Methods, devices and equipment for predicting sulfur dioxide emission concentrations from catalytic regeneration flue gas
By establishing a prediction model for the sulfur content of raw materials and the concentration of sulfur dioxide, the problem of the inability to judge the fluctuation of emission concentration in catalytic regeneration flue gas in a timely manner was solved, and the accurate prediction of sulfur dioxide concentration in regenerated flue gas was achieved, supporting the real-time adjustment of flue gas treatment facilities and improving the treatment effect and efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-18
- Publication Date
- 2026-03-10
AI Technical Summary
In existing technologies, fluctuations in sulfur dioxide emission concentrations from catalytic regeneration flue gas cannot be detected in a timely manner, leading to lags in the operation and adjustment of treatment facilities and making it easy for emissions to exceed standards.
By establishing a prediction model for the sulfur content of raw materials and a prediction model for the sulfur dioxide concentration, and by using a neural network to train the catalytic cracking reaction products and process parameters, the prediction of the sulfur content of the catalytic cracking reaction raw materials and the sulfur dioxide concentration of the regenerated flue gas can be achieved.
It enables accurate prediction of sulfur dioxide concentration in regenerated flue gas, supports real-time adjustment of flue gas treatment facilities, and improves treatment effectiveness and efficiency.
Smart Images
Figure CN117467464B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of regenerated flue gas treatment technology, specifically to a method for predicting sulfur dioxide emission concentration in catalytic regeneration flue gas, a device for predicting sulfur dioxide emission concentration in catalytic regeneration flue gas, and a terminal device. Background Technology
[0002] Catalytic cracking (FCC) is a crucial secondary crude oil processing technology, producing nearly 70% of gasoline and 30%–40% of diesel fuel. The catalytic cracking unit is a major source of emissions from refining operations, accounting for over 40% of the total emissions from the main refining units. However, with the increasing prominence of heavy and inferior crude oil and the growing market demand for light oil products, inferior feedstocks such as residual oil, coking wax oil, and shale oil are increasingly becoming part of FCC processing feedstocks to expand their availability. These inferior feedstocks often have high sulfur content, resulting in high and fluctuating sulfur dioxide concentrations in the catalytic regeneration flue gas. If treatment facilities are not timely in adjusting these concentrations, sulfur dioxide emissions can easily exceed standards.
[0003] To address the significant fluctuations in sulfur dioxide emission concentration at the outlet of catalyst regenerators, most current methods simply set the operating parameters and desulfurizing agent dosage of waste gas treatment facilities based on the maximum pollutant concentration. However, this approach fails to promptly detect sudden increases in pollutant concentrations, resulting in relatively delayed adjustments to the treatment facilities' operation and making it easy for sulfur dioxide concentrations at the outlet of the treatment facilities to exceed emission standards. Summary of the Invention
[0004] The purpose of this invention is to provide a method for predicting sulfur dioxide emission concentration in catalytic regeneration flue gas, a device for predicting sulfur dioxide emission concentration in catalytic regeneration flue gas, and a terminal device, so as to solve the problem that existing flue gas treatment facilities cannot determine the fluctuation of pollutant concentration in regenerated flue gas.
[0005] To achieve the above objectives, in a first aspect of the present invention, a method for predicting sulfur dioxide emission concentration from catalytic regeneration flue gas is provided, comprising:
[0006] The sulfur content of the catalytic cracking reaction product is obtained. Using the sulfur content of the catalytic cracking reaction product as input, the sulfur content of the catalytic cracking reaction feedstock is predicted by the feedstock sulfur content prediction model. The catalytic cracking reaction product is obtained by the catalytic cracking reaction feedstock through a catalytic cracking unit after catalytic cracking reaction.
[0007] The characteristic parameters of the catalytic cracking reaction are determined. Using the characteristic parameters and the sulfur content of the catalytic cracking reaction feedstock as inputs, the concentration of sulfur dioxide in the regenerated flue gas is predicted by a sulfur dioxide concentration prediction model.
[0008] The raw material sulfur content prediction model is obtained by training a first preset neural network with the sulfur content of catalytic cracking reaction products and the corresponding sulfur content of catalytic cracking reaction feedstock under different operating conditions. The sulfur dioxide concentration prediction model is obtained by training a second preset neural network with the sulfur content, characteristic parameters and corresponding sulfur dioxide concentration in regenerated flue gas under different operating conditions.
[0009] Optionally, the catalytic cracking reaction products include: dry gas, liquefied petroleum gas, gasoline, diesel, and sulfur-containing wastewater; the training process of the feedstock sulfur content prediction model includes:
[0010] Determine the initial network parameters of the first preset neural network;
[0011] Using the sulfur content of dry gas, liquefied petroleum gas, gasoline, diesel, and sulfur-containing wastewater obtained under different operating conditions as inputs, the sulfur content of the catalytic cracking reaction feedstock corresponding to the current operating condition is predicted by the first preset neural network.
