Comprehensive utilization device and method for raw gas

Abstract

The invention relates to the field of intelligent control, and provides a comprehensive utilization device and method for raw gas, and the method comprises the steps: collecting time queue data of a real-time temperature value and a real-time flow value of a second stream of converted gas of a water cooler through a temperature sensor and a flow sensor which are installed at the front end of the water cooler; and carrying out time sequence correlation analysis on the real-time temperature value and the real-time flow value by adopting a data processing technology based on deep learning, so as to automatically recommend a refrigerant flow velocity value according to collaborative significant interaction representation between real-time temperature time sequence correlation characteristics and real-time flow time sequence correlation characteristics. In this way, the flow speed of the refrigerant can be dynamically adjusted, and it is ensured that the optimal heat recovery effect can be achieved under different working conditions. And temperature and flow fluctuation caused by untimely manual adjustment can be reduced based on real-time data, so that the heat recovery process is more stable, and automatic control over the flow speed of the refrigerant is achieved.

Description

Technical Field

[0001] The present application relates to the field of intelligent control, and more specifically, to a device and method for comprehensive utilization of raw coal gas. Background Art

[0002] Raw coal gas usually refers to untreated raw coal gas, which comes from the coal gasification process in coal chemical industry, metallurgy and other industries. Raw coal gas contains many components, such as hydrogen (H 2 ), carbon monoxide (CO), carbon dioxide (CO 2 ), methane (CH 4 ) and a small amount of sulfide and other impurities. These components have high calorific value and chemical activity, so their comprehensive utilization is of great significance. The comprehensive utilization of waste gas can not only improve the utilization efficiency of resources and reduce environmental pollution, but also create additional economic benefits for enterprises.

[0003] In the process of treating raw coal gas, heat recovery is an important link. By utilizing low-pressure waste heat boilers, desalted water preheaters, water coolers and other equipment, the heat in the gas can be recovered, energy consumption can be reduced, and the energy efficiency of the entire process can be improved. In the process of heat recovery, the control of the refrigerant flow rate is particularly important. A reasonable refrigerant flow rate can not only ensure the maximum recovery of heat from the exhaust gas, but also optimize the heat transfer efficiency and maintain the thermal stability of the system.

[0004] However, traditional refrigerant flow rate control uses a fixed flow rate setting, which cannot be adjusted dynamically according to actual working conditions, resulting in low efficiency. This static setting cannot optimize the refrigerant flow rate in real time when facing changing working conditions, thus affecting the heat recovery efficiency. In addition, traditional methods are highly dependent on the experience and technical level of operators, lacking systematicity and scientificity. In particular, the operator's experience may be biased, and it is difficult to make accurate judgments quickly in complex or sudden situations, resulting in inaccurate control, making it difficult to ensure long-term stable and efficient operation of the system, especially in industrial environments with complex and changing working conditions.

[0005] Therefore, an optimized comprehensive utilization scheme of waste gas is desired. Summary of the invention

[0006] In view of the shortcomings in the prior art, the present application provides a device and method for comprehensive utilization of waste gas, which collects the time queue data of the real-time temperature value and the real-time flow value of the second stream of converted gas in the water cooler through the temperature sensor and flow sensor installed at the front end of the water cooler, and uses the data processing technology based on deep learning to perform time series correlation analysis on the real-time temperature value and the real-time flow value, so as to automatically recommend the refrigerant flow rate value according to the synergistic significant interaction expression between the real-time temperature time series correlation feature and the real-time flow time series correlation feature. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on the real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process more stable, thereby realizing the automatic control of the refrigerant flow rate.

[0007] According to one aspect of the present application, a method for comprehensive utilization of raw coal gas is provided, which comprises:

[0008] S1: compressing the raw raw gas to obtain compressed raw gas;

[0009] S2: converting the compressed raw coal gas to obtain converted gas;

[0010] S3: splitting the converted gas to obtain a first stream of converted gas and a second stream of converted gas;

[0011] S4: performing shift treatment, acid removal purification and pressure swing adsorption on the first converted gas to obtain hydrogen;

[0012] S5: The second converted gas is subjected to heat recovery, acid removal and pressure swing adsorption to obtain carbon monoxide.

[0013] According to another aspect of the present application, a device for comprehensive utilization of raw coal gas is provided, comprising:

[0014] A raw gas compression processing module is used to compress the raw raw gas introduced to obtain compressed raw gas;

[0015] A compressed raw coal gas conversion processing module, used for converting the compressed raw coal gas to obtain converted gas;

[0016] A gas splitting processing module, used for splitting the converted gas to obtain a first stream of converted gas and a second stream of converted gas;

[0017] A hydrogen preparation module, used for performing conversion treatment, acid removal purification and pressure swing adsorption on the first converted gas to obtain hydrogen;

[0018] The carbon monoxide preparation module is used to perform heat recovery, acid removal and pressure swing adsorption on the second converted gas to obtain carbon monoxide.

[0019] This application has significant technical effects due to the adoption of the above technical solutions:

[0020] The device and method for comprehensive utilization of waste gas provided by the present application collects the time queue data of the real-time temperature value and the real-time flow value of the second stream of converted gas of the water cooler through the temperature sensor and the flow sensor installed at the front end of the water cooler, and uses the data processing technology based on deep learning to perform time series correlation analysis on the real-time temperature value and the real-time flow value, so as to automatically recommend the refrigerant flow rate value according to the synergistic significant interaction expression between the real-time temperature time series correlation feature and the real-time flow time series correlation feature. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on the real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process more stable, thereby realizing the automatic control of the refrigerant flow rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying 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 of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0022] Figure 1 It is a flow chart of the method for comprehensive utilization of raw coal gas according to an embodiment of the present application.

