Intelligent control method and system for hydrogen production based on methanol cracking by using big data

Abstract

The application discloses a kind of methanol cracking hydrogen production intelligent control method based on big data, obtains hydrogen demand data, methanol raw material parameter data and catalyst parameter data;Hydrogen demand data is used to predict target hydrogen data in preset time period;Target hydrogen data, methanol raw material concentration data and catalyst parameter data are input into hydrogen production optimization process model, and target control parameter output by hydrogen production optimization process model is obtained;Based on target control parameter, adjust purification process parameter and hydrogen production process parameter;Based on the hydrogen production equipment after adjustment, real-time hydrogen data is obtained, and real-time hydrogen data is compared and analyzed with target hydrogen data, to judge whether it exceeds deviation threshold, if it exceeds deviation threshold, then real-time adjustment purification process parameter and hydrogen production process parameter.The application realizes intelligent quantitative control in cracking hydrogen production, as far as possible to avoid the waste of resources, and the problem of high production cost.

Description

Technical Field

[0001] This invention relates to the field of methanol-to-hydrogen, and more particularly to a smart control method and system for methanol cracking-to-hydrogen production based on big data. Background Technology

[0002] With the development of clean energy technologies, hydrogen energy, as a clean and efficient energy source, has received widespread attention globally. Among various hydrogen production technologies, methanol cracking for hydrogen production has gradually become a research hotspot due to its abundant raw materials and simple process. The methanol cracking hydrogen production process mainly decomposes methanol into hydrogen and carbon monoxide through a catalytic reaction, accompanied by small amounts of byproducts such as dimethyl ether. This technology not only efficiently generates hydrogen but also demonstrates great potential in industrial and mobile applications due to its low energy consumption and relatively simple operation.

[0003] Currently, hydrogen production commonly employs a production-driven sales model. This means that companies purchase a certain amount of methanol as raw material based on past production experience and estimated market demand, and then produce hydrogen according to a predetermined process. Under this model, because the production processes and equipment are relatively fixed, it's difficult to quickly adjust the product structure. This can lead to over-purification of methanol after obtaining the raw material, neglecting actual demand and resulting in resource waste. Furthermore, under this production-driven sales model, companies may continue operating hydrogen production equipment even if market demand is insufficient in order to maintain a certain production scale and output, leading to energy waste and increased production costs. Summary of the Invention

[0004] The purpose of this invention is to solve the problems in the prior art, and to propose a smart control method and system for methanol cracking to produce hydrogen based on big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A smart control method for methanol cracking to hydrogen production based on big data includes the following steps:

[0007] Obtain hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data;

[0008] Predict target hydrogen data within a preset time period based on the hydrogen demand data;

[0009] The target hydrogen data, methanol feedstock concentration data, and catalyst parameter data are input into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from sample data and its corresponding control parameter label results.

[0010] Adjust the purification process parameters and hydrogen production process parameters based on the target control parameters;

[0011] Based on the real-time hydrogen data obtained from the adjusted hydrogen production equipment, the real-time hydrogen data is compared and analyzed with the target hydrogen data to determine whether it exceeds the deviation threshold. If it exceeds the deviation threshold, the purification process parameters and hydrogen production process parameters are adjusted in real time.

[0012] A smart control system for methanol cracking to hydrogen production based on big data includes:

[0013] The acquisition unit is used to acquire hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data.

[0014] The prediction unit is used to predict target hydrogen data within a preset time period based on the hydrogen demand data.

[0015] The control parameter determination unit is used to input the target hydrogen data, methanol feedstock concentration data, and catalyst parameter data into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from sample data and its corresponding control parameter label results.

[0016] The output unit is used to adjust the purification process parameters and the hydrogen production process parameters based on the target control parameters.

[0017] The feedback adjustment unit is used to acquire real-time hydrogen data based on the adjusted hydrogen production equipment, compare and analyze the real-time hydrogen data with the target hydrogen data, determine whether the deviation threshold is exceeded, and if the deviation threshold is exceeded, adjust the purification process parameters and hydrogen production process parameters in real time.

[0018] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described intelligent control method for methanol cracking to hydrogen production based on big data.

[0019] The present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described intelligent control method for methanol cracking to hydrogen production based on big data.

[0020] Compared with the prior art, the present invention has the following advantages:

[0021] This invention provides a big data-based intelligent control method for methanol cracking to produce hydrogen. It accurately predicts target hydrogen data within a preset time period. Based on the predicted target hydrogen data, methanol feedstock parameters, and catalyst parameters, it determines the control parameters for the purification and hydrogen production processes. This initially achieves intelligent control of cracking to produce hydrogen under a production-to-order model, avoiding waste caused by excessive purification of methanol feedstock and excessive hydrogen production. Furthermore, it monitors real-time hydrogen production data and compares it with target hydrogen data to determine if any deviations occur. If deviations are found, the parameters of the purification and hydrogen production processes are adjusted accordingly, further achieving intelligent quantitative control of cracking to produce hydrogen, minimizing resource waste and high production costs. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a schematic flowchart of a smart control method for methanol cracking to hydrogen production based on big data, provided in an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the structure of the intelligent control system for methanol cracking to hydrogen production based on big data provided in an embodiment of the present invention;

[0025] Figure 3 An embodiment diagram of the electronic device provided in this invention;

[0026] Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0029] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0030] See Figure 1 , Figure 1 This is a flowchart illustrating a big data-based intelligent control method for methanol cracking to hydrogen production provided by the present invention. In this embodiment, the executing entity of the big data-based intelligent control method for methanol cracking to hydrogen production is an intelligent control system. Therefore, the big data-based intelligent control method for methanol cracking to hydrogen production includes:

[0031] Step 10: Obtain hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data.

[0032] Specifically, the intelligent control system in this embodiment of the invention uses a back-end server in a hydrogen production workshop. The hydrogen demand data includes historical hydrogen production data, historical hydrogen sales data, current pre-order data, and industry policy data. When acquiring hydrogen demand data, the intelligent control system collects data from multiple data sources, including market sales orders, industry reports, historical sales records, etc. After acquiring the relevant data, it stores the data in the storage device of the cloud platform or back-end server that is connected to the intelligent control system.

