CN114427684B - Control method and control device for combustion furnace in natural gas purification process - Google Patents

Abstract

The present invention relates to a control method and control device for a combustion furnace in a natural gas purification process, comprising the following steps: collecting acid gas parameters currently entering a sulfur recovery unit, optimizing through an intelligent algorithm to obtain key operating parameters corresponding to the optimal energy efficiency conditions under the acid gas parameter conditions; controlling the corresponding natural gas purification device according to the key operating parameters; the intelligent algorithm can use a genetic algorithm for iterative optimization; the energy efficiency conditions can be predicted by training a neural network model with historical data. The method of the present invention can scientifically and valuablely evaluate the operating level of the tail gas treatment unit in natural gas purification, improve the energy efficiency level of the combustion furnace, and reduce production costs.

Description

A control method and control device for a combustion furnace in a natural gas purification process

Technical Field

The invention relates to a control method and a control device for a combustion furnace in a natural gas purification process, belonging to the technical field of natural gas purification.

Background Art

The natural gas purification process is an important part of the natural gas development and utilization process, and it is also a high energy consumption and high material consumption link. Especially for the purification of high-sulfur natural gas, the fuel gas energy consumption of the tail gas incinerator and hydrogenation feed combustion furnace included in the tail gas treatment unit accounts for about 60% of the total energy consumption of the purification process, which is an important link in energy saving and consumption reduction. At the same time, the waste heat boiler after the tail gas incinerator can also recover a large amount of medium-pressure steam, and the medium-pressure steam can in turn be used by other links in the natural gas purification process, thereby reducing the energy consumption of the purification plant's public works (used to provide the necessary energy and energy-carrying working fluids for the purification process). The energy-saving measures in the production process are mainly based on the dispatcher's acid gas load, operating specifications and experience to adjust and control some operating parameters. It is impossible to accurately control and maximize energy utilization. In order to prevent insufficient energy supply from causing tail gas to fail to meet standards, it tends to adopt an oversupply method for production, which has energy waste problems. The main reason for this situation is the lack of a scientific and effective tail gas incinerator evaluation system to guide on-site operation evaluation and operation optimization. However, there is currently a lack of effective evaluation standards for combustion furnaces. Existing evaluation methods for industrial boilers, industrial combustion furnaces and heating furnaces all evaluate the thermal efficiency of the furnace from the perspective of thermal balance, which is not suitable for the evaluation and optimization of the operating performance of the above-mentioned combustion furnaces that are closely integrated with the natural gas purification process.

The difficulty and particularity of scientifically evaluating the operating performance of the natural gas purification process burner lies in the fact that the fuel gas consumption of the hydrogenation feed burner and the tail gas incinerator is affected by the acid gas flow and composition of the upstream desulfurization and deacidification unit, as well as the operating parameters of the upstream sulfur recovery device. At the same time, the influence of its own operating parameters on the fuel gas consumption cannot be ignored. Changes in operating parameters will also affect the amount of medium-pressure steam generated by heat recovery.

Acid gas parameters are constantly changing due to the influence of upstream process links, and have a complex coupling relationship with the burner fuel gas consumption and medium-pressure steam generation. At the same time, energy efficiency evaluation lacks scientific evaluation indicators, making it difficult to evaluate the operation of the burner in the natural gas purification process, and it is impossible to determine whether the current operation is energy-saving and efficient. It is even more difficult to further optimize the burner operation and achieve further energy-saving operation.

Summary of the invention

The purpose of the present invention is to provide a control method and a control device for a burner in a natural gas purification process, so as to solve the problem that the energy consumption of the burner in the natural gas purification process is difficult to evaluate and optimize.

To achieve the above object, the solution of the present invention includes:

A method for controlling a combustion furnace in a natural gas purification process of the present invention comprises the following steps:

1) Collecting the acid gas parameters currently entering the sulfur recovery unit, and optimizing the key operating parameters corresponding to the optimal energy efficiency conditions under the acid gas parameter conditions through intelligent algorithm B;

The energy efficiency conditions include one or more of the ratio of the fuel gas consumption of the tail gas treatment unit to the acid gas treatment amount entering the sulfur recovery unit, the ratio of the steam generated by the tail gas treatment unit to the acid gas treatment amount entering the sulfur recovery unit, and the ratio of the comprehensive energy consumption of the tail gas treatment unit to the acid gas treatment amount entering the sulfur recovery unit;

2) controlling the corresponding natural gas purification device according to the key operating parameters;

The optimization process of the intelligent algorithm B in step 1) includes: firstly initializing the population, the population including the key operating parameters; then calculating the fitness of the population, the fitness being the energy efficiency condition, and the energy efficiency condition being calculated by the prediction model A; finally changing the population and iterating multiple times to finally select the key operating parameters corresponding to the population with the best energy efficiency condition as the optimization result;

The prediction model A is a machine learning model, which is trained by the historical data of the acid gas parameters, key operating parameters and corresponding energy efficiency conditions.