[0012] The measured sulfur content of the catalytic cracking feedstock corresponding to the current operating condition is obtained. If the difference between the obtained sulfur content of the catalytic cracking feedstock and the measured sulfur content of the catalytic cracking feedstock is greater than the sulfur content threshold, the network parameters of the first preset neural network are updated until the difference between the obtained sulfur content of the catalytic cracking feedstock and the measured sulfur content of the catalytic cracking feedstock is not greater than the sulfur content threshold. The first preset neural network with updated network parameters is used as the feedstock sulfur content prediction model.
[0013] Optionally, characteristic parameters of the catalytic cracking reaction are determined, including:
[0014] Obtain the process parameters of the catalytic cracking reaction, and determine the process parameters as characteristic parameters of the catalytic cracking reaction;
[0015] The process parameters include the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas in the catalytic cracking unit.
[0016] Optionally, the training process of the sulfur dioxide concentration prediction model includes:
[0017] Determine the initial network parameters of the second preset neural network;
[0018] Using the sulfur content of the catalytic cracking feedstock, the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of stripping material in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas obtained under different operating conditions as inputs, the concentration of sulfur dioxide in the regenerated flue gas corresponding to the current operating condition is predicted by the second preset neural network.
[0019] The measured concentration of sulfur dioxide in the regenerated flue gas corresponding to the current operating condition is obtained. If the difference between the obtained concentration of sulfur dioxide in the regenerated flue gas and the measured concentration of sulfur dioxide in the regenerated flue gas is greater than the sulfur dioxide concentration threshold, the network parameters of the second preset neural network are updated until the difference between the obtained concentration of sulfur dioxide in the regenerated flue gas and the measured concentration of sulfur dioxide in the regenerated flue gas is no greater than the sulfur dioxide concentration threshold. The second preset neural network with updated network parameters is used as the sulfur dioxide concentration prediction model.
[0020] Optionally, characteristic parameters of the catalytic cracking reaction are determined, including:
[0021] The process parameters of the catalytic cracking reaction are obtained, and the sub-characteristic parameters of the catalytic cracking reaction are determined based on the sulfur content of the feedstock and the process parameters.
[0022] The process parameters and sub-characteristic parameters of the catalytic cracking reaction are determined to be characteristic parameters of the catalytic cracking reaction;
[0023] The process parameters include the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas in the catalytic cracking unit.
[0024] Optionally, the sub-characteristic parameters of the catalytic cracking reaction include:
[0025] The ratio of the sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas.
[0026] Optionally, the sub-characteristic parameters of the catalytic cracking reaction further include:
[0027] And the ratio of the amount of stripping steam to the amount of steam stored in the stripping section of the reactor.
[0028] In a second aspect of the present invention, a device for predicting the sulfur dioxide emission concentration of catalytic regeneration flue gas is provided, comprising:
[0029] The data acquisition module is configured to acquire the sulfur content of the catalytic cracking reaction products and determine the characteristic parameters of the catalytic cracking reaction.
[0030] The prediction module is configured to predict the sulfur content of the catalytic cracking feedstock using a feedstock sulfur content prediction model, taking the sulfur content of the catalytic cracking reaction products as input. The catalytic cracking reaction products are obtained by catalytic cracking of the catalytic cracking feedstock in a catalytic cracking unit.
[0031] Using the aforementioned characteristic parameters and the sulfur content of the catalytic cracking reaction feedstock as inputs, the concentration of sulfur dioxide in the regenerated flue gas is predicted by a sulfur dioxide concentration prediction model.
[0032] The raw material sulfur content prediction model is obtained by training a first preset neural network with the sulfur content of catalytic cracking reaction products and the corresponding sulfur content of catalytic cracking reaction feedstock under different operating conditions. The sulfur dioxide concentration prediction model is obtained by training a second preset neural network with the sulfur content, characteristic parameters and corresponding sulfur dioxide concentration in regenerated flue gas under different operating conditions.
[0033] Optionally, the device further includes a feature parameter determination module; the feature parameter determination module is configured to:
[0034] The process parameters of the catalytic cracking reaction are obtained, and the sub-characteristic parameters of the catalytic cracking reaction are determined based on the sulfur content of the feedstock and the process parameters.
[0035] The process parameters and sub-characteristic parameters of the catalytic cracking reaction are determined to be characteristic parameters of the catalytic cracking reaction;
[0036] The process parameters include the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas in the catalytic cracking unit.
[0037] Optionally, the sub-characteristic parameters of the catalytic cracking reaction include:
[0038] The ratio of the sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas.
[0039] Optionally, the sub-characteristic parameters of the catalytic cracking reaction further include:
[0040] The ratio of the amount of stripping steam to the amount of steam stored in the stripping section of the reactor.
[0041] In a third aspect of the invention, a computer-readable medium is provided, the computer-readable medium storing a computer program that, when executed by a processor, implements the above-described method for predicting sulfur dioxide emission concentrations in catalytic regeneration flue gas.
[0042] In a fourth aspect of the invention, a terminal device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for predicting sulfur dioxide emission concentrations from catalytic regeneration flue gas.
[0043] The above-mentioned technical solution of the present invention achieves soft measurement of the sulfur content of the feedstock by establishing a predictive model of the sulfur content of the catalytic cracking reaction feedstock, and uses the predicted sulfur content of the feedstock and the determined characteristic parameters that are strongly correlated with the sulfur dioxide concentration of the regenerated flue gas to predict the sulfur dioxide emission concentration of the regenerated flue gas, thereby facilitating the operation and adjustment of the flue gas treatment facility.