[0033] Furthermore, in addition to conventional concentration and purity data, methanol feedstock parameter data also includes information such as impurity composition. When acquiring methanol feedstock parameter data, the intelligent control system uses high-precision chemical analysis instruments (near-infrared spectrometer) for detection, and the detection results are also stored in the background server or cloud platform. Catalyst parameter data includes parameters such as catalyst activity, selectivity, and lifetime, and is measured using methods such as X-ray diffraction and scanning electron microscopy. The measurement results are also stored in the background server or cloud platform for easy access by the intelligent control system.

[0034] It is important to note that during the storage of hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data, the intelligent control system first carefully cleans the collected data to remove duplicates, errors, and missing values. It then employs methods such as logical verification and cross-validation to ensure data accuracy and consistency, thus preventing deviations in subsequent predictions due to poor data quality. Furthermore, for data collection involving multiple sub-data sets, an aggregate approach is used. For example, assuming the hydrogen demand data set is D... H1 It contains n data points of different types. but The methanol feedstock parameter data set is D MeOH It contains m parameters but The catalyst parameter data set is D cat It contains k parameters but This enables the storage of hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data, achieving comprehensive data acquisition and more accurately reflecting the actual situation, thus providing a more reliable foundation for subsequent prediction and control.

[0035] Step 20: Predict the target hydrogen data within a preset time period based on hydrogen demand data.

[0036] Specifically, when the intelligent control system acquires hydrogen demand data, it analyzes historical hydrogen production data, historical hydrogen sales data, current pre-purchase data, and industrial policy data to predict the target hydrogen production data required for the current production cycle (i.e., within a preset time period). This allows the prediction results to better reflect the actual situation from multiple data and factor outcomes, enhancing the system's flexibility. The specific process is described in steps 201-205.

[0037] Furthermore, by predicting the target hydrogen data, the intelligent control system can provide precise data support for the subsequent adjustment of equipment parameters in the methanol cracking hydrogen production control process, based on the accurately predicted production target, including the purity of the target hydrogen and the corresponding quantity of purity.

[0038] Step 30: Input the target hydrogen data, methanol feedstock concentration data, and catalyst parameter data into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from the sample data and its corresponding control parameter label results.

[0039] Specifically, after acquiring target hydrogen data, methanol feedstock concentration data, and catalyst parameter data, the intelligent control system processes the acquired data based on a pre-trained hydrogen production optimization process model to balance economic benefits and environmental protection goals. This process ultimately yields the target control parameters for the equipment, including the control parameters for the equipment used in the hydrogen production process and the equipment used in the methanol feedstock purification process, thereby achieving precise control of the overall equipment.

[0040] Furthermore, the hydrogen production optimization process model employs a deep learning model, such as a neural network, trained with a large number of samples to ultimately obtain a fully trained hydrogen production optimization process model. The specific training process is described in step 301. Using a deep learning model for training enables the automatic learning of complex relationships in the data, improving the model's accuracy and generalization ability.

[0041] Step 40: Adjust the purification process parameters and hydrogen production process parameters based on the target control parameters.

[0042] Specifically, the intelligent control system adopts an automated control system. After receiving the target control parameters, it automatically adjusts the purification process parameters (such as temperature, pressure, flow rate, etc.) and hydrogen production process parameters (such as reaction temperature, catalyst dosage, etc.) according to the target control parameters. During the control and adjustment process, the overall parameters and the hydrogen production process are monitored in real time to facilitate real-time adjustment of parameters based on subsequent feedback. This improves the efficiency and accuracy of adjustment while reducing interference from human factors. The specific adjustment process is described in steps 401-405.

[0043] Step 50: Obtain real-time hydrogen data based on the adjusted hydrogen production equipment, and compare and analyze the real-time hydrogen data with the target hydrogen data to determine whether the deviation threshold is exceeded. If the deviation threshold is exceeded, adjust the purification process parameters and hydrogen production process parameters in real time.

[0044] Specifically, the intelligent control system employs high-precision sensors and video monitoring during the adjustment of target control parameters and hydrogen production. For hydrogen production, sensors monitor real-time data such as hydrogen output and purity. Simultaneously, the intelligent control system, based on the target hydrogen data acquired in step 20 (which also includes hydrogen output and purity data within a specific timeframe), compares and analyzes the real-time hydrogen data with the target hydrogen data. A reasonable deviation threshold is set to determine if the real-time hydrogen data exceeds this threshold. If it does not exceed the threshold, the overall production process is proceeding stably according to the preset production plan. If it exceeds the threshold, it indicates an unexpected situation in the overall production plan. Therefore, the purification and hydrogen production process parameters can be adjusted in real-time using a feedback control algorithm (such as PID control) to ensure stable production according to the preset calculations. This allows for timely detection of abnormalities in the process, ensuring production stability and improving the anti-interference capability of the intelligent control system.

[0045] Furthermore, the specific analysis process for comparing and analyzing real-time hydrogen data with target hydrogen data is described in steps 501-503.

[0046] This invention relates to the field of methanol-to-hydrogen technology and proposes a big data-based intelligent control method for methanol cracking to produce hydrogen. The proposed intelligent control method accurately predicts target hydrogen data within a preset time period. Based on the predicted target hydrogen data, methanol feedstock parameters, and catalyst parameters, it determines the control parameters for the purification and hydrogen production processes. This initially achieves intelligent control of cracking to produce hydrogen under a production-to-order model, avoiding waste caused by excessive purification of methanol feedstock and excessive hydrogen production. Furthermore, it monitors real-time hydrogen production data and compares it with target hydrogen data to determine if any deviations occur. If deviations are found, the parameters of the purification and hydrogen production processes are adjusted accordingly, further achieving intelligent quantitative control of cracking to produce hydrogen, minimizing resource waste and high production costs.

