walk semantic encoding vectors of gas components and the local temporal hidden layer context dynamic random walk semantic encoding vectors of gas components to obtain the gas component dynamic temporal context aggregation feature vectors.

[0043] Specifically, performing dynamic random walks of feature sequences of graph convolution on the sequences of the local temporal semantic causal association topological feature matrix of gas components and the local temporal association feature vectors of gas components to obtain local temporal surface context dynamic random walk semantic encoding vectors of gas components, which can be expressed by the local temporal surface formula as: ; where is graph convolution processing, It is the semantic encoding vector of the local temporal surface context dynamic walk of gas components. It should be understood that after the causal trigger network deeply optimizes the local temporal semantic causal association topological structure of gas components, the feature sequence dynamic walk of graph convolution generates the semantic encoding vector of the local temporal surface context dynamic walk of gas components and the semantic encoding vector of the local temporal hidden layer context dynamic walk of gas components hierarchically through dynamic graph representation learning based on the graph convolutional neural network (GCN). Specifically, to generate the semantic encoding vector of the local temporal surface context dynamic walk of gas components, the dynamic walk mechanism simulates the propagation of features in the topological structure, from local to global, and recursively aggregates the explicit semantics between nodes.

[0044] Specifically, perform the feature sequence dynamic walk of the graph convolution on the sequence of the local temporal semantic causal association topological feature matrix of the gas components and the local temporal association depth implicit feature encoding vector of the gas components to obtain the semantic encoding vector of the local temporal hidden layer context dynamic walk of the gas components, which is expressed by the local temporal hidden layer formula of the gas components as: ; where is the semantic encoding vector of the local temporal hidden layer context dynamic walk of the gas components. It should be understood that for the semantic encoding vector of the local temporal surface context dynamic walk of the gas components, the dynamic walk mechanism conducts a deeper semantic exploration for the implicit features. The hidden layer semantics enables the potential embedding of features. Therefore, it is necessary to pay attention to the multi-hop propagation of high-order information and the distributed decoupling of deep features during the walk process to avoid the over-smoothing phenomenon caused by the propagation of deep topological features. This hidden layer representation provides greater generalization ability for feature expression by capturing far-reaching temporal dependencies and complex semantic patterns. Specifically, through graph convolution operations, features can be propagated and aggregated in the topological structure to generate deeper feature representations. This feature representation not only contains the information of the original causal relationship but also considers the global information in the topological structure, thus providing a more comprehensive and accurate feature representation. The dynamic walk mechanism allows the model to perform multi-hop propagation in the topological structure and gradually capture high-order dependency relationships. This is particularly important for capturing complex causal relationships because many causal relationships are not simple binary relationships but involve interactions between multiple nodes. Through the dynamic walk mechanism, these complex causal relationships can be gradually revealed, thereby improving the robustness and generalization ability of the system.

[0045] Specifically, fuse the semantic encoding vector of the local temporal surface context dynamic walk of the gas components and the semantic encoding vector of the local temporal hidden layer context dynamic walk of the gas components to obtain the dynamic temporal context aggregation feature vector of the gas components, which is expressed by the dynamic temporal fusion formula of the gas components as: ; where is the fusion weighting parameter, It is the dynamic time - series context aggregation feature vector of the gas components. It should be understood that the semantic representations of the gas component local time - series surface context dynamic walk semantic encoding vector and the gas component local time - series hidden - layer context dynamic walk semantic encoding vector respectively capture explicit and implicit semantic information. Among them, the surface context dynamic walk semantic encoding vector mainly reflects the interaction and local features between direct neighbor nodes, while the hidden - layer context dynamic walk semantic encoding vector captures deeper global features and high - order dependencies. Considering the complementarity of these two feature dimensions, where the surface features can capture local details and immediate changes, while the hidden - layer features can reveal global patterns and long - term trends. By fusing these two features, the complex relationships between gas components can be better understood, so feature fusion is required in the final stage. In specific implementation, strategies such as weighted summation, attention mechanism, or multi - head fusion can be used to dynamically adjust the importance distribution of the gas component local time - series surface context dynamic walk semantic encoding vector and the gas component local time - series hidden - layer context dynamic walk semantic encoding vector during the fusion process to ensure that the comprehensive features have stronger discriminative ability and representational integrity. Specifically, in weighted summation, a weight coefficient is assigned to each feature vector, and these weights determine the importance of each feature vector in the final comprehensive feature vector. When processing the gas component data in the methane recovery system, a higher weight can be given to the surface features to capture immediate changes, while a certain weight is given to the hidden - layer features to reflect long - term trends. The attention mechanism determines the influence degree of each feature vector on the final result by calculating the attention weights of each feature vector. Specifically, the attention mechanism first generates query (Query), key (Key), and value (Value) matrices according to the current context, and then generates attention weights by calculating the similarity between the query and the key. These weights reflect the relative importance of different feature vectors in a specific context and are used for weighted summation to generate the final comprehensive feature vector. When processing gas component data, the attention mechanism can automatically identify which features are most critical for the current task, thus improving the performance of the model. And multi - head fusion captures information in different dimensions through multiple parallel attention heads. Among them, each attention head operates independently, focusing on different feature sub - spaces or context information, so that the model can comprehensively understand the data from multiple perspectives. Multi - head fusion can capture local details and global patterns simultaneously, enhancing the expressive ability and generalization ability of the model. Among them, one attention head may focus on short - term fluctuations, and another attention head focuses on long - term trends. The combination of multiple attention heads can provide a richer and more diverse feature representation.