[0023] Figure 2 The present invention is a flow chart of the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, in which heat recovery, acid removal and pressure swing adsorption are performed on the second converted gas to obtain carbon monoxide.

[0024] Figure 3 This is a flow chart of the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, in which the second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain the cooled second converted gas.

[0025] Figure 4 In the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, the data flow diagram of the second converted and treated gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain the cooled second converted and treated gas.

[0026] Figure 5In the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, multi-feature collaborative interactive fusion is performed on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain a flowchart representing the significant interactive fusion of heat flow temperature and flow.

[0027] Figure 6 It is a block diagram of a device for comprehensive utilization of raw coal gas according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0029] Raw coal gas, that is, raw coal gas without purification, mainly comes from the coal gasification process and is commonly used in industries such as coal chemical industry and metallurgy. This gas contains hydrogen, carbon monoxide, carbon dioxide, methane, and trace amounts of sulfides and other impurities. Due to its high calorific value and chemical activity, the comprehensive utilization of raw coal gas can not only improve the utilization efficiency of resources and reduce environmental pollution, but also bring additional economic benefits to enterprises.

[0030] In the process of treating raw coal gas, heat recovery is a key step. By using low-pressure waste heat boilers, desalted water preheaters and water coolers, heat can be effectively recovered from the gas, energy consumption can be reduced, and the energy efficiency of the entire process can be improved. In this process, it is crucial to control the flow rate of the refrigerant, because a reasonable flow rate can not only maximize the recovery of heat from the exhaust gas, but also improve the heat transfer efficiency and maintain the thermal stability of the system.

[0031] However, the traditional refrigerant flow rate control method uses a fixed flow rate and lacks adaptability to changes in actual operating conditions, resulting in low efficiency. This static setting cannot achieve real-time optimization of the refrigerant flow rate when facing changes in operating conditions, thus affecting the effect of heat recovery. In addition, the traditional method relies heavily on the experience and technical level of the operator and lacks systematic and scientific support. The operator's experience may be biased, and it is difficult to make accurate judgments quickly in complex or emergency situations, which limits the accuracy of the control and makes it difficult to ensure the long-term stability and efficient operation of the system, especially in a changing industrial environment.

[0032] Based on this, the present application proposes a method for comprehensive utilization of waste gas. Figure 1 FIG. 1 is a flow chart of a method for comprehensive utilization of raw coal gas according to an embodiment of the present application. Figure 1As shown, the method for comprehensive utilization of raw coal gas according to the embodiment of the present application includes: S1: compressing the raw raw coal gas introduced to obtain compressed raw coal gas; S2: converting the compressed raw coal gas to obtain converted gas; S3: splitting the converted gas to obtain a first converted gas and a second converted gas; S4: performing transformation treatment, acid depuration and pressure swing adsorption on the first converted gas to obtain hydrogen; S5: performing heat recovery, acid depuration and pressure swing adsorption on the second converted gas to obtain carbon monoxide.

[0033] In step S1, the raw raw gas is compressed to obtain compressed raw gas. It should be understood that the raw raw gas needs to be compressed to increase its pressure, so as to prepare for subsequent conversion treatment. After compression, the density of the raw gas is increased, which is conducive to increasing the rate of subsequent chemical reactions.

[0034] In step S2, the compressed raw gas is subjected to conversion treatment to obtain converted gas. Specifically, the conversion treatment is to convert the methane in the raw gas into synthesis gas (CO and H2 in this application) through a non-catalytic partial oxidation process, and at the same time, the organic sulfur, tar, HCN, benzene, naphthalene and other impurities in the raw gas are decomposed or converted through a high-temperature conversion reaction (carried out at 1200-1300°C) to avoid affecting the subsequent processes and product quality.

[0035] In step S3, the converted gas is subjected to a shunt treatment to obtain a first stream of converted gas and a second stream of converted gas. It should be understood that the converted gas usually contains multiple components, including hydrogen, carbon monoxide, carbon dioxide, etc. In order to facilitate subsequent specific treatment of different gas components, specifically hydrogen and carbon monoxide, the converted gas needs to be shunt treated. The shunt treatment allows each gas to undergo a specially designed treatment process, which is conducive to improving the recovery rate and purity of the gas.

[0036] In step S4, the first converted gas is subjected to conversion treatment, acid depuration and pressure swing adsorption to obtain hydrogen. It should be understood that the conversion treatment in the present application specifically refers to the conversion reaction of CO and H20 in the first converted gas to generate H2 and CO2 and release heat (Q) under the catalytic action of the catalyst and reaction conditions of certain temperature and pressure, and the reaction equation is as follows:

[0037]

[0038] Through the above-mentioned reactions, the CO in the first stream of converted gas is reduced to the required content, and the H2 in the first stream of converted gas is increased to the required content.

[0039] Accordingly, considering that the gas after the conversion treatment still contains acidic gases (such as CO2 and H2S), in order to ensure that the content of acidic gases in the hydrogen finally obtained is reduced to a minimum to meet the needs of industrial applications, it is necessary to perform heat recovery and acid removal purification on the first gas after the conversion treatment in the present application. In particular, in the embodiment of the present application, a low-temperature methanol washing method can be specifically used to achieve acid removal and purification of the raw gas. Next, in order to separate and extract hydrogen, the purified gas needs to be subjected to pressure swing adsorption. Pressure swing adsorption can separate hydrogen by utilizing the difference in adsorption capacity of different gases on solid adsorbents to separate hydrogen.