[0044] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description
[0045] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:
[0046] Figure 1 This is a flowchart of a method for predicting sulfur dioxide emission concentration in catalytic regeneration flue gas according to a preferred embodiment of the present invention;
[0047] Figure 2 This is a schematic block diagram of a catalytic regeneration flue gas sulfur dioxide emission concentration prediction device provided by a preferred embodiment of the present invention;
[0048] Figure 3 This is a schematic block diagram of a terminal device provided by a preferred embodiment of the present invention.
[0049] Explanation of reference numerals in the attached figures
[0050] 10 - Terminal device, 100 - Processor, 101 - Memory, 102 - Computer program. Detailed Implementation
[0051] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.
[0052] like Figure 1 As shown, the first aspect of this embodiment provides a method for predicting the sulfur dioxide emission concentration of catalytic regeneration flue gas, including:
[0053] The sulfur content of the catalytic cracking reaction products is obtained. Using the sulfur content of the catalytic cracking reaction products as input, the sulfur content of the feedstock for catalytic cracking reaction is predicted by the feedstock sulfur content prediction model. The catalytic cracking reaction products are obtained by passing the feedstock for catalytic cracking reaction through a catalytic cracking unit.
[0054] The characteristic parameters of the catalytic cracking reaction are determined. Using the characteristic parameters and the sulfur content of the catalytic cracking feedstock as inputs, the concentration of sulfur dioxide in the regenerated flue gas is predicted by a sulfur dioxide concentration prediction model.
[0055] The feedstock sulfur content prediction model is obtained by training a first preset neural network with the sulfur content of catalytic cracking reaction products and the corresponding sulfur content of catalytic cracking reaction feedstock under different operating conditions. The sulfur dioxide concentration prediction model is obtained by training a second preset neural network with the sulfur content of catalytic cracking reaction feedstock, characteristic parameters, and the corresponding sulfur dioxide concentration in regenerated flue gas under different operating conditions.
[0056] Thus, this embodiment achieves soft measurement of the sulfur content of the feedstock by establishing a predictive model of the sulfur content of the catalytic cracking reaction feedstock, and uses the predicted sulfur content of the feedstock and the determined characteristic parameters that are strongly correlated with the sulfur dioxide concentration of the regenerated flue gas to predict the sulfur dioxide emission concentration of the regenerated flue gas, thereby facilitating the operation and adjustment of the flue gas treatment facility.
[0057] Catalytic cracking is a petroleum refining process that involves the cracking of heavy oil under the action of heat and catalyst, transforming it into cracked gas, gasoline, and diesel. Catalytic cracking units typically include a regeneration system, a fractionation system, and an absorption-stabilization system. During catalytic cracking, the generated regenerated flue gas is discharged from the top outlet of the regenerator tower and then enters the flue gas treatment facility for further treatment. However, existing flue gas treatment facilities cannot determine the sulfur dioxide content of the regenerated flue gas at the inlet. The operating parameters and desulfurizer dosage of the flue gas treatment facility are usually fixed at a certain concentration. When the sulfur dioxide content in the regenerated flue gas flues, such as a sudden increase, the operating parameters of the treatment facility cannot be dynamically adjusted according to the fluctuations in the sulfur dioxide content, thus affecting the treatment effect and efficiency of the regenerated flue gas.
[0058] Since the total sulfur content of the feedstock for catalytic cracking is difficult to accurately detect before the catalytic cracking reaction, while the sulfur content of the products can be accurately detected, this embodiment establishes a feedstock sulfur content prediction model based on a neural network to characterize the correlation between the sulfur content of the products and the sulfur content of the feedstock. This enables soft measurement of the total sulfur content in the feedstock based on the sulfur content of the products. Furthermore, this embodiment uses a neural network-based sulfur dioxide concentration prediction model to predict the concentration of sulfur dioxide in the regenerated flue gas, based on the correlation between the sulfur content of the feedstock and relevant characteristic parameters of the catalytic cracking reaction, such as process parameters, and the concentration of sulfur dioxide in the regenerated flue gas. The products of the catalytic cracking reaction include, but are not limited to, dry gas, liquefied petroleum gas, gasoline, diesel, and sulfur-containing wastewater. The first and second preset neural networks can be, but are not limited to, convolutional neural networks, backpropagation neural networks, etc.
[0059] In this embodiment, both the first and second preset neural networks use backpropagation (BP) neural networks, and the training process of the raw material sulfur content prediction model includes:
[0060] Determine the initial network parameters of the first preset neural network;
[0061] Using the sulfur content of dry gas, liquefied petroleum gas, gasoline, diesel, and sulfur-containing wastewater obtained under different operating conditions as inputs, the sulfur content of the catalytic cracking reaction feedstock corresponding to the current operating condition is predicted by the first preset neural network.