[0046] Specifically, the separation technology parameter aggregation subunit 1323 is configured to perform feature aggregation on the gas component dynamic time-series context aggregation feature vector, the physical property dynamic time-series context aggregation feature vector, and the process parameter dynamic time-series context aggregation feature vector to obtain the separation technology parameter multi-dimensional integration vector. It should be understood that the gas component dynamic time-series context aggregation feature vector, the physical property dynamic time-series context aggregation feature vector, and the process parameter dynamic time-series context aggregation feature vector are closely related to the selection of separation technologies. Specifically, the gas component dynamic time-series context aggregation feature provides basic information about the "quality" of the gas for the selection of separation technologies, helping to determine the technologies that can effectively separate target components and remove impurities, and is one of the core bases for the selection of separation technologies. The physical property dynamic time-series context aggregation feature provides a basis for the operating conditions and equipment selection of separation technologies, helping to judge which technology can operate efficiently under the current physical property conditions, and is an important factor in ensuring the feasibility and economy of separation technologies. The process parameter dynamic time-series context aggregation feature affects the final selection of separation technologies from the perspectives of the feasibility and economy of technology implementation, ensuring that the selected technology can maximize cost-effectiveness while meeting production requirements. Based on this, in order to comprehensively consider multiple factors and provide a richer and more comprehensive basis for the selection and parameter adjustment of separation technologies, the present application performs feature aggregation on the gas component dynamic time-series context aggregation feature vector, the physical property dynamic time-series context aggregation feature vector, and the process parameter dynamic time-series context aggregation feature vector to obtain the separation technology parameter multi-dimensional integration vector. In particular, in a specific example of the present application, the separation technology parameter multi-dimensional integration vector can be obtained by performing cascade processing on the gas component dynamic time-series context aggregation feature vector, the physical property dynamic time-series context aggregation feature vector, and the process parameter dynamic time-series context aggregation feature vector.

[0047] In an embodiment of the present application, the operation parameter adjustment unit 133 is configured to: input the multi-dimensional integrated vector of the separation technology parameters into a classifier-based selector to obtain the separation technology recommendation result, where the separation technology recommendation result includes membrane separation technology or pressure swing adsorption technology; based on the separation technology recommendation result, dynamically adjust the operation parameters during the separation process. That is, perform classification processing on the multi-dimensional integrated vector of the separation technology parameters obtained by aggregating features using the gas component dynamic time-series context aggregation feature vector, the physical property dynamic time-series context aggregation feature vector, and the process parameter dynamic time-series context aggregation feature vector to intelligently recommend the separation technology recommendation result. It should be understood that the classifier-based selector can analyze and process this integrated information. According to existing knowledge and models, using the discrimination rules and pattern recognition capabilities trained by it, map the multi-dimensional information to the appropriate separation technology category to more accurately determine which technology is more suitable under the current working conditions. For example, when the impurity content in the gas component is low and the requirement for separation efficiency is high, the classifier may recommend membrane separation technology; while when the gas treatment volume is large and the requirement for methane purity is not extremely high, pressure swing adsorption technology may be recommended, thereby improving the overall performance and efficiency of the system.