[0040] In step S5, the second converted gas is subjected to heat recovery, acid removal and pressure swing adsorption to obtain carbon monoxide. Specifically, Figure 2 The present invention is a flow chart of the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, wherein the second gas after conversion treatment is subjected to heat recovery, acid removal and pressure swing adsorption to obtain carbon monoxide. Figure 2 As shown, the second converted gas is subjected to heat recovery, acid removal and pressure swing adsorption to obtain carbon monoxide, including: S51: the second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain a cooled second converted gas; S52: the cooled second converted gas is subjected to the acid removal and pressure swing adsorption to obtain the carbon monoxide.

[0041] In step S51, the second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain a cooled second converted gas. It should be understood that the second converted gas obtained after the conversion is a high-temperature gas. In order to utilize the waste heat in the gas, the second converted gas needs to be sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler. By introducing the second converted gas into a low-pressure waste heat boiler, the heat in the gas can be used to heat the water in the boiler to generate steam or hot water, which can not only recover heat energy but also improve the overall energy efficiency of the comprehensive utilization of raw coal gas. Subsequently, the gas discharged from the low-pressure waste heat boiler continues to pass through the desalted water preheater, and the waste heat is further used to preheat the incoming water, thereby reducing the energy consumption of subsequent heating. Next, the gas passes through the water cooler to further cool the gas to a temperature suitable for subsequent acid removal and purification treatment, which helps to improve the efficiency of subsequent treatment.

[0042] Accordingly, in the case where the second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain the second converted gas after cooling, the technical concept of the present application is to install a temperature sensor and a flow sensor at the front end of the water cooler and collect the time queue of the real-time temperature value and the time queue of the real-time flow value of the second converted gas of the water cooler, and use a data processing technology based on deep learning to perform a time series correlation analysis on the real-time temperature value and the real-time flow value, so as to automatically recommend the refrigerant flow rate value according to the synergistic significant interaction representation between the real-time temperature time series correlation feature and the real-time flow time series correlation feature. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process smoother, thereby realizing the automatic control of the refrigerant flow rate.

[0043] Specifically, Figure 3 This is a flow chart of the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, in which the second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain the cooled second converted gas. Figure 4 The data flow diagram of the second stream of converted and treated gas is shown in the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, wherein the second stream of converted and treated gas passes through a low-pressure waste heat boiler, a desalted water preheater and a water cooler in sequence to obtain the cooled second stream of converted and treated gas. Figure 3 and Figure 4 As shown, the second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain a second converted gas after cooling, including: S511: installing a temperature sensor and a flow sensor at the front end of the water cooler; S512: collecting a time queue of real-time temperature values ​​and a time queue of real-time flow values ​​of the second converted gas flowing into the water cooler through the temperature sensor and the flow sensor; S513: performing sequence encoding on the time queue of the real-time temperature values ​​and the time queue of the real-time flow values ​​to obtain a real-time temperature time series associated feature vector and a real-time flow time series associated feature vector; S514: performing multi-feature collaborative interactive fusion on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain a heat flow temperature-flow significant interactive fusion representation; S515: based on the heat flow temperature-flow significant interactive fusion representation, obtaining a control instruction, the control instruction including a decoded value of a recommended refrigerant flow rate value.

[0044] In step S511 and step S512, a temperature sensor and a flow sensor are installed at the front end of the water cooler, and the time queue of the real-time temperature value and the time queue of the real-time flow value of the second stream of converted gas flowing into the water cooler are collected by the temperature sensor and the flow sensor. It should be understood that by installing a temperature sensor and a flow sensor at the front end of the water cooler, the temperature and flow of the gas can be monitored in real time. By using the sensor to obtain the time queue data of the real-time temperature value and the real-time flow value of the second stream of converted gas, and performing a time series analysis on it, the time-varying trend information of the real-time temperature and real-time flow of the gas can be captured, and the time series characteristics of the two parameters are correlated and fused, and finally the recommended value of the refrigerant flow rate is obtained based on the decoding of the correlated fusion characteristics. In this way, accurate control of the refrigerant flow rate can be achieved, which is conducive to a more efficient and stable heat recovery process.

[0045] In step S513, the time queue of the real-time temperature value and the time queue of the real-time flow value are serially encoded to obtain a real-time temperature time series associated feature vector and a real-time flow time series associated feature vector. Specifically, in an embodiment of the present application, the time queue of the real-time temperature value and the time queue of the real-time flow value are serially encoded to obtain a real-time temperature time series associated feature vector and a real-time flow time series associated feature vector, including: inputting the time queue of the real-time temperature value and the time queue of the real-time flow value into a sequence encoder based on a 1D-CNN model to obtain the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector. It should be understood that the real-time temperature value and the real-time flow value have different time series characteristics in different local time periods, such as continuous changes in a local time period and temperature fluctuations or flow change patterns in a short period of time. Based on this, in order to better understand and capture the local time series features in the time queue of the real-time temperature values ​​and the time queue of the real-time flow values, in the technical solution of the present application, the time queue of the real-time temperature values ​​and the time queue of the real-time flow values ​​are input into a sequence encoder based on the 1D-CNN model to utilize the convolution kernel to slide within the local time window to capture the local time series features in the data, and obtain the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector.

[0046] In step S514, the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are subjected to multi-feature collaborative interaction fusion to obtain a heat flow temperature-flow significant interaction fusion representation. Specifically, Figure 5 The present invention is a flowchart of the method for comprehensive utilization of raw coal gas according to an embodiment of the present application, in which the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are subjected to multi-feature collaborative interactive fusion to obtain a heat flow temperature-flow significant interactive fusion representation. Figure 5As shown, the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are subjected to multi-feature collaborative interactive fusion to obtain a heat flow temperature-flow significant interactive fusion representation, including: S5141: implicit feature combination of the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain temperature-flow conditional features; S5142: using the temperature-flow conditional features as conditional information, performing significant interaction based on semantic contribution on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain the heat flow temperature-flow significant interactive fusion representation.