[0062] The measured sulfur content of the catalytic cracking feedstock corresponding to the current operating condition is obtained. If the difference between the obtained sulfur content of the catalytic cracking feedstock and the measured sulfur content of the catalytic cracking feedstock is greater than the sulfur content threshold, the network parameters of the first preset neural network are updated until the difference between the obtained sulfur content of the catalytic cracking feedstock and the measured sulfur content of the catalytic cracking feedstock is no greater than the sulfur content threshold. The first preset neural network with updated network parameters is used as the feedstock sulfur content prediction model.
[0063] Specifically, training and validation sets are constructed using the sulfur content of dry gas, liquefied petroleum gas, gasoline, diesel, and sulfur-containing wastewater obtained under different operating conditions, as well as the sulfur content of the raw materials corresponding to each operating condition. Based on the constructed training and validation sets, the neural network is trained through the following steps:
[0064] Step S1: Set the initial weight values k and b of the BP neural network.
[0065] Step S2: Calculate the hidden layer values by weighted summation of the input layer nodes of the BP neural network using the following formula: Ff,j=∑n n·knj;
[0066] Where j is the numerical sequence number of the hidden layer, and n is the process parameter under various operating conditions, namely the sulfur content in dry gas, liquefied gas, gasoline, diesel and sulfur content in sulfur-containing wastewater.
[0067] Step S3: Execute the sigmoid activation formula:
[0068] Step S4: Calculate the predicted sulfur content of the catalytic cracking feedstock using the following formula:
[0069]
[0070] Step S5: Calculate the error between the predicted value and the actual value. Convergence occurs when the error approaches zero.
[0071] Step S6: If the error value does not meet the convergence criterion, then the weight values k and b of the BP neural network are calculated and updated in reverse using the following formula:
[0072]
[0073]
[0074] Where X is the set learning rate.
[0075] Step S7: Repeat steps S1 to S6 using the updated weight values until convergence, and obtain the trained raw material sulfur content prediction model.
[0076] In a specific example of this embodiment, determining the characteristic parameters of the catalytic cracking reaction includes: obtaining the process parameters of the catalytic cracking reaction and determining the process parameters as the characteristic parameters of the catalytic cracking reaction; the process parameters include the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of stripping material in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas.
[0077] The training process for the sulfur dioxide concentration prediction model includes:
[0078] Determine the initial network parameters of the second preset neural network;
[0079] Using the sulfur content of the catalytic cracking feedstock, the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas obtained under different operating conditions as inputs, the concentration of sulfur dioxide in the regenerated flue gas corresponding to the current operating condition is predicted by the second preset neural network.
[0080] The measured concentration of sulfur dioxide in the regenerated flue gas corresponding to the current operating condition is obtained. If the difference between the obtained concentration of sulfur dioxide in the regenerated flue gas and the measured concentration of sulfur dioxide in the regenerated flue gas is greater than the sulfur dioxide concentration threshold, the network parameters of the second preset neural network are updated until the difference between the obtained concentration of sulfur dioxide in the regenerated flue gas and the measured concentration of sulfur dioxide in the regenerated flue gas is no greater than the sulfur dioxide concentration threshold. The second preset neural network with updated network parameters is used as the sulfur dioxide concentration prediction model.
[0081] Specifically, training and validation sets are constructed based on the sulfur content of the catalytic cracking feedstock, the oxygen content and carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas obtained under different operating conditions. A backpropagation (BP) neural network is then trained using these sets to obtain a sulfur dioxide concentration prediction model. Understandably, the specific training steps for the sulfur dioxide concentration prediction model are the same as those for the feedstock sulfur content prediction model, and will not be repeated here.
[0082] In this embodiment, the method for constructing the training set and the test set is as follows:
[0083] Collect historical process parameter data of the catalytic cracking unit within a set time period, such as the sulfur content of historical catalytic cracking feedstock, oxygen content at the regenerator outlet, carbon monoxide content at the regenerator outlet, reactor stripping section reserves, stripping steam volume, and standard state flow rate of regenerated flue gas. Then, process the raw data through the following steps:
[0084] The collected historical data is preprocessed, including deleting data from periods of shutdown and extracting data from periods of normal operation.
[0085] Threshold calculations are performed on the process parameters after the previous cleaning step to determine the confidence interval of each process parameter, obtain an accurate value range, and adjust the value range of each process parameter in combination with the properties of each process parameter and their actual rated operating conditions to determine the value range of each process parameter.
[0086] Anomaly mining was performed on all data. By analyzing the sulfur dioxide concentration data at the inlet of the flue gas treatment facility, the extreme values of all data were obtained. The maximum and minimum values were analyzed in conjunction with all other process parameters to determine whether all extreme values were abnormal data. The abnormal data were removed, and the process parameters that caused the sulfur dioxide concentration fluctuation under normal operating conditions were retained. Through further analysis, process parameters that were strongly correlated with the sulfur dioxide concentration at the inlet of the flue gas treatment facility were obtained.
[0087] Critical parameter analysis is performed on the relevant process parameters processed in the previous step, and training and testing sets are constructed based on the identified critical parameters. For example, the above analysis shows that process parameters such as the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas in the catalytic cracking unit are highly correlated with the sulfur dioxide emission concentration at the inlet of the flue gas treatment facility. Therefore, these process parameters are identified as characteristic parameters of the catalytic cracking reaction.