[0047] In one embodiment, steps 201-205 are described as follows:

[0048] Step 201: Match historical hydrogen production data and historical hydrogen sales data based on time series data to obtain a hydrogen sales and production dataset.

[0049] Specifically, after determining historical hydrogen production and sales data, the intelligent control system first sorts the data chronologically to ensure consistency of timestamps. For example, using the Dynamic Time Warping (DTW) algorithm, it flexibly matches the two time series data along the time axis to find the optimal matching path, addressing potential inconsistencies in time intervals between production and sales data. Ultimately, it obtains historical hydrogen production and sales data at the same timestamp.

[0050] Furthermore, the intelligent control system will also combine the corresponding production data and sales data into a data pair based on each matched time point, and all data pairs will constitute a hydrogen sales and production dataset.

[0051] Furthermore, in one embodiment, if the historical hydrogen production data sequence is P li ={p1,p2,…,p q The historical hydrogen sales data series is S. li ={s1,s2,…,s w After matching using the DTW algorithm, the resulting hydrogen sales and production dataset was obtained. Where i x and j x These are the indices of production and sales data within their respective sequences, and their correspondence was determined using the DTW algorithm. This allows for matching data across different time intervals using a dynamic time warping algorithm, avoiding erroneous data associations caused by time alignment issues, and ensuring that the resulting hydrogen sales and production dataset more accurately reflects the relationship between production and sales.

[0052] Step 202: The hydrogen sales production dataset is segmented based on a preset production cycle interval to obtain multiple inventory backlog rates.

[0053] Specifically, the intelligent control system presets a production cycle interval of T0 (e.g., T0 can be one week or one month). Starting from the start time of the acquired hydrogen sales and production dataset, the system is divided into multiple time periods with T0 as the interval. For each time period, the inventory backlog rate is calculated. The formula for calculating the inventory backlog rate is: Where P t S represents the amount of hydrogen produced during time period t. t Let ∑ represent the hydrogen sales volume within time period t, and se represent a specific time period obtained after segmenting the hydrogen sales and production dataset. t∈se This represents the summation of data at all points in time within the specified time period.

[0054] Furthermore, in one embodiment, if the preset production cycle interval is one month, for the first month, the daily hydrogen production and sales data for that month are substituted into the above formula to calculate the inventory backlog rate for that month. Similarly, the entire hydrogen sales and production dataset is segmented according to preset intervals to obtain multiple inventory backlog rates IR1, IR2, ..., IR... z By calculating the inventory backlog rate in segments according to the production cycle intervals, it is possible to clearly see the changes in inventory over different time periods, which helps to analyze the cyclical patterns and trends of inventory backlog.

[0055] Step 203: Based on the changing patterns of multiple inventory backlog rates, predict the current inventory backlog rate, and determine the fault tolerance rate within the current production cycle based on the current inventory backlog rate.

[0056] Specifically, after obtaining multiple inventory backlog rates over a period of time, the intelligent control system can use various analysis methods to determine the changing patterns among these rates. Based on these patterns, it can predict the inventory backlog rate for the current time period. During production, companies typically set up reserve quantities for their production plans to avoid unexpected or sudden events and improve their tolerance for errors during production and sales. Therefore, by fully utilizing the time correlation and changing patterns of historical inventory backlog rate data along with reserve quantities, the system can more accurately predict capacity rates, providing important references for production decisions and helping companies rationally arrange production to cope with potential inventory fluctuations. However, it should be noted that the inventory backlog rate and reserve quantities for hydrogen production and sales are often affected by seasonal factors. Therefore, it is necessary to comprehensively consider seasonal factors to more accurately determine the tolerance rate. The specific process is described in steps 2031-2034.

[0057] Step 204: Determine the policy impact index based on the degree of impact of industrial policy data on hydrogen demand.

[0058] Specifically, in determining the target hydrogen data, the intelligent control system should also consider the impact of industrial policies on actual demand. This is because the research and development and use of hydrogen are both affected by economic factors, which in turn are affected by industrial policies. For example, if industrial policies encourage the development of hydrogen fuel cell vehicles, it will stimulate automakers to increase their investment in the research and development and use of hydrogen fuel cells, thereby affecting the demand for hydrogen. Therefore, by comprehensively considering the impact of industrial policies and determining the policy impact index, the target hydrogen data can be predicted more accurately. The specific process of determining the policy impact index is described in steps 2041-2044.

[0059] Step 205: Integrate the current pre-purchase data with the fault tolerance rate and policy impact index within the current production cycle to obtain the target hydrogen data.

[0060] Specifically, after determining the fault tolerance rate and government influence index within the current production cycle, the intelligent control system uses a weighted average method to integrate the three factors that affect the target hydrogen based on the initial current estimated data. Specifically, the target hydrogen data THD = ω1*OD + ω2*FR + ω3*PI, where OD represents the current estimated data, FR represents the fault tolerance rate within the current production cycle, PI represents the government influence index, and ω1, ω2, and ω3 are the corresponding weights, and ω1 + ω2 + ω3 = 1.

[0061] Furthermore, in determining ω1, ω2, and ω3, the intelligent control system uses the AHP method to construct a judgment matrix A, where A... ij To represent the importance of factor i relative to factor j, the largest eigenvalue λ of the judgment matrix is ​​calculated. max The weight vector is obtained by normalizing the corresponding feature vector W0. This ensures that the weights can reasonably reflect the relative importance of each factor.

[0062] This invention comprehensively considers various factors such as historical production and sales data, current pre-order data, inventory status, and industrial policies to predict target hydrogen data, making the prediction results more comprehensive and accurate in reflecting actual market demand. In addition, through time series analysis and real-time data processing, it can promptly capture the impact of market changes and policy adjustments on hydrogen demand, and achieve dynamic prediction and adjustment.

[0063] In one embodiment, steps 2031-2034 are described as follows:

[0064] Step 2031: Based on the changing patterns of multiple inventory backlog rates, determine the changing trend, and predict the current inventory backlog rate based on the changing trend.