[0047] It should be understandable that there is a complex temporal interaction relationship and a high-level intrinsic interaction relationship between the real-time temperature time series correlation feature vector and the real-time flow time series correlation feature vector. Therefore, it is possible to more accurately characterize and highlight the dependency between the change patterns of temperature and flow, and extract information that can reflect the significant interaction between the two. In the technical solution of the present application, the real-time temperature time series correlation feature vector and the real-time flow time series correlation feature vector are subjected to multi-feature collaborative interaction fusion to obtain a heat flow temperature-flow significant interaction fusion representation vector as a heat flow temperature-flow significant interaction fusion representation.

[0048] In detail, first, the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are implicitly combined to obtain the temperature-flow condition feature. That is, implicit feature combination can produce a higher-level feature representation, and these features may be more expressive than the original features, and can capture relationships that cannot be reflected by a single feature. In other words, they can better reflect the complex relationships within the system and provide richer information for subsequent analysis. Specifically, the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are jointly processed by using a joint implicit feature capture network to capture and extract the common time series pattern between temperature and flow, while highlighting the implicit information of the original single feature, and obtaining a real-time temperature-flow joint implicit feature vector. Then, Sigmoid is used to activate the features to further extract the implicit information, and appropriate weights are assigned to them to obtain the temperature-flow condition feature vector that is most critical for describing the comprehensive conditions.

[0049] Then, taking the temperature-flow conditional feature vector as the conditional information, the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are significantly interacted based on the semantic contribution to obtain the heat flow temperature-flow significant interaction fusion representation vector. That is, by performing significant interaction based on the semantic contribution, we can have a deeper understanding of the degree of influence of temperature and flow characteristics on the entire conditional information, so as to highlight those interactive information with significant influence. Specifically, the semantic contribution of the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector relative to the temperature-flow conditional feature vector is calculated respectively, so as to evaluate the correlation and contribution degree between the temperature feature and the flow feature to the conditional feature. Then, each semantic contribution is normalized to convert it to the same scale to ensure that the importance of different features can be fairly compared. After that, the original real-time temperature time series associated feature and the real-time flow time series associated feature are weighted modulated based on each normalized semantic contribution to emphasize the key part of the conditional vector, and the modulated real-time temperature time series associated feature vector and the modulated real-time flow time series associated feature vector are obtained. Finally, the real-time temperature time series associated feature vector after modulation is used as the query vector, the real-time flow time series associated feature vector after modulation is used as the key vector and the temperature-flow condition feature vector is used as the value vector. The feature significant guided interaction module based on the converter structure is input to promote the interaction and fusion between different features, so that the model can have a deeper understanding of the temporal interaction relationship between features, thereby adaptively focusing on the most critical features, and finally generating a heat flow temperature-flow significant interaction fusion representation vector reflecting the comprehensive interaction state.

[0050] Specifically, in an embodiment of the present application, an implicit feature combination is performed on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain a temperature-flow condition feature, including: inputting the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector into a joint implicit feature capture network to obtain a real-time temperature-flow joint implicit feature vector; and using a Sigmoid function to perform feature activation on the real-time temperature-flow joint implicit feature vector to obtain a temperature-flow condition feature vector as the temperature-flow condition feature.

[0051] More specifically, in an embodiment of the present application, the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are input into a joint implicit feature capture network to obtain a real-time temperature-flow joint implicit feature vector, including: adding the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector by position point, multiplying the obtained real-time temperature-flow sum vector with the weight matrix and then adding it by position with the bias vector to obtain a real-time temperature-flow joint interaction vector; using a tanh function to process the real-time temperature-flow joint interaction vector to obtain the real-time temperature-flow joint implicit feature vector.

[0052] Specifically, in an embodiment of the present application, taking the temperature-flow condition feature as conditional information, the real-time temperature time series association feature vector and the real-time flow time series association feature vector are significantly interacted based on semantic contribution to obtain the heat flow temperature-flow significant interaction fusion representation, including: calculating the real-time temperature semantic contribution of the real-time temperature time series association feature vector relative to the temperature-flow condition feature vector; calculating the real-time flow semantic contribution of the real-time flow time series association feature vector relative to the temperature-flow condition feature vector; normalizing the real-time temperature semantic contribution and the real-time flow semantic contribution, and using the normalized real-time temperature semantic contribution and the normalized real-time flow semantic contribution to The real-time temperature time series association feature vector and the real-time flow time series association feature vector are weighted modulated to obtain the modulated real-time temperature time series association feature vector and the modulated real-time flow time series association feature vector; using the modulated real-time temperature time series association feature vector as the query vector, the modulated real-time flow time series association feature vector as the key vector and the temperature-flow condition feature vector as the value vector, the modulated real-time temperature time series association feature vector, the modulated real-time flow time series association feature vector and the temperature-flow condition feature vector are input into the feature-to-feature significant guided interaction module based on the converter structure to obtain the heat flow temperature-flow significant interaction fusion representation vector as the heat flow temperature-flow significant interaction fusion representation.