[0088] In another specific example of this embodiment, in order to further improve the predicted value of sulfur dioxide concentration in regenerated flue gas, characteristic parameters of the catalytic cracking reaction are determined, including:
[0089] The process parameters of the catalytic cracking reaction are obtained, and the sub-characteristic parameters of the catalytic cracking reaction are determined based on the sulfur content of the feedstock and the process parameters. The process parameters and sub-characteristic parameters of the catalytic cracking reaction are determined as the characteristic parameters of the catalytic cracking reaction. The process parameters include the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of stripping material in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas.
[0090] To further improve prediction accuracy, the sub-characteristic parameters of the catalytic cracking reaction in this embodiment include: the ratio of sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas, and the ratio of stripping steam quantity to the amount of fuel in the reactor stripping section. The ratio of sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas is determined as the first sub-characteristic parameter, and the ratio of stripping steam quantity to the amount of fuel in the reactor stripping section is determined as the second sub-characteristic parameter. This embodiment, through analysis of the catalytic cracking process flow, determines that the ratio of sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas can effectively characterize the amount of sulfur dioxide generated; therefore, this ratio is used as the first sub-characteristic parameter. On the other hand, this embodiment, through analysis of the catalytic cracking process flow, determines that the ratio of stripping steam quantity to the amount of fuel in the reactor stripping section can effectively characterize the amount of stripping steam used per unit amount of catalyst fuel; therefore, this ratio can also be used as the second sub-characteristic parameter. Understandably, the first sub-characteristic parameter and the second sub-characteristic parameter can be used individually as one of the characteristic parameters of the catalytic cracking reaction. For example, the characteristic parameters of the catalytic cracking reaction can be the process parameters and the first sub-characteristic parameter, or they can be the process parameters and the second sub-characteristic parameter. To improve the accuracy of prediction, this embodiment uses the process parameters, the first sub-characteristic parameter, and the second sub-characteristic parameter of the catalytic cracking reaction as the characteristic parameters of the catalytic cracking reaction.
[0091] The sub-characteristic parameters of the catalytic cracking reaction can also be determined through the following steps:
[0092] Using the ratio of any two of the following as initial sub-characteristic parameters—sulfur content of catalytic cracking feedstock, oxygen content at regenerator outlet, carbon monoxide content at regenerator outlet, amount of stripping material in reactor stripping section, amount of stripping steam, and standard flow rate of regenerated flue gas—the Pearson correlation coefficient between the initial sub-characteristic parameter and the concentration of sulfur dioxide in regenerated flue gas was calculated.
[0093] Initial sub-characteristic parameters whose Pearson correlation coefficient with the concentration of sulfur dioxide in the regenerated flue gas is greater than the Pearson correlation coefficient threshold are determined as sub-characteristic parameters of the catalytic cracking reaction. Alternatively, the initial sub-characteristic parameters corresponding to the top n Pearson correlation coefficients (arranged from largest to smallest) are determined as sub-characteristic parameters of the catalytic cracking reaction. For example, if the ratio of the sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas is greater than the preset Pearson correlation coefficient threshold, then the ratio of the sulfur content of the catalytic cracking feedstock to the standard flow rate of the regenerated flue gas is determined as a sub-characteristic parameter of the catalytic cracking reaction, and so on.
[0094] Pearson correlation calculations of relevant parameters yielded the following results: the Pearson correlation coefficient between the sulfur dioxide concentration at the flue gas treatment facility inlet and the total feed rate was 0.132; the Pearson correlation coefficient between the sulfur dioxide concentration at the flue gas treatment facility inlet and the standard flow rate of the flue gas at the flue gas treatment facility inlet was 0.103; and the Pearson correlation coefficient between the sulfur dioxide concentration at the flue gas treatment facility inlet and the first sub-feature parameter was 0.233. Furthermore, the first sub-feature parameter more directly expresses the sulfur dioxide generation per unit feed rate. Using this feature parameter as the input parameter for the sulfur dioxide concentration prediction model can effectively improve the accuracy of predicting the sulfur dioxide emission concentration at the flue gas treatment facility inlet.
[0095] The second sub-feature parameter is the ratio of stripping steam volume to the amount of catalyst stored in the stripping section of the reactor, which represents the amount of stripping steam used per unit amount of catalyst. Through Pearson correlation calculations of relevant parameters, the following results are obtained: the Pearson correlation coefficient between the sulfur dioxide concentration at the flue gas treatment facility inlet and the stripping steam volume is 0.210; the Pearson correlation coefficient between the sulfur dioxide concentration at the flue gas treatment facility inlet and the amount of catalyst stored in the stripping section of the reactor is 0.16; and the Pearson correlation coefficient between the sulfur dioxide concentration at the flue gas treatment facility inlet and the second sub-feature parameter is 0.320. Moreover, this parameter can more intuitively represent the unit energy consumption of the catalyst. Using the feature parameter as the input parameter of the sulfur dioxide concentration prediction model can effectively improve the accuracy of predicting the sulfur dioxide emission concentration at the flue gas treatment facility inlet.