[0065] Specifically, after receiving multiple inventory backlog rates, the intelligent control system can use time series analysis to process the data. First, a moving average is calculated; let the inventory backlog rate sequence be I. o ={y1,y2,…,y i If the moving average period is u, then the moving average sequence MA = {ma1, ma2, ..., ma} i-u+1},in h = 1, 2, ..., i-u+1, and then observe the trend of the moving average sequence to determine the initial trend.

[0066] Furthermore, the inventory backlog rate sequence is fitted using a linear regression model, with the linear regression equation y = y. to =a+bto+∈ t0 , where yto Let a represent the inventory backlog rate at time t0, a represent the intercept, and b represent the slope. t0 Let represent the error term. Then, the parameters a and b are estimated using the least squares method, i.e., by minimizing ... Determine the values ​​of a and b, and based on the sign of b, determine whether the inventory backlog rate is rising or falling. Make a forecast based on the determined trend; if the trend is linearly increasing, predict the current inventory backlog rate. If the trend is linearly downward, the same formula can be used for prediction. On the other hand, for complex trends, more advanced time series models such as ARIMA (Autoregressive Integral Moving Average) can be used. First, the data is differencing to make it stationary, and then the model parameters p are determined. q (autoregressive order), d q (difference order), q q (Moving average order) is used to make predictions after training the model with historical data.

[0067] Step 2032: Obtain the reserve quantity for the current production cycle and determine the current reserve rate based on the reserve quantity.

[0068] Specifically, the intelligent control system connects to the enterprise's inventory management system to obtain the reserved hydrogen reserve quantity S for the current production cycle. q Calculated output P during the current production cycle ji The current reserve quantity

[0069] Step 2033: Adjust the current inventory backlog rate and the current reserve rate based on seasonal factors to obtain the target inventory backlog rate and the target reserve rate.

[0070] Specifically, the intelligent control system collects inventory backlog rate and reserve rate data from multiple past production cycles, categorizing them by season. It then calculates the average inventory backlog rate for each season. and the average reserve ratio and the overall mean and Regarding the current inventory backlog rate Seasonal adjustment factor Target inventory backlog rate If it is currently summer, the average summer inventory backlog rate is 0.2, the overall average is 0.15, and the predicted current inventory backlog rate is 0.18, then the seasonal adjustment factor... Target inventory backlog rate For the current reserve ratio SR, the seasonal adjustment factor Target Reserve Ratio (SR) tar =SR*C SR .

[0071] Step 2034: Compare the target inventory backlog rate with the target reserve rate. If the target inventory backlog rate is greater than or equal to the target reserve rate, the current inventory backlog rate is determined as the fault tolerance rate in the current production cycle. If the target inventory backlog rate is less than the target reserve rate, the target reserve rate is determined as the fault tolerance rate in the current production cycle.

[0072] Specifically, the intelligent control system receives the target inventory backlog rate I tar and target reserve ratio SR tar Then, it is judged, when I tar ≥SR tar When this occurs, it indicates that the current inventory backlog is more severe than the reserve situation, and at this point, the tolerance rate FR within the current production cycle is I. tar . When I tar <SR tar This indicates that the backup situation is relatively more sufficient, and the fault tolerance rate FR = SR in the current production cycle. tar Ultimately, the fault tolerance rate is determined through comparison, taking into account the relative relationship between inventory backlog and reserve conditions, and providing a more realistic reference indicator for production decisions.

[0073] This invention comprehensively analyzes and determines the tolerance rate by taking into account multiple aspects such as the trend of inventory backlog rate changes, reserve capacity, and seasonal factors, making the results more reliable. In addition, it grasps the trend of inventory backlog rate changes through time series analysis and combines seasonal adjustments to adapt to the production and operation characteristics of different periods and provide a dynamic and realistic tolerance rate.

[0074] In one embodiment, steps 2041-2044 are described as follows:

[0075] Step 2041 involves semantic parsing and classification of the industrial policy data to obtain multiple sub-policies.

[0076] Specifically, the intelligent control system uses natural language processing technology, employing pre-trained language models (such as BERT) to encode industrial policy text data, transforming the text into a computer-understandable vector form. Through lexical, syntactic, and semantic analysis, key information from the policy text is extracted, such as the policy subject, policy object, policy measures, and policy objectives. Based on the extracted key information, classification rules are established to categorize the policies. For example, according to the policy's operational stage, it can be divided into hydrogen production policies, hydrogen storage policies, hydrogen transportation policies, and hydrogen utilization policies; according to the policy's nature, it can be divided into incentive policies, regulatory policies, and restrictive policies.

[0077] Step 2042: Compare each sub-policy with the policy quantification database to obtain the quantification score for each sub-policy.

[0078] Specifically, the intelligent control system first constructs a policy quantification database. This database pre-stores various common sub-policies, their corresponding quantification score ranges, and quantification criteria. The quantification criteria can include the scale of funding involved in the policy, its coverage, and the strength of policy implementation. For example, for a hydrogen production subsidy sub-policy, a higher quantification score corresponds to a certain amount of subsidy funding; a larger number of companies covered by the subsidy will also result in a higher quantification score. For each sub-policy, a matching quantification standard is searched in the database based on its specific content. For example, a hydrogen production subsidy sub-policy stipulates a subsidy of 500 yuan per ton of green hydrogen. The database is searched for the correspondence between the green hydrogen subsidy amount and the quantification score; assuming a subsidy of 500 yuan corresponds to a quantification score of 8 points. For some sub-policies that are difficult to match directly, an expert scoring method combined with reference standards in the database is used for quantification. Experts assign quantification scores within the range specified in the database based on the actual impact of the sub-policy.