[0053] More specifically, in an embodiment of the present application, the real-time temperature semantic contribution of the real-time temperature time series associated feature vector relative to the temperature-flow condition feature vector is calculated, including: calculating the positional division of the real-time temperature time series associated feature vector and the temperature-flow condition feature vector to obtain a real-time temperature semantic contribution vector; calculating the base-two logarithmic function value of the absolute value of each eigenvalue of the real-time temperature semantic contribution vector to obtain a real-time temperature semantic contribution logarithmic vector; calculating the positional dot product of the real-time temperature time series associated feature vector and the real-time temperature semantic contribution logarithmic vector, and performing position-by-position multiplication on the obtained dot product vector to obtain a real-time temperature semantic contribution value; calculating an exponential function with the natural constant e as the base and the real-time temperature semantic contribution value as the exponent to obtain the real-time temperature semantic contribution. Similarly, the processing method of the real-time flow semantic contribution is consistent with the processing method of the real-time temperature semantic contribution.

[0054] More specifically, in an embodiment of the present application, the real-time temperature semantic contribution and the real-time traffic semantic contribution are normalized, and the normalized real-time temperature semantic contribution and the normalized real-time traffic semantic contribution are used to perform weighted modulation on the real-time temperature time series association feature vector and the real-time traffic time series association feature vector to obtain a modulated real-time temperature time series association feature vector and a modulated real-time traffic time series association feature vector, including: calculating the sum of the real-time temperature semantic contribution and the real-time traffic semantic contribution to obtain a total sum of real-time temperature-traffic semantic contributions; dividing The real-time temperature semantic contribution and the real-time traffic semantic contribution are divided by the total sum of the real-time temperature-traffic semantic contributions to obtain the normalized real-time temperature semantic contribution and the normalized real-time traffic semantic contribution; the real-time temperature time series association feature vector is multiplied by the normalized real-time temperature semantic contribution to obtain the modulated real-time temperature time series association feature vector; the real-time traffic time series association feature vector is multiplied by the normalized real-time traffic semantic contribution to obtain the modulated real-time traffic time series association feature vector.

[0055] More specifically, in an embodiment of the present application, the post-modulation real-time temperature time series associated feature vector is used as a query vector, the post-modulation real-time flow time series associated feature vector is used as a key vector, and the temperature-flow condition feature vector is used as a value vector. The post-modulation real-time temperature time series associated feature vector, the post-modulation real-time flow time series associated feature vector, and the temperature-flow condition feature vector are input into a significant guided interaction module between features based on a converter structure to obtain a heat flow temperature-flow significant interaction fusion representation vector as the heat flow temperature-flow significant interaction fusion representation, including: inputting the post-modulation real-time temperature time series associated feature vector and the post-modulation real-time flow time series associated feature vector into a significant guided interaction module between features based on a converter structure to obtain a heat flow temperature-flow significant interaction fusion representation vector as the heat flow temperature-flow significant interaction fusion representation. The real-time temperature-flow timing association feature matrix is ​​obtained by multiplying the transposed vector of the time-flow timing association feature vector to obtain the real-time temperature-flow timing association feature matrix; the real-time temperature-flow timing association feature matrix is ​​divided by the square root of the length of the modulated real-time flow timing association feature vector to obtain the real-time temperature-flow timing association modulation feature matrix; the real-time temperature-flow timing association modulation feature matrix is ​​input into the softmax function to obtain the real-time temperature-flow timing association activation feature matrix; the real-time temperature-flow timing association activation feature matrix is ​​multiplied by the temperature-flow condition feature vector to obtain the heat flow temperature-flow significant interaction fusion representation vector.

[0056] In the embodiment of the present application, specifically, the multi-feature collaborative interactive fusion is performed on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector, which can be expressed by the formula:

[0057]

[0058] V c =sigmoid(V J )

[0059]

[0060] V 1 ′=V 1 w 1

[0061] V 2 ′=V 2 w 2

[0062]

[0063] Among them, V 1 and V 2 are respectively the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector, To add by position, W VT and b VTare the weight matrix and bias vector respectively, tanh(·) is the tanh function, V J is the real-time temperature-flow joint implicit feature vector, sigmoid(·) is the sigmoid function, V c represents the temperature-flow condition eigenvector, is the eigenvalue of each position in the real-time temperature time series associated eigenvector, is the characteristic value of each position in the real-time traffic time series associated characteristic vector, is the eigenvalue of each position in the temperature-flow condition eigenvector, L is the number of eigenvalues ​​in the real-time temperature time series associated eigenvector and the real-time flow time series associated eigenvector, log represents the logarithmic function value with base 2, exp(·) represents the exponential function with base e as the natural constant, CI 1 is the real-time temperature semantic contribution, CI 2 is the semantic contribution of real-time traffic, w 1 and w 2 are the normalized semantic contribution of real-time traffic and the normalized semantic contribution of real-time traffic, V 1 ' and V 2 ' are respectively the temporal correlation feature vector of real-time temperature after modulation and the temporal correlation feature vector of real-time flow after modulation, V 2 ' T Yes V 2 ', d is the transposed vector of V 2 'length, represents vector multiplication, Softmax(·) is the Softmax function, V r is the heat flow temperature-flow rate significant interaction fusion representation vector.

[0064] In step S515, based on the heat flow temperature-flow significant interactive fusion representation, a control instruction is obtained, and the control instruction includes a decoded value of the recommended refrigerant flow rate value. Specifically, in an embodiment of the present application, based on the heat flow temperature-flow significant interactive fusion representation, a control instruction is obtained, including: inputting the heat flow temperature-flow significant interactive fusion representation vector into a refrigerant flow rate controller based on a decoder to obtain the control instruction. That is, the heat flow temperature-flow significant interactive fusion representation obtained by significantly interactively fusion of the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector is used for decoding processing to automatically recommend the refrigerant flow rate value. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process more stable, thereby realizing the automatic control of the refrigerant flow rate.