[0096] The following is a specific example to illustrate this implementation method:
[0097] First, over 20,000 raw data points from May 5th to May 31st, 2019, for a catalytic cracking unit were collected for data analysis and correlation analysis. The process parameters retained were: sulfur content of catalytic cracking reaction products (dry gas, liquefied gas, gasoline, diesel), sulfur content of sulfur-containing wastewater, oxygen content at the top outlet of the regenerator tower, carbon monoxide content at the top outlet of the regenerator tower, standard flow rate of flue gas at the inlet of the flue gas treatment facility, and the amount of stripping steam in the reactor stripping section.
[0098] Secondly, data was deleted according to the respective value ranges of the process parameters, and the values within the value range of the process parameters were retained. The data from the continuous period of May 1, 2019 to May 10, 2019 were divided into two parts: 3360 data points from May 1, 2019 to May 7, 2019 were used as model training data, and 1440 data points from May 8, 2019 to May 10, 2019 were used as model testing data.
[0099] Next, the sulfur content of the aforementioned products (dry gas, liquefied gas, gasoline, and diesel) and the sulfur content of sulfur-containing wastewater are used as two process parameters as inputs to the BP neural network model. That is, the number of neurons in the input layer of the BP neural network model is 5, the number of hidden layers is 2, the model structure of the BP neural network is 5-2-1, and the output is the total sulfur content of the raw materials, thus predicting the total sulfur content of the raw materials.
[0100] The aforementioned five process parameters—oxygen content at the top outlet of the regenerator tower, carbon monoxide content at the top outlet of the regenerator tower, standard flow rate of flue gas at the inlet of the flue gas treatment facility, amount of steam stored in the stripping section of the reactor, and amount of stripping steam—along with the predicted sulfur content of the raw material, are used as inputs to the BP neural network model. This results in a model with 6 neurons in the input layer, 2 hidden layers, and a 6-2-1 model structure. The output is the sulfur dioxide emission concentration at the inlet of the flue gas treatment facility. Conversely, the same six process parameters—oxygen content at the top outlet of the regenerator tower, carbon monoxide content at the top outlet of the regenerator tower, standard flow rate of flue gas at the inlet of the flue gas treatment facility, amount of steam stored in the stripping section of the reactor, and amount of stripping steam—along with the predicted sulfur content of the raw material, and the first and second sub-feature parameters, are used as inputs to the BP neural network model. This results in a model with 8 neurons in the input layer, 2 hidden layers, and an 8-2-1 model structure. The output is the sulfur dioxide concentration of the regenerated flue gas at the inlet of the treatment facility.
[0101] Finally, comparing the two, the prediction results show that the BP neural network model with eight input parameters has a prediction accuracy of 97.5%, while the prediction accuracy with the original six input parameters is only 88.3%. That is, the prediction accuracy obtained by using the first and second sub-feature parameters of this embodiment to predict the sulfur dioxide emission concentration at the inlet of the flue gas treatment facility is significantly higher than the prediction obtained by using only the original process parameters with high correlation.
[0102] like Figure 2 As shown, in a second aspect of the present invention, a device for predicting the sulfur dioxide emission concentration of catalytic regeneration flue gas is provided, comprising:
[0103] The data acquisition module is configured to acquire the sulfur content of the catalytic cracking reaction products and determine the characteristic parameters of the catalytic cracking reaction.
[0104] The prediction module is configured to use the sulfur content of the catalytic cracking reaction products as input, and predict the sulfur content of the catalytic cracking reaction feedstock using a feedstock sulfur content prediction model. The catalytic cracking reaction products are obtained by passing the catalytic cracking reaction feedstock through a catalytic cracking unit.
[0105] Using characteristic parameters and the sulfur content of the catalytic cracking feedstock as inputs, the sulfur dioxide concentration in the regenerated flue gas is predicted by a sulfur dioxide concentration prediction model.
[0106] The feedstock sulfur content prediction model is obtained by training a first preset neural network with the sulfur content of catalytic cracking reaction products and the corresponding sulfur content of catalytic cracking reaction feedstock under different operating conditions. The sulfur dioxide concentration prediction model is obtained by training a second preset neural network with the sulfur content of catalytic cracking reaction feedstock, characteristic parameters, and the corresponding sulfur dioxide concentration in regenerated flue gas under different operating conditions.
[0107] Optionally, the device further includes a feature parameter determination module; the feature parameter determination module is configured to:
[0108] The process parameters of the catalytic cracking reaction are obtained, and the sub-characteristic parameters of the catalytic cracking reaction are determined based on the sulfur content of the feedstock and the process parameters.
[0109] The process parameters and sub-characteristic parameters of the catalytic cracking reaction are determined as the characteristic parameters of the catalytic cracking reaction;
[0110] The process parameters include the oxygen content at the regenerator outlet, the carbon monoxide content at the regenerator outlet, the amount of steam stored in the reactor stripping section, the amount of stripping steam, and the standard flow rate of the regenerated flue gas in the catalytic cracking unit.
[0111] Optionally, the sub-characteristic parameters of the catalytic cracking reaction include:
[0112] The ratio of the sulfur content of the feedstock for catalytic cracking to the standard flow rate of the regenerated flue gas.
[0113] Optionally, the sub-characteristic parameters of the catalytic cracking reaction also include:
[0114] The ratio of stripping steam volume to the amount of steam stored in the stripping section of the reactor.