[0048] Subsequently, based on the separation technology recommendation result, dynamically adjust the operation parameters during the separation process. It should be understood that different separation technologies (such as membrane separation technology and pressure swing adsorption technology) have their own unique working principles and optimal operating conditions. For example, the separation effect of membrane separation technology is greatly affected by factors such as the membrane material, pressure difference, and temperature; pressure swing adsorption technology is sensitive to operation parameters such as adsorption pressure, adsorption time, and desorption conditions. Dynamically adjusting the operation parameters based on the separation technology recommendation result is to enable the system to adapt to the characteristics of the selected technology, give full play to the advantages of the technology, and ensure the efficient progress of the separation process. That is to say, reasonably adjusting the operation parameters according to the separation technology recommendation result can make the separation process more efficient. For example, for membrane separation technology, adjusting the appropriate pressure difference and temperature can increase the membrane permeation flux and improve the separation efficiency of methane; for pressure swing adsorption technology, optimizing the adsorption time and desorption conditions can improve the utilization rate of the adsorbent, thereby improving the product purity.

[0049] In summary, the recovery system in the methane recovery and purification system of the intelligent dredging device based on the embodiments of the present application is elucidated. The time series of gas components (methane concentration, carbon dioxide ratio, and nitrogen ratio), the time series of physical properties (temperature, pressure, and humidity), and the time series of process parameters (processing volume and recovery rate) of the preliminarily purified gas are obtained through sensors. The time series of the gas components, the time series of the physical properties, and the time series of the process parameters are respectively locally encoded in time series using an artificial intelligence-based data analysis and feature extraction method. Then, causal context time series aggregation is performed on the sequences of the locally encoded time series features of each parameter. Based on the multi-dimensional representation between the aggregated dynamic time series context aggregation features of the gas components, physical properties, and process parameters, the separation technology recommendation results are intelligently recommended, and the operating parameters in the separation process are dynamically adjusted based on this result. The present application abandons subjective experience judgment, can monitor the changes in gas components, physical properties, and process parameters in real time, and can intelligently recommend the most suitable separation technology for the current working conditions based on these changes, greatly improving the accuracy and objectivity of the selection to more precisely match the gas characteristics and process requirements. At the same time, dynamically adjusting the operating parameters ensures that the separation process is always in the optimal state.

[0050] Figure 8 It is a flowchart of the purification module in the methane recovery and purification system of the intelligent dredging device according to the embodiments of the present application. As Figure 8As shown, in the methane recovery and purification system 100 of the above intelligent cleaning equipment, the purification module 140 is used to perform multi-stage purification treatment on the separated methane to obtain purified methanol. It should be understood that during the methane recovery process, although methane can be preliminarily purified through various separation technologies (such as membrane separation, adsorption separation, etc.), the separated methane may still contain a certain amount of impurities, such as carbon dioxide, hydrogen sulfide, water vapor, and other trace gas components. These impurities not only affect the purity of methane and reduce its value as a fuel or chemical raw material, but may also cause corrosion or blockage problems to subsequent use equipment. Therefore, in order to ensure the quality of methane and meet the requirements of different application scenarios, it is necessary to perform further purification treatment on it. The core of multi-stage purification treatment lies in gradually removing various impurities in methane to ensure that the final product meets the required purity standard. Specifically, in the technical solution of this application, the purification module performs in-depth purification treatment on the recovered methane to further improve its purity. First, the separated methane passes through a primary purification device, and methods such as chemical adsorption (such as activated carbon, alumina, etc.) or catalytic oxidation are used to remove impurities such as organic sulfur and inorganic sulfur in methane. Then, through a secondary purification device, carbon monoxide, carbon dioxide and other gas impurities in methane are further removed through catalytic oxidation reactions to purify methane. Next, through a tertiary purification device, a precision filter (such as a polytetrafluoroethylene membrane filter) is used to further remove tiny particles and residual impurities in the gas, ensuring that the purity of the purified methanol reaches more than 99%. Finally, the purified methanol is output through the discharge port.

[0051] In summary, the methane recovery and purification system 100 in the intelligent cleaning equipment based on the embodiments of this application is elucidated. It collects methane-containing gas from the cleaning equipment, removes moisture and tiny particles in the gas to obtain a preliminarily purified gas, then uses separation technology to separate and recover methane, and obtains purified methanol through multi-stage purification treatment. Among them, the methane recovery technology needs to obtain the time series of gas components, the time series of physical properties, and the time series of process parameters, and perform local time series encoding and causal context time series aggregation to recommend the separation technology recommendation result and dynamically adjust the operation parameters. This can abandon subjective experience, monitor in real time and accurately match the gas characteristics with the process requirements, improve the accuracy and objectivity of the separation technology selection, and ensure that the separation process is always in the optimal state. In this way, the value of methane as a fuel or chemical raw material can be improved, equipment corrosion and blockage problems can be avoided, the stable operation of the system can be guaranteed, the requirements of different application scenarios can be met, and the resource utilization efficiency and economic benefits can be improved. Finally, high-purity methanol is output through the discharge port, ensuring its reliability and quality in subsequent applications.