[0065] In particular, when the real-time temperature time series correlation feature vector and the real-time flow rate time series correlation feature vector represent one-dimensional local time series correlation features of real-time temperature values ​​and real-time flow rate values ​​respectively, when performing significant guided interaction between features with joint implicit features as conditional information, the heat flow temperature-flow rate significant interaction fusion representation vector will also cause diversity of significant guided interaction based on joint conditions due to differences in time series implicit correlation feature patterns, affecting the iterative consistency of the decoding mapping and reducing the accuracy of the decoding results.

[0066] Preferably, the heat flow temperature-flow significant interaction fusion representation vector is input into the decoder-based refrigerant flow rate controller to obtain the control instruction, including: respectively calculating the first characteristic sum value of all eigenvalues ​​of the heat flow temperature-flow significant interaction fusion representation vector, the second characteristic sum value of all eigenvalues ​​of the real-time temperature time series associated feature vector, and the third characteristic sum value of all eigenvalues ​​of the real-time flow time series associated feature vector; dividing the first characteristic sum value by the second characteristic sum value and the third characteristic sum value, respectively, and calculating the square root of the sum of the squares of the two quotients to obtain the heat flow temperature-flow significant interaction fusion distribution probability value; calculating the heat flow temperature-flow significant interaction fusion distribution probability value. The product of the first heat flow temperature-flow significant interactive fusion characteristic value of the mutual fusion representation vector and the heat flow temperature-flow significant interactive fusion distribution probability value is obtained to obtain the first heat flow temperature-flow significant interactive fusion characteristic distribution probability product; the product of the second heat flow temperature-flow significant interactive fusion characteristic value of the heat flow temperature-flow significant interactive fusion representation vector and the absolute value of the difference between one and the heat flow temperature-flow significant interactive fusion distribution probability value is calculated to obtain the second heat flow temperature-flow significant interactive fusion characteristic inverse distribution probability product; the first heat flow temperature-flow significant interactive fusion characteristic value is calculated divided by one minus the heat flow temperature-flow significant interactive fusion distribution probability value. The quotient of the absolute value of the difference between the values ​​is used to obtain the inverse distribution probability quotient of the first heat flow temperature-flow significant interaction fusion feature; the quotient of the second heat flow temperature-flow significant interaction fusion feature value and the heat flow temperature-flow significant interaction fusion distribution probability value is calculated to obtain the second heat flow temperature-flow significant interaction fusion feature distribution probability quotient; the product of the first heat flow temperature-flow significant interaction fusion feature distribution probability and the product of the second heat flow temperature-flow significant interaction fusion feature inverse distribution probability are added, and then the difference between the first heat flow temperature-flow significant interaction fusion feature inverse distribution probability quotient and the second heat flow temperature-flow significant interaction fusion feature distribution probability quotient is subtracted to obtain Heat flow temperature-flow significant interaction fusion correction characteristic value; for each first heat flow temperature-flow significant interaction fusion characteristic value and second heat flow temperature-flow significant interaction fusion characteristic value of the heat flow temperature-flow significant interaction fusion representation vector, matrix multiplication of the heat flow temperature-flow significant interaction fusion correction characteristic matrix composed of the heat flow temperature-flow significant interaction fusion correction characteristic values ​​and the heat flow temperature-flow significant interaction fusion representation vector is performed to obtain a corrected heat flow temperature-flow significant interaction fusion representation vector; the corrected heat flow temperature-flow significant interaction fusion representation vector is input into the decoder-based refrigerant flow rate controller to obtain the control instruction.

[0067] The correction process of the heat flow temperature-flow significant interaction fusion representation vector is expressed as:

[0068]

[0069] m i,j ∈M,v i ∈V,v j(j≠i) ∈V

[0070]

[0071] and,

[0072]

[0073] Wherein, V is the heat flow temperature-flow significant interaction fusion representation vector, V 1 is the real-time temperature time series associated feature vector, V 2 is the real-time traffic time series correlation feature vector, v i is the eigenvalue of the ith position in V, v 1i is the eigenvalue of the ith position in the real-time temperature time series associated eigenvector, v 2i is the eigenvalue of the ith position in the real-time traffic time series associated eigenvector, p is the heat flow temperature-flow significant interaction fusion distribution probability value, v j is the eigenvalue of the jth position in the heat flow temperature-flow significant interaction fusion representation vector, m i,j represents the heat flow temperature-flow significant interaction fusion correction feature value at the (i, j)th position in the heat flow temperature-flow significant interaction fusion correction feature matrix, M is the heat flow temperature-flow significant interaction fusion correction feature matrix, represents matrix multiplication, V′ represents the corrected heat flow temperature-flow significant interaction fusion representation vector.

[0074] That is, the present application is aimed at the region-boundary integral relationship of the high-dimensional characteristic flow type of the heat flow temperature-flow significant interaction fusion representation vector, and by defining the spatial probability distribution of the feature set of the heat flow temperature-flow significant interaction fusion representation vector as a manifold constraint boundary relative to the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector, to approximate the high-dimensional manifold of the feature set of the heat flow temperature-flow significant interaction fusion representation vector, which is composed of a single connected region representation of eigenvalue pairs, so as to avoid the mapping ambiguity of the diversified feature representation of the heat flow temperature-flow significant interaction fusion representation vector to the local manifold representation in the high-dimensional feature decoding regression convergence space, improve the execution iteration consistency of each local feature distribution in the mapping task, improve the feature decoding regression convergence effect, and improve the accuracy of the control instruction obtained by the decoder-based refrigerant flow rate controller. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process more stable, thereby realizing the automatic control of the refrigerant flow rate.