[0079] Furthermore, let the set of sub-policies be... The policy quantification database is DS, which stores sub-policy types t. t With quantified score q f The correspondence (t) t q f For sub-policies s i Determine its type Search for the corresponding quantified score in the DS database. If there is no direct match in the database, let the expert scoring function be f(s). i Experts, based on sub-policies i The situation falls within the score range specified in the database. The internal quantitative score is given. and

[0080] Step 2043: Add the quantitative scores of multiple sub-policies together to obtain the quantitative value of the comprehensive impact of the current policy on hydrogen demand.

[0081] Specifically, the intelligent control system assigns a quantitative score to each sub-policy it obtains. The summation operation is performed to obtain the comprehensive impact quantification value.

[0082] Step 2044: Based on the preset benchmark value, the comprehensive impact quantification value is judged to obtain the policy impact index.

[0083] Specifically, the intelligent control system presets a baseline value Q0. This baseline value represents a quantitative value indicating that the policy has no significant impact on hydrogen demand. For example, if the quantitative score range is 0-10, the baseline value Q0 can be set to 5. Policy Impact Index I zh The calculation method is as follows: when Qz ≥Q0, When Q z < Q0, The policy impact index ranges from [0, 1]. The larger the value, the stronger the promoting effect of the policy on hydrogen demand.

[0084] In the embodiments of the present invention, a complete system is formed from policy text parsing, sub-policy classification, quantification to final index determination, comprehensively and systematically evaluating the impact degree of industrial policies on hydrogen demand. With the help of natural language processing technology and policy quantification databases, the interference of human subjective factors is reduced, making the evaluation process and results more objective and accurate.

[0085] In one embodiment, the description of step 301 is as follows:

[0086] Step 301, obtain sample data; the sample data includes target hydrogen data, methanol raw material concentration data, and catalyst parameter data; input the sample data into a pre-trained model to obtain the prediction result output by the pre-trained model; obtain the loss value based on the loss function of the pre-trained model according to the pre-trained result; adjust the optimization function in the pre-trained model based on the loss value; predict the prediction result of the sample data based on the adjusted pre-trained module until the loss value output by the loss function is greater than zero, less than or equal to the first preset value (such as 0.05), at least five consecutive loss values decrease in sequence, and the difference between adjacent loss values is less than or equal to the second preset value (such as 0.005).

[0087] Specifically, the intelligent control system collects sample data from multiple channels such as enterprise historical production records, experimental data, and market research. When training the pre-trained model, deep neural network (DNN), recurrent neural network (RNN), and its variant long short-term memory network (LSTM) can be selected. For example, when using LSTM as the training model, the target hydrogen data, methanol raw material concentration data, and catalyst parameter data are preprocessed and transformed into a tensor form suitable for LSTM input. The processed sample data is input into the LSTM model batch by batch. After the input layer in the model receives the data, multiple LSTM units learn the time series features and complex nonlinear relationships in the data. Each LSTM unit contains an input gate, a forget gate, and an output gate, which can effectively handle information transmission and memory problems in long sequence data. After layer-by-layer calculations in the hidden layer, the output layer finally outputs the prediction result, which includes the prediction of various parameters in the hydrogen production process, such as reaction temperature, pressure, hydrogen production, and purity.

[0088] This invention, through a comprehensive assessment of multiple conditions, more accurately determines the convergence state of model training, avoiding premature or late training termination and improving training quality. The first preset value is determined based on the precision requirements and experience of actual hydrogen production processes, ensuring the loss value remains within an acceptable range. When at least five consecutive sets of loss values ​​show a decreasing trend, it indicates that the model is effectively converging towards optimization. The largest difference between adjacent loss values ​​refers to the absolute difference between two consecutive loss values. The second preset value is also set empirically; when the largest difference between adjacent loss values ​​is less than or equal to this value, it indicates that the change in loss value tends to stabilize, and the model training is approaching convergence.

[0089] In one embodiment, steps 401-404 are described as follows:

[0090] Step 401: Construct a correlation matrix between the target control parameters, hydrogen production process parameters, and purification process parameters.

[0091] Specifically, the intelligent control system first determines the target control parameter set CIC = {cic1, cic2, ..., cic...} i The set of hydrogen production process parameters H2 = {H21, H22, ..., H2} j The set of purification process parameters PC = {PC1, PC2, ..., PC} l Then, through statistical analysis of historical production data, knowledge of process principles, and experimental design, the correlation between various parameters was determined. For example, the synchronous changes of target control parameters (such as hydrogen purity and yield) with hydrogen production process parameters (such as reaction temperature, pressure, and catalyst dosage) and purification process parameters (such as adsorbent temperature and regeneration time) in a large amount of historical data were analyzed, and a correlation matrix A was constructed. g The correlation matrix A g It is an (i*(j+l)) matrix with elements Indicates the target control parameter cic γ The degree of correlation with process parameters (hydrogen production or purification process parameters) δ, with values ​​ranging from [-1, 1]. Positive values ​​indicate positive correlation, negative values ​​indicate negative correlation, and the larger the absolute value, the stronger the correlation.

[0092] Step 402: Based on the correlation matrix, match the data in the target control parameters with the data in the purification process parameters and the hydrogen production process parameters to obtain a parameter correspondence list.

[0093] Specifically, the intelligent control system traverses the association matrix A g For each target control parameter cic γ Identify the process parameters with a large absolute value of correlation and preset a management threshold ε, such as ε = 0.5, when |A gγδ When |>ε, confirm the target control parameter cicγ It has a strong correlation with the process parameter δ. Then, the associated target control parameter and process parameter are paired to form a correspondence pair. All these correspondence pairs constitute the parameter correspondence list L. c If the target control parameter cic1 is strongly correlated with the hydrogen production process parameter H22 and the purification process parameter PC3, then list L c It includes correspondences such as (cic1, H22) and (cic1, PC3).

[0094] Step 403: Evaluate the influence of each process parameter on the target control parameter based on the parameter correspondence list, and obtain the sensitivity coefficient of each process parameter.