[0052] As described above, the methane recovery and purification system 100 in the intelligent cleaning and dredging device according to the embodiments of the present application can be implemented in various terminal devices. In one example, the methane recovery and purification system 100 in the intelligent cleaning and dredging device can be integrated into the terminal device as a software module and / or a hardware module. For example, the methane recovery and purification system 100 in the intelligent cleaning and dredging device can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the methane recovery and purification system 100 in the intelligent cleaning and dredging device can also be one of the many hardware modules of the terminal device.

[0053] Alternatively, in another example, the methane recovery and purification system 100 in the intelligent cleaning and dredging device and the terminal device can also be separate devices, and the methane recovery and purification system 100 in the intelligent cleaning and dredging device can be connected to the terminal device through a wired and / or wireless network, and transmit and interact information in accordance with a predefined data format.

[0054] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0055] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0056] In addition, in each embodiment of the present invention, the functional modules can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0057] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms.

[0058] Therefore, in any aspect, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Thus, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced by the present invention. Any reference signs in the claims should not be construed as limiting the claims concerned.

[0059] In addition, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of elements or devices recited in the system claims may also be implemented by one element or device through software or hardware. The terms such as "second" are used to denote names and do not denote any particular order.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit of the technical solutions of the present invention.

Claims

1. A methane recovery and purification system in an intelligent cleaning device, characterized in that: include: A methane collection module, used to collect methane-containing gas from the cleaning equipment; A pre-processing module, used for performing preliminary purification treatment on the collected methane-containing gas to remove moisture and tiny particles in the methane-containing gas to obtain pre-purified gas; A recovery module, used for separating and treating the preliminarily purified gas by separation technology to recover the separated methane, wherein the separation technology is one or a combination of membrane separation technology or pressure swing adsorption technology; The purification module is used to perform multi-stage purification treatment on the separated methane to obtain purified methanol.

2. The methane recovery and purification system in the intelligent cleaning equipment according to claim 1 is characterized in that: The recycling module comprises: a preliminary purified gas data acquisition unit, used to acquire a time series of gas components, a time series of physical properties, and a time series of process parameters of the gas after the preliminary purification through a sensor, wherein the gas components include a methane concentration, a carbon dioxide ratio, and a nitrogen ratio, the physical properties include temperature, pressure, and humidity, and the process parameters include a processing volume and a recovery rate; A separation parameter multi-dimensional integration unit, used for performing dynamic separation parameter multi-dimensional integration based on time series characteristics on the time series of the gas components, the time series of the physical properties and the time series of the process parameters to obtain a separation technical parameter multi-dimensional integration vector; The operating parameter adjustment unit is used to obtain a separation technology recommendation result based on the multi-dimensional integrated vector of the separation technology parameters, and dynamically adjust the operating parameters in the separation process based on the separation technology recommendation result.

3. The methane recovery and purification system in the intelligent cleaning equipment according to claim 2 is characterized in that: The separation parameter multi-dimensional integration unit comprises: A data local time series encoding subunit, used for respectively performing local time series encoding on the time series of the gas components, the time series of the physical properties and the time series of the process parameters to obtain a sequence of local time series associated feature vectors of the gas components, a sequence of local time series associated feature vectors of the physical properties and a sequence of local time series associated feature vectors of the process parameters; A temporal causal context aggregation subunit, used for performing dynamic separation parameter temporal causal context aggregation on the sequence of the gas component local temporal association feature vectors, the sequence of the physical property local temporal association feature vectors and the sequence of the process parameter local temporal association feature vectors respectively to obtain a gas component dynamic temporal context aggregation feature vector, a physical property dynamic temporal context aggregation feature vector and a process parameter dynamic temporal context aggregation feature vector; The separation technology parameter aggregation subunit is used to perform feature aggregation on the gas component dynamic time series context aggregation feature vector, the physical property dynamic time series context aggregation feature vector and the process parameter dynamic time series context aggregation feature vector to obtain the separation technology parameter multi-dimensional integrated vector.