[0075] In step S52, the second converted gas after cooling is subjected to the acid removal purification and pressure swing adsorption to obtain the carbon monoxide. It should be understood that there are still acidic gases in the gas after cooling. In order to purify the second converted gas after cooling, it is necessary to perform acid removal purification on the second converted gas after cooling. In the embodiment of the present application, the acid removal purification operation of the second converted gas after cooling is still carried out by the low-temperature methanol washing method. Low-temperature methanol washing is based on Raoult's law and Henry's law. It is based on the selectivity that methanol in a low temperature state has a large solubility and absorption of acidic gases such as H2S and CO2, and a small solubility and absorption of H2 and CO, to remove acidic gases such as H2S and CO2 from the second converted gas after cooling. In order to separate and extract carbon monoxide from the gas after acid removal purification, it is necessary to perform pressure swing adsorption on the gas. Under high pressure, the adsorbent adsorbs impurity components, while the ideal component (here refers to CO) passes through the adsorbent bed because of its small adsorption capacity. Then, under low pressure, the adsorbent releases the impurity components, completing the regeneration of the adsorbent. Through periodic pressure changes, continuous separation and purification of CO can be achieved.

[0076] In summary, a method for comprehensive utilization of waste gas based on an embodiment of the present application is explained, which collects the time queue data of the real-time temperature value and the real-time flow value of the second stream of converted gas of the water cooler through the temperature sensor and the flow sensor installed at the front end of the water cooler, and uses a data processing technology based on deep learning to perform a time series correlation analysis on the real-time temperature value and the real-time flow value, so as to automatically recommend the refrigerant flow rate value according to the synergistic significant interaction representation between the real-time temperature time series correlation feature and the real-time flow time series correlation feature. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process smoother, thereby realizing the automatic control of the refrigerant flow rate.

[0077] Figure 6 FIG. 1 is a block diagram of a device for comprehensive utilization of raw coal gas according to an embodiment of the present application. Figure 6 As shown, the device for comprehensive utilization of raw coal gas 100 according to the embodiment of the present application includes: a raw coal gas compression processing module 110, which is used to compress the raw raw coal gas introduced to obtain compressed raw coal gas; a compressed raw coal gas conversion processing module 120, which is used to convert the compressed raw coal gas to obtain converted gas; a gas diversion processing module 130, which is used to divert the converted gas to obtain a first stream of converted gas and a second stream of converted gas; a hydrogen preparation and generation module 140, which is used to perform conversion treatment, acid depuration and pressure swing adsorption on the first stream of converted gas to obtain hydrogen; a carbon monoxide preparation and generation module 150, which is used to perform heat recovery, acid depuration and pressure swing adsorption on the second stream of converted gas to obtain carbon monoxide.

[0078] Here, those skilled in the art will appreciate that the specific functions and operations of the various units and modules in the above-mentioned waste gas comprehensive utilization device 100 have been described in detail above. Figures 1 to 5 The method for comprehensive utilization of waste gas has been described in detail, and therefore, its repeated description will be omitted.

[0079] In summary, the waste gas comprehensive utilization device 100 based on the embodiment of the present application is explained, which collects the time queue data of the real-time temperature value and the real-time flow value of the second stream of converted gas of the water cooler through the temperature sensor and the flow sensor installed at the front end of the water cooler, and uses the data processing technology based on deep learning to perform time series correlation analysis on the real-time temperature value and the real-time flow value, so as to automatically recommend the refrigerant flow rate value according to the synergistic significant interaction representation between the real-time temperature time series correlation feature and the real-time flow time series correlation feature. In this way, the refrigerant flow rate can be dynamically adjusted to ensure that the best heat recovery effect can be achieved under different working conditions. And based on the real-time data, the temperature and flow fluctuations caused by untimely manual adjustment can be reduced, making the heat recovery process smoother, thereby realizing the automatic control of the refrigerant flow rate.

Claims

1. A method for comprehensive utilization of waste gas, characterized in that: include: S1: compressing the raw raw gas to obtain compressed raw gas; S2: converting the compressed raw coal gas to obtain converted gas; S3: splitting the converted gas to obtain a first stream of converted gas and a second stream of converted gas; S4: performing shift treatment, acid removal purification and pressure swing adsorption on the first converted gas to obtain hydrogen; S5: The second converted gas is subjected to heat recovery, acid removal and pressure swing adsorption to obtain carbon monoxide.

2. The method for comprehensive utilization of waste gas according to claim 1, characterized in that: The second converted gas is subjected to heat recovery, acid removal and pressure swing adsorption to obtain carbon monoxide, including: Passing the second converted gas through a low-pressure waste heat boiler, a desalted water preheater and a water cooler in sequence to obtain a cooled second converted gas; The cooled second converted gas is subjected to the acid removal and pressure swing adsorption to obtain the carbon monoxide.

3. The method for comprehensive utilization of waste gas according to claim 2, characterized in that: The second converted gas is sequentially passed through a low-pressure waste heat boiler, a desalted water preheater and a water cooler to obtain a cooled second converted gas, comprising: Install a temperature sensor and a flow sensor at the front end of the water cooler; The temperature sensor and the flow sensor are used to collect a time queue of a real-time temperature value and a time queue of a real-time flow value of the second stream of converted gas flowing into the water cooler; Performing sequence encoding on the time queue of the real-time temperature value and the time queue of the real-time flow value to obtain a real-time temperature time series associated feature vector and a real-time flow time series associated feature vector; The real-time temperature time series associated feature vector and the real-time flow time series associated feature vector are subjected to multi-feature collaborative interactive fusion to obtain a heat flow temperature-flow significant interactive fusion representation, including: performing implicit feature combination on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain a temperature-flow condition feature; taking the temperature-flow condition feature as condition information, performing significant interaction based on semantic contribution on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain the heat flow temperature-flow significant interactive fusion representation; Based on the heat flow temperature-flow significant interaction fusion representation, a control instruction is obtained, and the control instruction includes a decoded value of a recommended refrigerant flow rate value.