The key to achieving optimized operation of the burner is to establish burner evaluation indicators in combination with the natural gas purification process, and then establish a relationship model between acid gas parameters and corresponding key operating parameters and evaluation indicators. The model of the present invention is based on the operating principle of the natural gas purification process and combines the actual operating performance of the on-site purification equipment. Moreover, the included key operating parameters can determine the adjustable key operating parameter value guidance value corresponding to the evaluation index benchmark value, i.e., the optimal value, which can be used for optimizing the on-site operation. The evaluation and optimization work are organically combined to achieve the energy efficiency evaluation and energy-saving optimization of the natural gas purification process burner.

Furthermore, the comprehensive energy consumption of the tail gas treatment unit is obtained by weighting the sum of the fuel gas consumption of the tail gas treatment unit and the steam generation of the tail gas treatment unit according to their respective energy conversion coefficients.

In addition to considering the energy consumption of the burner, the evaluation indicators also take into account the output of the burner's medium-pressure steam, an energy-carrying working fluid. It is not limited to improving its own energy efficiency, but has a longer-term vision for energy-saving operation, which is beneficial to energy saving in the entire natural gas purification process.

Furthermore, the prediction model A is a neural network model.

The prediction model A can be any machine learning model. The machine learning model is trained by a large amount of historical data. Based on the model, the selected evaluation indicators under different acid gas parameters and related key operating parameters can be predicted.

Furthermore, the intelligent algorithm B is a genetic algorithm.

Intelligent algorithm B can use iterative optimization algorithms such as genetic algorithm or particle swarm algorithm.

Furthermore, the tail gas treatment unit includes a hydrogenation feed combustion furnace and a tail gas incinerator.

Furthermore, the acid gas parameters include acid gas flow rate, hydrogen sulfide content, carbon dioxide content, water content, methane content, and hydroxyl sulfur content.

A control device for a burner in a natural gas purification process of the present invention comprises a controller and a memory, wherein the controller executes instructions stored in the memory to implement the control method for the burner in a natural gas purification process as described above.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a flow chart of the method of the present invention;

FIG2 is a purification process flow chart of a high-sulfur natural gas purification plant combined unit;

Fig. 3 is a schematic diagram of the structure of a typical artificial neural network model;

4 is a schematic diagram showing a comparison between the predicted results of the unit fuel gas consumption of the hydrogenation feed combustion furnace using the method of the present invention and the actual operation results;

5 is a relative error distribution diagram of the prediction results of the unit fuel gas consumption of the hydrogenation feed combustion furnace by the method of the present invention;

6 is a schematic diagram showing a comparison between the prediction result of the unit fuel gas consumption of the tail gas incinerator using the method of the present invention and the actual operation result;

7 is a relative error distribution diagram of the prediction results of the unit fuel gas consumption of the tail gas incinerator by the method of the present invention;

FIG8 is a schematic diagram showing a comparison between the prediction result of the medium-pressure steam output of the waste heat boiler of the tail gas incinerator using the method of the present invention and the actual operation result;

9 is a relative error distribution diagram of the prediction results of the medium-pressure steam output of the waste heat boiler of the tail gas incinerator by the method of the present invention;

10 is a schematic diagram of the technical route for operation evaluation and optimization of the combustion furnace of the present invention using a neural network and a genetic algorithm;

Figure 11 is a schematic diagram of a multi-objective Pareto front solution under certain acid gas parameters;

FIG. 12 is a schematic diagram of a control device for a combustion furnace in the natural gas purification process of the present invention.

DETAILED DESCRIPTION

The present invention will be further described in detail below in conjunction with the accompanying drawings.

Method Example:

The present invention provides a control method for a burner in a natural gas purification process. Based on the on-site natural gas purification process and device operation data, a multivariate evaluation index for the burner in a natural gas purification plant is established and related acid gas parameters and corresponding key operating parameters are determined. The method is shown in FIG1 and specifically includes the following steps:

1) First, establish the evaluation index of the combustion furnace, which includes unit fuel gas consumption, unit medium-pressure steam production and unit comprehensive energy consumption.