[0095] Specifically, the intelligent control system uses the parameter correspondence list L c Each parameter pair in (cic) γ ,xx δ (xx) δ (This refers to process parameters for hydrogen production or purification), analyzed using a local sensitivity method. With other process parameters kept constant, the process parameter xx is analyzed. δ Make a small perturbation Δxx δ Observe the target control parameter CIC γ The change Δcic γ Then the sensitivity coefficient Indicates process parameter xx δ When the relative change is one unit, the target control parameter cic γ The relative change. Then, through multiple experiments or numerical simulations based on historical data, the sensitivity coefficient of each process parameter to its associated target control parameter is calculated.

[0096] Step 404: Based on the process principles of purification and hydrogen production, a qualitative analysis is performed on the relationship between the target control parameters and process parameters to obtain auxiliary analysis results for each process parameter.

[0097] Specifically, the intelligent control system analyzes the direction and mechanism of influence of each process parameter on the target control parameter from the perspective of process principles such as chemical kinetics and thermodynamics. For example, in the hydrogen production process, according to the principles of chemical reaction kinetics, increasing the reaction temperature usually accelerates the reaction rate, thereby increasing hydrogen production, but may affect the purity of hydrogen. Therefore, for each process parameter, based on process principles and actual production experience, a qualitative analysis conclusion is given, such as "increasing the reaction temperature will increase hydrogen production, but may reduce hydrogen purity." Ultimately, the given conclusions constitute the auxiliary analysis results for each process parameter.

[0098] Step 405: Adjust the purification process parameters and hydrogen production process parameters based on the auxiliary analysis results of each process parameter, the sensitivity coefficient of each process parameter, and the target control parameters.

[0099] Specifically, the intelligent control system provides control parameters cic for each target parameter. γ According to its expected adjustment value Δcic γ ,tar, combined with the sensitivity coefficient SS of the associated process parameters γδ Based on the auxiliary analysis results, determine the direction and amount of adjustment for process parameters.

[0100] Furthermore, process parameters with high sensitivity coefficients should be adjusted first. Simultaneously, auxiliary analysis results should be consulted to avoid adverse effects on other target control parameters due to adjustments in one process parameter. For example, when increasing hydrogen production is required, for process parameters with high sensitivity coefficients where increasing the parameter according to process principle analysis (such as reaction temperature) will increase hydrogen production, their values ​​should be appropriately increased; however, it is important to observe whether this will negatively impact other target control parameters such as hydrogen purity. The target control parameters should be gradually adjusted to approach the desired values ​​through iterative adjustments. After each adjustment, the actual changes in the target control parameters should be reassessed, and the process parameters should be further adjusted based on the feedback results.

[0101] The embodiments of the present invention comprehensively consider the correlation between parameters, the degree of influence, and the process principle, making parameter adjustment more scientific and comprehensive. By combining quantitative analysis of sensitivity coefficients with qualitative analysis of process principles, the adjustment direction and amount of process parameters can be accurately determined, thereby improving the adjustment accuracy of target control parameters.

[0102] In one embodiment, steps 501-503 are described as follows:

[0103] Step 501: Align the real-time hydrogen data with the target hydrogen data according to the time series to obtain a time series list.

[0104] Specifically, the intelligent control system first checks the timestamp formats of the real-time hydrogen data and the target hydrogen data to ensure they are consistent. If the timestamp formats are different, a format conversion is performed to unify the different precision time representations into second-level timestamps.

[0105] Furthermore, the intelligent control system can use methods such as linear interpolation or spline interpolation to perform time alignment processing on the data. For example, if the real-time hydrogen data sequence is... The corresponding timestamp sequence is The target hydrogen data sequence is The corresponding timestamp sequence is Based on real-time data timestamps, for target data, if at a certain real-time data time point tr δ If there is no corresponding target data time point, the target data value for that time point is calculated through interpolation. For example, the linear interpolation formula is: If th γ <tr δ <th γ+1 , then tr δ Target data interpolation at time step (hq γ+1 -hq γ The aligned real-time hydrogen data and target hydrogen data are arranged in chronological order to form a time series list.

[0106] Step 502: Calculate the deviation index between real-time hydrogen data and target hydrogen data based on the time series list.

[0107] Specifically, deviation indicators include absolute deviation indicators and relative deviation indicators. The intelligent control system first calculates the absolute deviation indicator, such as the mean absolute deviation (MAD). Assuming the time series list length is N, then... in This indicates real-time hydrogen data. This indicates the target hydrogen data. For relative deviation indicators, such as mean relative deviation (MRD), then...

[0108] Step 503: Compare and analyze the deviation index with the preset deviation threshold to determine whether it exceeds the deviation threshold.

[0109] Specifically, the preset deviation threshold of the intelligent control system is determined based on factors such as the precision requirements of the hydrogen production process, production targets, and historical experience. For example, for the indicator of hydrogen purity, an absolute deviation threshold Δ is set according to product quality standards. abs and relative deviation threshold Δ rel The calculated deviation index is compared with a preset threshold. If the calculated mean absolute deviation (MAD) > Δ... abs Or the mean relative deviation MRD > Δ rel If the real-time hydrogen data exceeds the deviation threshold, then it is determined that the real-time hydrogen data exceeds the deviation threshold.

[0110] This invention ensures data comparison on the same time scale through time alignment, comprehensively measures deviation through multiple deviation indicators, and improves the accuracy of judging the deviation between real-time hydrogen data and target hydrogen data by combining reasonable threshold judgment. In addition, a complete process is formed from data processing to deviation calculation and threshold judgment, with each step cooperating with each other to enhance the reliability of the judgment results and provide a solid basis for whether to adjust process parameters in the future.

[0111] Optional, refer to Figure 2 , Figure 2 This is a schematic diagram of the intelligent control system for methanol cracking to hydrogen production based on big data provided by the present invention. The intelligent control system for methanol cracking to hydrogen production based on big data includes:

[0112] Acquisition unit 210 is used to acquire hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data;

[0113] Prediction unit 220 is used to predict target hydrogen data within a preset time period based on hydrogen demand data;

[0114] The control parameter determination unit 230 is used to input the target hydrogen data, methanol feedstock concentration data, and catalyst parameter data into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from the sample data and its corresponding control parameter label results.