4. The methane recovery and purification system in the intelligent cleaning equipment according to claim 3 is characterized in that: The data local time series encoding subunit is used to: input the time series of the gas components, the time series of the physical properties and the time series of the process parameters into a local time series encoder based on one-dimensional convolution respectively to obtain a sequence of local time series associated feature vectors of the gas components, a sequence of local time series associated feature vectors of the physical properties and a sequence of local time series associated feature vectors of the process parameters.

5. The methane recovery and purification system in the intelligent cleaning equipment according to claim 4 is characterized in that: The temporal causal context aggregation subunit includes: A gas component implicit feature mining secondary subunit is used to perform implicit feature mining on each gas component local time series correlation feature vector in the sequence of gas component local time series correlation feature vectors to obtain a sequence of gas component local time series correlation deep implicit feature coding vectors; A gas component topological feature construction secondary subunit is used to construct the semantic causal association topological features of the sequence of the gas component local temporal association deep implicit feature encoding vectors to obtain a gas component local temporal semantic causal association topological feature matrix; The gas component dynamic temporal fusion secondary subunit is used to dynamically fuse the feature sequences of the sequence of the gas component local temporal association deep implicit feature coding vectors and the sequence of the gas component local temporal association deep implicit feature coding vectors by graph convolution based on the gas component local temporal semantic causal association topological feature matrix to obtain the gas component dynamic temporal context aggregation feature vector.

6. The methane recovery and purification system in the intelligent cleaning equipment according to claim 5 is characterized in that: The gas composition topological characteristics construct a secondary subunit, including: The gas component extension semantic causal factor calculation tertiary subunit is used to calculate the semantic causal association factor between any two gas component local temporal association depth implicit feature coding vectors in the sequence of the gas component local temporal association depth implicit feature coding vectors to obtain a gas component local temporal semantic causal association topological matrix composed of multiple gas component local temporal semantic causal association factors; The gas component gated causal trigger activation three-level sub-unit is used to perform gated causal trigger activation on the gas component local temporal semantic causal association topological matrix to obtain the gas component local temporal semantic causal association topological feature matrix.

7. The methane recovery and purification system in the intelligent cleaning equipment according to claim 6 is characterized in that: The gas composition extension semantic causal factor calculation three-level subunit is used to: Calculating the correlation matrix between any two gas component local temporal correlation depth implicit feature coding vectors in the sequence of gas component local temporal correlation depth implicit feature coding vectors to obtain a sequence of gas component local temporal correlation matrices; Calculating the semantic causal association factor of each gas component local temporal association matrix in the sequence of the gas component local temporal association matrices to obtain the gas component local temporal semantic causal association topological matrix composed of a plurality of gas component local temporal semantic causal association factors, wherein the gas component local temporal semantic causal association factor is related to the mean, variance, maximum value and gas component causal association bias value of the corresponding gas component local temporal association matrix; Wherein, in response to the variance of the gas component local temporal correlation matrix being greater than or equal to a predetermined threshold, a weighted average of the distances between any two gas component local temporal correlation depth implicit feature coding vectors in the sequence of the gas component local temporal correlation depth implicit feature coding vectors is used as the gas component causal correlation bias value; In response to the variance of the gas component local time series correlation matrix being smaller than the predetermined threshold, a weighted mean of the gas component local time series correlation matrix is ​​used as the gas component causal correlation bias value.

8. The methane recovery and purification system in the intelligent cleaning equipment according to claim 7 is characterized in that: The gas composition dynamic time series fusion secondary subunit is used for: Performing a feature sequence dynamic walk of graph convolution on the sequence of the gas component local temporal semantic causal association topological feature matrix and the gas component local temporal association feature vector to obtain a gas component local temporal surface context dynamic walk semantic encoding vector; Performing the feature sequence dynamic walk of the graph convolution on the sequence of the gas component local temporal semantic causal association topological feature matrix and the gas component local temporal association deep implicit feature coding vector to obtain the gas component local temporal hidden layer context dynamic walk semantic coding vector; The gas component local temporal surface context dynamic wandering semantic coding vector and the gas component local temporal hidden context dynamic wandering semantic coding vector are fused to obtain the gas component dynamic temporal context aggregation feature vector.

9. The methane recovery and purification system in the intelligent cleaning equipment according to claim 8 is characterized in that: The operating parameter adjustment unit is used to: Inputting the multidimensional integrated vector of separation technology parameters into a classifier-based selector to obtain the separation technology recommendation result, wherein the separation technology recommendation result includes membrane separation technology or pressure swing adsorption technology; Based on the separation technology recommendation results, the operating parameters of the separation process are dynamically adjusted.

Images