[0115] In a third aspect of the invention, a computer-readable medium is provided, which stores a computer program that, when executed by a processor, implements the above-described method for predicting sulfur dioxide emission concentrations in catalytic regeneration flue gas.
[0116] like Figure 3 As shown, in a fourth aspect of the present invention, a terminal device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for predicting sulfur dioxide emission concentration in catalytic regeneration flue gas.
[0117] like Figure 3 The diagram shown is a schematic representation of a terminal device provided in an embodiment of the present invention. Figure 3 As shown, the terminal device 10 of this embodiment includes a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, it implements the steps in the above method embodiments. Alternatively, when the processor 100 executes the computer program 102, it implements the functions of each module / unit in the above device embodiments.
[0118] For example, computer program 102 can be divided into one or more modules / units, one or more of which are stored in memory 101 and executed by processor 100 to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 102 in terminal device 10. For example, computer program 102 can be divided into a data acquisition module and a prediction module.
[0119] Terminal device 10 may be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. Terminal device 10 may include, but is not limited to, a processor 100 and a memory 101. Those skilled in the art will understand that... Figure 3 This is merely an example of terminal device 10 and does not constitute a limitation on terminal device 10. It may include more or fewer components than shown, or combine certain components, or different components. For example, terminal device may also include input / output devices, network access devices, buses, etc.
[0120] The processor 100 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.
[0121] The memory 101 can be an internal storage unit of the terminal device 10, such as a hard disk or RAM of the terminal device 10. The memory 101 can also be an external storage device of the terminal device 10, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal device 10. Furthermore, the memory 101 can include both internal and external storage units of the terminal device 10. The memory 101 is used to store computer programs and other programs and data required by the terminal device 10. The memory 101 can also be used to temporarily store data that has been output or will be output.
[0122] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.
[0123] In summary, this implementation method collects, processes, and filters data from the operation of the catalytic cracking unit. First, it models the sulfur content in dry gas, liquefied petroleum gas, gasoline, diesel, and sulfur-containing wastewater as input variables and the sulfur content of the feedstock as the output value, achieving soft measurement of the sulfur content in the feedstock. Second, it selects process parameters and characteristic parameters related to the sulfur dioxide emission concentration in the regenerated flue gas at the inlet of the flue gas treatment facility and uses them as input variables for model training. The trained model then predicts the sulfur dioxide emission concentration in the regenerated flue gas at the inlet of the flue gas treatment facility. Through the above steps, it predicts the sulfur dioxide emission concentration at the inlet of the flue gas treatment facility in real time during the actual production process of catalytic cracking, enabling timely adjustment of the operating conditions of the catalytic cracking unit and the treatment facility.
[0124] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention.
[0125] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not describe the various possible combinations separately.
[0126] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.
[0127] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, and should also be regarded as the content disclosed by the embodiments of the present invention.
Claims
1. A method of predicting flue gas sulfur dioxide emission concentration for catalytic regenerators, characterized by, The method comprises the following steps: obtaining the sulfur content of the catalytic cracking reaction product, and predicting the sulfur content of the catalytic cracking reaction raw material through a raw material sulfur content prediction model with the sulfur content of the catalytic cracking reaction product as input; determining the characteristic parameters of the catalytic cracking reaction, and predicting the concentration of sulfur dioxide in the regenerated flue gas through a sulfur dioxide concentration prediction model with the characteristic parameters and the sulfur content of the catalytic cracking reaction raw material as input; the raw material sulfur content prediction model is obtained by training a first preset neural network through the sulfur content of the catalytic cracking reaction product and the corresponding sulfur content of the catalytic cracking reaction raw material under different working conditions, and the sulfur dioxide concentration prediction model is obtained by training a second preset neural network through the sulfur content of the catalytic cracking reaction raw material, the characteristic parameters and the corresponding concentration of sulfur dioxide in the regenerated flue gas under different working conditions; the training process of the raw material sulfur content prediction model comprises: determining the initial network parameters of the first preset neural network; inputting the sulfur content of dry gas, the sulfur content of liquefied gas, the sulfur content of gasoline, the sulfur content of diesel and the sulfur content of sulfur-containing sewage obtained under different working conditions into the first preset neural network to predict the sulfur content of the catalytic cracking reaction raw material corresponding to the current working condition; obtaining the measured sulfur content of the catalytic cracking reaction raw material corresponding to the current working condition, and updating the network parameters of the first preset neural network in the case that the difference between the obtained sulfur content of the catalytic cracking reaction raw material and the measured sulfur content of the catalytic cracking reaction raw material is greater than the sulfur content threshold until the difference between the obtained sulfur content of the catalytic cracking reaction raw material and the measured sulfur content of the catalytic cracking reaction raw material is not greater than the sulfur content threshold, and taking the first preset neural network with the updated network parameters as the raw material sulfur content prediction model; determining the characteristic parameters of the catalytic cracking reaction, comprising: obtaining the process parameters of the catalytic cracking reaction, and determining the process parameters as the characteristic parameters of the catalytic cracking reaction; the process parameters comprise the oxygen content at the outlet of the regenerator, the carbon monoxide content at the outlet of the regenerator, the amount of the reactor stripping section, the amount of stripping steam and the flue gas standard state flow of the regenerated flue gas in the catalytic cracking device; the training process of the sulfur dioxide concentration prediction model comprises: determining the initial network parameters of the second preset neural network; inputting the sulfur content of the catalytic cracking reaction raw material, the oxygen content at the outlet of the regenerator, the carbon monoxide content at the outlet of the regenerator, the amount of the reactor stripping section, the amount of stripping steam and the flue gas standard state flow of the regenerated flue gas obtained under different working conditions into the second preset neural network to predict the concentration of sulfur dioxide in the regenerated flue gas corresponding to the current working condition; obtaining the measured concentration of sulfur dioxide in the regenerated flue gas corresponding to the current working condition, and updating the network parameters of the second preset neural network in the case that the difference between the obtained concentration of sulfur dioxide in the regenerated flue gas and the measured concentration of sulfur dioxide in the regenerated flue gas is greater than the sulfur dioxide concentration threshold until the difference between the obtained concentration of sulfur dioxide in the regenerated flue gas and the measured concentration of sulfur dioxide in the regenerated flue gas is not greater than the sulfur dioxide concentration threshold, and taking the second preset neural network with the updated network parameters as the sulfur dioxide concentration prediction model.