2) Establish prediction model A: Based on historical data, establish prediction model A for evaluation indicators, acid gas parameters and corresponding key operating parameters. Select standard historical data to train the prediction model.

3) Establish optimization algorithm model B; based on prediction model A, establish optimization algorithm models for both acid gas parameters and corresponding key operating parameters when the corresponding evaluation index is optimal; at the same time, it can also calculate the energy consumption benchmark value of the corresponding evaluation index under different acid gas parameters, that is, the theoretical optimal energy consumption value under the corresponding acid gas parameters. The difference between the energy consumption benchmark value and the actual value on site reflects the energy-saving potential and can also effectively evaluate the operating level of the burner in the natural gas purification process. While determining the benchmark value of the evaluation index, the key operating parameters of the burner and related equipment obtained by the optimization algorithm model B can be used as the optimization guidance value of the corresponding burner control parameters.

4) Energy-saving control of natural gas purification process combustion furnace: By adjusting and controlling the combustion furnace and related upstream and downstream equipment, the control parameters are made to reach the key operating parameter guidance values or the operating parameters are made to move closer to the key operating parameter guidance values, so as to achieve energy saving and consumption reduction of the natural gas purification combustion furnace and achieve the purpose of optimizing energy consumption.

The following is a more detailed explanation of each step of the present invention in conjunction with examples.

The natural gas purification process selected in this embodiment is a typical high-sulfur natural gas purification process, which mainly includes five main units: MDEA desulfurization and decarbonization (deacidification), TEG dehydration, conventional Claus sulfur recovery and hydrogenation reduction tail gas treatment, and low-pressure acid water stripping.

Taking a high-sulfur natural gas purification plant as an example, the main process flow of the purification process of the purification plant is shown in Figure 2, which shows a process flow of a combined device with two purification series (I and II). The prediction model A and the optimization algorithm B are both for a complete purification series. The purification process of high-sulfur natural gas is as follows: the raw natural gas is dehydrated by the deacidification unit using MDEA lean amine liquid to remove hydrogen sulfide, part of organic sulfur and carbon dioxide, etc.; then it is dehydrated by the dehydration unit to meet the product gas requirements and is transported through the long-distance pipeline network; the acid gas generated by the deacidification unit enters the sulfur recovery unit, mixed with air and enters the reactor to react to recover the sulfur element as liquid sulfur, and then industrial sulfur is produced by the sulfur forming unit. The reaction gas in this process is passed through the waste heat boiler to generate medium-pressure saturated steam; the tail gas of the sulfur recovery unit is treated by the tail gas treatment unit, and the flue gas that meets environmental protection requirements is discharged through the chimney; the acid water generated in the purification process is sent to the acid water stripping unit, and the purified water generated is recycled, and the acid gas generated by stripping is sent to the tail gas treatment unit for treatment. The tail gas treatment unit includes a combustion furnace, which includes a hydrogenation feed combustion furnace and a tail gas incinerator for consuming fuel gas, and a tail gas incinerator waste heat boiler for generating medium-pressure steam.

1. Establish and select evaluation indicators for combustion furnace operation performance.

According to the on-site operation requirements, we must pay attention to the fuel gas consumption and medium-pressure steam production respectively, and also hope to obtain the optimal result of comprehensive energy consumption. Therefore, three evaluation indicators are set for the burner.

(1) Unit fuel gas consumption.

It represents the ratio of the fuel gas consumption of the hydrogenation feed combustion furnace and the tail gas incinerator to the acid gas treatment amount entering the sulfur recovery unit, which is represented by e ₁ here, and the unit is Nm ³ /10 ⁴ Nm ³ . The calculation expression is:

Wherein, Macid _gas is the acid gas treatment capacity of the sulfur recovery unit, in units of 10 ⁴ Nm ³ /t; Mfuel _gas is the sum of the fuel gas consumption of the hydrogenation feed combustion furnace and the tail gas incinerator in the tail gas treatment unit, in units of Nm ³ /t; t is the time unit, which can be hours, days or other time units, wherein the time units in the acid gas treatment capacity and the fuel gas consumption must be consistent.

(2) Unit medium-pressure steam output

It represents the ratio of the medium-pressure steam produced by the tail gas incinerator to the acid gas treatment amount entering the sulfur recovery unit. It is represented by e ₂ here, and the unit is ton/10 ⁴ Nm ^3. The calculation expression is:

Wherein, Macid _gas is the acid gas treatment capacity of the sulfur recovery unit, in units of 10 ⁴ Nm ³ /t; _Msteam is the medium-pressure superheated steam generated by the tail gas incinerator in the tail gas treatment unit, in units of ton/t; t is the time unit, which can be hours, days or other time units, wherein the time unit of the acid gas treatment capacity and the medium-pressure steam generation should be consistent.