4. The method for comprehensive utilization of waste gas according to claim 3, characterized in that: The time queue of the real-time temperature value and the time queue of the real-time flow value are serially encoded to obtain a real-time temperature time series associated feature vector and a real-time flow time series associated feature vector, including: inputting the time queue of the real-time temperature value and the time queue of the real-time flow value into a sequence encoder based on a 1D-CNN model to obtain the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector.

5. The method for comprehensive utilization of waste gas according to claim 4, characterized in that: Performing implicit feature combination on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain temperature-flow condition feature, including: Inputting the real-time temperature time series associated feature vector and the real-time flow rate time series associated feature vector into a joint implicit feature capture network to obtain a real-time temperature-flow rate joint implicit feature vector; The real-time temperature-flow joint implicit feature vector is activated using a Sigmoid function to obtain a temperature-flow condition feature vector as the temperature-flow condition feature.

6. The method for comprehensive utilization of waste gas according to claim 5, characterized in that: Taking the temperature-flow condition feature as condition information, performing significant interaction based on semantic contribution on the real-time temperature time series associated feature vector and the real-time flow time series associated feature vector to obtain the heat flow temperature-flow significant interaction fusion representation, including: Calculating the real-time temperature semantic contribution of the real-time temperature time series associated feature vector relative to the temperature-flow condition feature vector; Calculating the real-time traffic semantic contribution of the real-time traffic time series association feature vector relative to the temperature-traffic condition feature vector; Normalizing the real-time temperature semantic contribution and the real-time traffic semantic contribution, and using the normalized real-time temperature semantic contribution and the normalized real-time traffic semantic contribution to weighted modulate the real-time temperature time series association feature vector and the real-time traffic time series association feature vector to obtain a modulated real-time temperature time series association feature vector and a modulated real-time traffic time series association feature vector; Taking the post-modulation real-time temperature time series associated feature vector as the query vector, the post-modulation real-time flow time series associated feature vector as the key vector and the temperature-flow condition feature vector as the value vector, the post-modulation real-time temperature time series associated feature vector, the post-modulation real-time flow time series associated feature vector and the temperature-flow condition feature vector are input into a feature-to-feature significant guided interaction module based on a converter structure to obtain a heat flow temperature-flow significant interaction fusion representation vector as the heat flow temperature-flow significant interaction fusion representation.

7. The method for comprehensive utilization of waste gas according to claim 6, characterized in that: Calculating the real-time temperature semantic contribution of the real-time temperature time series association feature vector relative to the temperature-flow condition feature vector includes: Calculating the positional division of the real-time temperature time series association feature vector and the temperature-flow condition feature vector to obtain a real-time temperature semantic contribution vector; Calculating a base-two logarithmic function value of the absolute value of each eigenvalue of the real-time temperature semantic contribution vector to obtain a real-time temperature semantic contribution logarithmic vector; Calculate the positional dot product of the real-time temperature time series associated feature vector and the real-time temperature semantic contribution logarithm vector, and perform positional multiplication on the obtained dot product vector to obtain the real-time temperature semantic contribution value; An exponential function with the natural constant e as the base and the real-time temperature semantic contribution value as the exponent is calculated to obtain the real-time temperature semantic contribution degree.

8. The method for comprehensive utilization of waste gas according to claim 7, characterized in that: The real-time temperature semantic contribution and the real-time flow semantic contribution are normalized, and the real-time temperature time series association feature vector and the real-time flow time series association feature vector are weighted modulated using the normalized real-time temperature semantic contribution and the normalized real-time flow semantic contribution to obtain a modulated real-time temperature time series association feature vector and a modulated real-time flow time series association feature vector, including: Calculating the sum of the real-time temperature semantic contribution and the real-time flow semantic contribution to obtain a total sum of real-time temperature-flow semantic contributions; Respectively dividing the real-time temperature semantic contribution and the real-time traffic semantic contribution by the total sum of the real-time temperature-traffic semantic contributions to obtain the normalized real-time temperature semantic contribution and the normalized real-time traffic semantic contribution; Multiplying the real-time temperature time series associated feature vector by the normalized real-time temperature semantic contribution by position point to obtain the modulated real-time temperature time series associated feature vector; The real-time traffic time series correlation feature vector and the normalized real-time traffic semantic contribution are multiplied by position points to obtain the modulated real-time traffic time series correlation feature vector.

9. The method for comprehensive utilization of waste gas according to claim 8, characterized in that: Based on the heat flow temperature-flow significant interaction fusion representation, a control instruction is obtained, including: inputting the heat flow temperature-flow significant interaction fusion representation vector into a decoder-based refrigerant flow rate controller to obtain the control instruction.

10. A device for comprehensive utilization of waste gas, characterized in that: include: A raw gas compression processing module is used to compress the raw raw gas introduced to obtain compressed raw gas; A compressed raw coal gas conversion processing module, used for converting the compressed raw coal gas to obtain converted gas; A gas splitting processing module, used for splitting the converted gas to obtain a first stream of converted gas and a second stream of converted gas; A hydrogen preparation module, used for performing conversion treatment, acid removal purification and pressure swing adsorption on the first converted gas to obtain hydrogen; The carbon monoxide preparation module is used to perform heat recovery, acid removal and pressure swing adsorption on the second converted gas to obtain carbon monoxide.

Images