2. The catalytic regenerative flue gas sulfur dioxide emission concentration prediction method of claim 1, wherein, determining the characteristic parameters of the catalytic cracking reaction, comprising: obtaining process parameters of the catalytic cracking reaction, and determining sub-feature parameters of the catalytic cracking reaction according to sulfur content of the catalytic cracking reaction feedstock and the process parameters; determining the process parameters of the catalytic cracking reaction and the sub-feature parameters as feature parameters of the catalytic cracking reaction; the process parameters include oxygen content at the outlet of a regenerator in the catalytic cracking device, carbon monoxide content at the outlet of the regenerator, reactor stripping section inventory, stripping steam amount, and flue gas standard state flow rate of the regenerator flue gas; the sub-feature parameters of the catalytic cracking reaction include: a ratio of the sulfur content of the catalytic cracking reaction feedstock to the flue gas standard state flow rate of the regenerator flue gas.
3. The catalytic regenerative flue gas sulfur dioxide emission concentration prediction method of claim 2, wherein, the sub-feature parameters of the catalytic cracking reaction further include: a ratio of the stripping steam amount to the reactor stripping section inventory.
4. A catalytic regenerable flue gas sulfur dioxide emission concentration prediction device applying the catalytic regenerable flue gas sulfur dioxide emission concentration prediction method according to any one of claims 1 to 3, characterized by, the device includes: a data acquisition module configured to obtain sulfur content of a catalytic cracking reaction product, and determine feature parameters of the catalytic cracking reaction; a prediction module configured to, with the sulfur content of the catalytic cracking reaction product as input, predict sulfur content of a catalytic cracking reaction feedstock through a feedstock sulfur content prediction model, the catalytic cracking reaction product being obtained from the catalytic cracking reaction feedstock through a catalytic cracking reaction in a catalytic cracking device; and with the feature parameters and the sulfur content of the catalytic cracking reaction feedstock as input, predict concentration of sulfur dioxide in the regenerator flue gas through a sulfur dioxide concentration prediction model; the feedstock sulfur content prediction model is obtained by training a first preset neural network through sulfur content of a catalytic cracking reaction product and corresponding sulfur content of a catalytic cracking reaction feedstock under different working conditions, and the sulfur dioxide concentration prediction model is obtained by training a second preset neural network through sulfur content of a catalytic cracking reaction feedstock, feature parameters, and corresponding concentration of sulfur dioxide in the regenerator flue gas under different working conditions; the device further includes a feature parameter determination module; the feature parameter determination module is configured to: obtain process parameters of the catalytic cracking reaction, and determine sub-feature parameters of the catalytic cracking reaction according to sulfur content of the catalytic cracking reaction feedstock and the process parameters; determine the process parameters of the catalytic cracking reaction and the sub-feature parameters as feature parameters of the catalytic cracking reaction; the process parameters include oxygen content at the outlet of a regenerator in the catalytic cracking device, carbon monoxide content at the outlet of the regenerator, reactor stripping section inventory, stripping steam amount, and flue gas standard state flow rate of the regenerator flue gas; the sub-feature parameters of the catalytic cracking reaction include: a ratio of the sulfur content of the catalytic cracking reaction feedstock to the flue gas standard state flow rate of the regenerator flue gas.
5. The catalytic regenerative flue gas sulfur dioxide emission concentration prediction apparatus of claim 4, wherein, the sub-feature parameters of the catalytic cracking reaction further include: a ratio of the stripping steam amount to the reactor stripping section inventory.
6. A computer readable medium storing a computer program, characterized in that, the computer program, when executed by a processor, implements the catalytic regenerator flue gas sulfur dioxide emission concentration prediction method of any one of claims 1-3.
7. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, the processor, when executing the computer program, implements the catalytic regenerator flue gas sulfur dioxide emission concentration prediction method of any one of claims 1-3.