(3) Unit comprehensive energy consumption

It represents the ratio of the _{comprehensive} energy consumption E of the hydrogenation feed combustion furnace and the tail gas incinerator to the acid gas treatment volume M _{acid gas} entering the sulfur recovery unit, which is represented by e ₃ here, and the unit is MJ/10 ⁴ Nm ^3. The calculation expression is:

Among them, E _{comprehensive energy consumption} is the comprehensive energy consumption composed of gas consumption and medium-pressure steam generation, M _{fuel gas} is the sum of fuel gas consumption of hydrogenation feed combustion furnace and tail gas incinerator in tail gas treatment unit, and the unit is Nm ³ /t; M _steam is the medium-pressure steam generation of tail gas incinerator in tail gas treatment unit, and the unit is ton/t; c ₁ and c ₂ are energy conversion coefficients of fuel gas and medium-pressure superheated steam according to national standards, and the units are Nm ³ /t and ton/t respectively; M _{acid gas} is the acid gas treatment capacity of sulfur recovery unit, and the unit is 10 ⁴ Nm ³ /t; t is the time unit, which can be hours, days or other time units, and the time units of various parameters must be consistent.

2. Use historical big data to establish a relationship prediction model A between the acid gas parameters, key operating parameters and corresponding evaluation indicators that affect the operating performance of the combustion furnace in the natural gas purification plant, and predict the evaluation index values under different operating conditions.

(1) Determine the acid gas parameters and key operating parameters that affect the evaluation indicators.

In addition to the uncontrollable acid gas parameters such as acid gas flow, pressure, hydrogen sulfide content and carbon dioxide content, the parameters that affect the evaluation index of the combustion furnace also include key operating parameters that can be adjusted and controlled on site. Among the evaluation indicators, the key issue is to determine the key operating parameters that affect the evaluation indicators, which are controllable parameters in the evaluation system. According to the process flow of the sulfur recovery unit and the tail gas treatment unit, the working principle of the combustion furnace and the field operation experience, the key operating parameters corresponding to the evaluation indicators are obtained. Table 1 shows the energy price indicators and the corresponding key operating parameters.

Table 1 Evaluation indexes and key operating parameters of the combustion furnace of the tail gas treatment unit

The above key operating parameters can all be obtained from the operating data of the natural gas purification plant for operation performance evaluation.

(2) Based on historical data, intelligent algorithm machine learning is used to establish a prediction model A between acid gas parameters, key operating parameters of the combustion furnace in the natural gas purification process, and evaluation indicators.

① Select a period for collecting historical data, such as February to May 2019, to collect acid gas parameters and corresponding key operating parameters as well as the energy consumption and material consumption of the combustion furnace at the corresponding time. Calculate the performance evaluation index based on the collected data.

② Based on the collected historical big data, a relationship prediction model A is established between different acid gas parameters, key operating parameters and corresponding evaluation indicators.

Excluding shutdown and abnormal data (if any) in historical data, according to the valid historical data of a single purification series (historical data of qualified natural gas produced after purification), an artificial intelligence algorithm is used to obtain a prediction model A with key operating parameters as independent variables and evaluation indicators as dependent variables. This embodiment uses an artificial neural network to establish a prediction model A. The artificial neural network is a certain simplification, abstraction and simulation of the biological neural structure, and is an empirical modeling tool. It can learn the complex relationship between input and output from the data collected from a specific problem domain, and has good performance in accurate prediction and classification. It has been widely used in many fields of engineering applications. The multilayer perceptron artificial neural network is one of the most widely used neural networks in various fields of engineering problems. It includes an input layer, a hidden layer and an output layer, and each layer is composed of neurons. The number of neurons in the input layer is equal to the number of input parameters, i.e., acid gas parameters and key operating parameters; the number of neurons in the output layer is equal to the number of prediction targets, where the prediction target is the selected evaluation indicator. The hidden layer can be composed of one or more layers, and the number of layers and neurons in the hidden layer can be obtained by trial and error or combined with intelligent algorithm optimization.