[0115] Output unit 240 is used to adjust purification process parameters and hydrogen production process parameters based on target control parameters;

[0116] The feedback adjustment unit 250 is used to acquire real-time hydrogen data based on the adjusted hydrogen production equipment, compare and analyze the real-time hydrogen data with the target hydrogen data, determine whether the deviation threshold is exceeded, and if the deviation threshold is exceeded, adjust the purification process parameters and hydrogen production process parameters in real time.

[0117] This invention, through precise prediction of target hydrogen data within a preset time period, determines the control parameters for the purification and hydrogen production processes based on the predicted target hydrogen data, methanol feedstock parameters, and catalyst parameters. This initially achieves intelligent control of cracking hydrogen production under a production-to-order model, avoiding waste caused by excessive purification of methanol feedstock and excessive hydrogen production. Furthermore, the real-time hydrogen production data is monitored and compared with the target hydrogen data to determine if any deviations occur. If deviations are found, the parameters of the purification and hydrogen production processes are adjusted accordingly, further achieving intelligent quantitative control of cracking hydrogen production and minimizing resource waste and high production costs.

[0118] Please see Figure 3 , Figure 3 An embodiment diagram of an electronic device provided in accordance with the present invention. For example... Figure 3 As shown, an embodiment of the present invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it performs the following steps:

[0119] Obtain hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data;

[0120] Predict target hydrogen data for a preset time period based on hydrogen demand data;

[0121] The target hydrogen data, methanol feedstock concentration data, and catalyst parameter data are input into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from the sample data and its corresponding control parameter label results.

[0122] Adjust the purification process parameters and hydrogen production process parameters based on the target control parameters;

[0123] Based on the real-time hydrogen data obtained from the adjusted hydrogen production equipment, the real-time hydrogen data is compared and analyzed with the target hydrogen data to determine whether it exceeds the deviation threshold. If it exceeds the deviation threshold, the purification process parameters and hydrogen production process parameters are adjusted in real time.

[0124] Please see Figure 4 , Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with an embodiment of the present invention is shown. Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it performs the following steps:

[0125] Obtain hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data;

[0126] Predict target hydrogen data for a preset time period based on hydrogen demand data;

[0127] The target hydrogen data, methanol feedstock concentration data, and catalyst parameter data are input into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from the sample data and its corresponding control parameter label results.

[0128] Adjust the purification process parameters and hydrogen production process parameters based on the target control parameters;

[0129] Based on the real-time hydrogen data obtained from the adjusted hydrogen production equipment, the real-time hydrogen data is compared and analyzed with the target hydrogen data to determine whether it exceeds the deviation threshold. If it exceeds the deviation threshold, the purification process parameters and hydrogen production process parameters are adjusted in real time.

[0130] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute a big data-based intelligent control method for methanol cracking to hydrogen production provided by the above methods, the method comprising:

[0131] Obtain hydrogen demand data, methanol feedstock parameter data, and catalyst parameter data;

[0132] Predict target hydrogen data for a preset time period based on hydrogen demand data;

[0133] The target hydrogen data, methanol feedstock concentration data, and catalyst parameter data are input into the hydrogen production optimization process model to obtain the target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained from the sample data and its corresponding control parameter label results.

[0134] Adjust the purification process parameters and hydrogen production process parameters based on the target control parameters;

[0135] Based on the real-time hydrogen data obtained from the adjusted hydrogen production equipment, the real-time hydrogen data is compared and analyzed with the target hydrogen data to determine whether it exceeds the deviation threshold. If it exceeds the deviation threshold, the purification process parameters and hydrogen production process parameters are adjusted in real time.

[0136] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0137] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0138] 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 foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A big data-based intelligent control method for hydrogen production by methanol cracking, characterized in that, The method comprises the following steps: obtaining hydrogen demand data, methanol raw material parameter data and catalyst parameter data; predicting target hydrogen data in a preset time period based on the hydrogen demand data; inputting the target hydrogen data, methanol raw material concentration data and catalyst parameter data into a hydrogen production optimization process model to obtain target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained by sample data and corresponding control parameter label results; adjusting purification process parameters and hydrogen production process parameters based on the target control parameters; comprising: constructing a correlation matrix of the target control parameters, the hydrogen production process parameters and the purification process parameters; based on the correlation matrix, finding process parameters with an absolute correlation degree greater than a preset management threshold with the target control parameters, and one-to-one corresponding data in the target control parameters to data in the purification process parameters and the hydrogen production process parameters to obtain a parameter correspondence list; based on the parameter correspondence list, evaluating the influence degree of each process parameter on the target control parameter to obtain a sensitivity coefficient of each process parameter, and the sensitivity coefficient formula is as follows: ； wherein, is expressed as a sensitivity coefficient, is expressed as a target control parameter, is expressed as a target control parameter is expressed as a change in the amount of, is expressed as a hydrogen production or purification process parameter, is expressed as a process parameter a small perturbation value; based on the process principles of the purification process and the hydrogen production process, qualitatively analyzing the relationship between the target control parameters and the process parameters to obtain an auxiliary analysis result of each process parameter; the auxiliary analysis result is the influence direction and mechanism of each process parameter on the target control parameter; adjusting the purification process parameters and the hydrogen production process parameters based on the auxiliary analysis result of each process parameter, the sensitivity coefficient of each process parameter and the target control parameter; based on the adjusted hydrogen production equipment, obtaining real-time hydrogen data, and comparing and analyzing the real-time hydrogen data with the target hydrogen data to determine whether the deviation threshold is exceeded, and if the deviation threshold is exceeded, the purification process parameters and the hydrogen production process parameters are adjusted in real time.

2. The big data based intelligent control method for hydrogen production by methanol cracking according to claim 1, characterized in that, The hydrogen demand data includes historical hydrogen production data, historical hydrogen sales data, current purchase data and industrial policy data, and the target hydrogen data in a preset time period is predicted based on the hydrogen demand data, comprising: matching the historical hydrogen production data and the historical hydrogen sales data based on time series to obtain a hydrogen sales and production data set; segmenting the hydrogen sales and production data set based on a preset production cycle interval to obtain a plurality of inventory accumulation rates; based on the change rule of the plurality of inventory accumulation rates, predicting the current inventory accumulation rate, and determining the fault tolerance rate in the current production cycle based on the current inventory accumulation rate; determining a policy influence index based on the influence degree of the industrial policy data on hydrogen demand; fusing the current purchase data, the fault tolerance rate in the current production cycle and the policy influence index to obtain the target hydrogen data. 3.The big data-based intelligent control method for hydrogen production by methanol cracking according to claim 2, characterized in that, based on the change rule of the plurality of inventory accumulation rates, predicting the current inventory accumulation rate, and determining the fault tolerance rate in the current production cycle based on the current inventory accumulation rate, comprising: based on the change rule of the plurality of inventory accumulation rates, determining a change trend, and predicting the current inventory accumulation rate based on the change trend; acquire a backup amount of a current production cycle, and determine a current backup rate based on the backup amount; correct the current inventory backlog rate and the current backup rate based on seasonal factors respectively to obtain a target inventory backlog rate and a target backup rate; compare the target inventory backlog rate with the target backup rate to determine whether the target inventory backlog rate is greater than or equal to the target backup rate, and if so, determine the previous inventory backlog rate as a fault tolerance rate in the current production cycle; or if the target inventory backlog rate is less than the target backup rate, determine the target backup rate as the fault tolerance rate in the current production cycle. 4.The big data-based intelligent control method for hydrogen production by methanol cracking according to claim 2, characterized in that, The policy influence index is determined based on the influence degree of the industrial policy data on hydrogen demand, including: performing semantic analysis and classification on the industrial policy data to obtain a plurality of sub-policies; comparing each of the sub-policies with a policy quantification database to obtain a quantification score corresponding to each of the sub-policies; adding the quantification scores of the plurality of sub-policies to obtain a comprehensive influence quantification value of the current policy on hydrogen demand; judging the comprehensive influence quantification value based on a preset reference value to obtain the policy influence index.

5. The big data based intelligent control method for hydrogen production by methanol cracking according to claim 1, wherein, The training steps of the hydrogen production optimization process model include: acquiring sample data; the sample data includes target hydrogen data, methanol raw material concentration data, and catalyst parameter data; inputting the sample data into a pre-training model to obtain a prediction result output by the pre-training model; obtaining a loss value based on a loss function of the pre-training model according to the prediction result; adjusting an optimization function in the pre-training model based on the loss value; predicting the prediction result of the sample data based on the adjusted pre-training module, until the loss value output by the loss function is greater than zero and less than or equal to a first preset value, and at least five consecutive loss values successively decrease and the difference between adjacent loss values is less than or equal to a second preset value.

6. The big data based intelligent control method for hydrogen production by methanol cracking according to claim 1, wherein, The comparison and analysis of the real-time hydrogen data and the target hydrogen data to determine whether the deviation threshold is exceeded includes: aligning the real-time hydrogen data and the target hydrogen data according to a time sequence to obtain a time sequence list; calculating a deviation index between the real-time hydrogen data and the target hydrogen data based on the time sequence list; comparing and analyzing the deviation index with a preset deviation threshold to determine whether the deviation threshold is exceeded.

7. A big data-based intelligent control system for methanol cracking hydrogen production, applied to a big data-based intelligent control method for methanol cracking hydrogen production according to any one of claims 1 to 6; the big data-based intelligent control system for methanol cracking hydrogen production includes: an acquisition unit for acquiring hydrogen demand data, methanol raw material parameter data, and catalyst parameter data; a prediction unit for predicting target hydrogen data in a preset time period based on the hydrogen demand data; a control parameter determination unit for inputting the target hydrogen data, methanol raw material concentration data, and catalyst parameter data into a hydrogen production optimization process model to obtain target control parameters output by the hydrogen production optimization process model; the hydrogen production optimization process model is trained by sample data and corresponding control parameter label results. An output unit is configured to adjust the purification process parameters and the hydrogen production process parameters based on the target control parameter; and the output unit comprises: An association matrix of the target control parameter, the hydrogen production process parameters and the purification process parameters is constructed; Based on the association matrix, process parameters associated with the target control parameter with an absolute value greater than a preset management threshold are found, and data in the target control parameter and data in the purification process parameters and the hydrogen production process parameters are one-to-one corresponding to obtain a parameter corresponding list; Based on the parameter corresponding list, an influence degree of each process parameter on the target control parameter is evaluated to obtain a sensitivity coefficient of each process parameter, and the sensitivity coefficient formula is as follows: ； wherein, is expressed as a sensitivity coefficient, is expressed as a target control parameter, is expressed as a target control parameter a change in, is expressed as a hydrogen production or purification process parameter, is expressed as a process parameter a small perturbation value; Based on the process principles of the purification process and the hydrogen production process, a qualitative analysis of the relationship between the target control parameter and the process parameters is performed to obtain an auxiliary analysis result of each process parameter; the auxiliary analysis result is an influence direction and an influence mechanism of each process parameter on the target control parameter; Based on the auxiliary analysis result of each process parameter, the sensitivity coefficient of each process parameter and the target control parameter, the purification process parameters and the hydrogen production process parameters are adjusted; A feedback adjustment unit is configured to obtain real-time hydrogen data based on the adjusted hydrogen production equipment, and compare and analyze the real-time hydrogen data with the target hydrogen data to determine whether the deviation threshold is exceeded, and if the deviation threshold is exceeded, the purification process parameters and the hydrogen production process parameters are adjusted in real time.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the intelligent control method for methanol cracking hydrogen production based on big data according to any one of claims 1 to 6. 9.A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the intelligent control method for methanol cracking hydrogen production based on big data according to any one of claims 1 to 6.

Images