Figure 3 shows a typical fully connected network structure. The model receives data input through the input layer nodes/neurons, passes it to the hidden layer nodes, and finally passes the information to the output nodes. Neurons are connected to any neuron in the next layer through communication links associated with connection weights (Synaptic Weight). Each neuron receives the output of each neuron in the previous layer, and sums it with the connection weight. On this basis, the bias is added to calculate the single output of the neuron through the activation function. In order to adapt to specific data/problems, it is necessary to configure and train the corresponding artificial neural network model. The training process can be regarded as reducing and minimizing the error between the expected output and the actual output of the model. The training of the artificial neural network model is to introduce randomly selected samples with input data and expected output into the artificial neural network configuration model, determine the error between the expected output value and the actual output of the model, and minimize the error by modifying/optimizing the connection weights and biases of the neurons.

Taking the unit fuel gas consumption of the hydrogenation feed burner as an evaluation index to establish prediction model A as an example, the input of the prediction model is 7 key operating parameters that affect the unit fuel gas consumption of the hydrogenation feed burner, specifically: acid gas flow, tail gas _H2S / _SO2 , primary reactor inlet temperature, secondary reactor inlet temperature, final reactor inlet temperature, hydrogenation feed burner air ratio and hydrogenation feed burner outlet temperature. The samples used to establish the model come from historical operating data from February to May 2019, with one group extracted every hour, excluding some shutdown or abnormal data. The sample data obtained is randomly divided into three categories: training set, validation set and test set, accounting for 80%, 10% and 10% respectively.

Before inputting the sample data into the neural network model for training, the minmax function in Matlab was used to normalize the input data to the range of [-1,1]. The transfer function of the hidden layer was the S-type function "tansig", and the transfer function of the output layer was the linear function "purelin". The training function of the network was the "trainlm" function, which updated the weights and bias values based on the Levenberg-Marquardt algorithm. A single-layer hidden layer network was used to establish the model, and the trial-and-error method was used to find the optimal number of hidden layer neurons. The network performance function used the mean square error (MSE), and the regression R value was used to measure the correlation between the expected data and the actual output of the network. Since the samples used to train the network and the initial values of the network weights and biases were randomly selected and generated, the training results of the neural network with the same structure were different each time. Therefore, each structure neural network was trained 5 times and the mean square error was averaged. The average mean square error values calculated by the neural networks with different structures were different. When the number of hidden layer neurons was 5, the mean square error value and correlation coefficient value of the test set were relatively ideal, so it was selected for subsequent research.

In order to quantify the difference between the predicted results of the energy-carrying medium consumption and the energy and material consumption prediction model A and the actual value, the average relative deviation (AAD%) is defined and calculated by the following formula, where _yi , _xi and n represent the actual value, the calculated value of the network model and the number of samples respectively.

In this embodiment, a prediction model is established by taking the establishment of a relationship model between key operating parameters and three evaluation indexes as an example, and a neural network algorithm is used to establish a prediction model. Figure 4 is a comparison chart between the predicted value and the actual historical value of the unit fuel gas (fuel gas) consumption of the hydrogenation feed combustion furnace under different working conditions, and Figure 5 is the distribution of the relative error of the predicted value, and most samples are within ±4%.

Figure 6 is a comparison chart between the predicted value and the actual historical value of the unit fuel gas consumption of the tail gas incinerator under different working conditions, and Figure 7 is the distribution of the relative error of the predicted value, most of the samples are within ±6%. Figure 8 is a comparison chart between the predicted value and the actual historical value of the unit medium-pressure steam production of the waste heat boiler of the tail gas incinerator under different working conditions, and Figure 9 is the distribution of the relative error of the predicted value, most of the samples are within ±3%.

3. Based on the relationship prediction model A established in step 2, an optimization algorithm model B between the evaluation indicators and the corresponding key operating parameters and acid gas parameters is established to determine the benchmark values of the evaluation indicators and the guidance values of the key operating parameters.

Based on the established relationship prediction model A between acid gas parameters, key operating parameters and evaluation indicators, an optimization algorithm model B between acid gas parameters and corresponding key operating parameters when the evaluation indicators are optimal is established. Under each acid gas parameter value, by changing the values of the adjustable key operating parameters, the optimal value of the energy price indicator, that is, the benchmark value with the lowest energy consumption, is obtained. When the benchmark value is determined, the guidance values of the corresponding adjustable key operating parameters are also determined. These guidance values can be used to guide the on-site parameter adjustment of key operating parameters to achieve the lowest process energy consumption of the tail gas treatment unit in the natural gas purification process.

The evaluation index can select one or more of the above three evaluation indexes. After selecting the evaluation index, the key operating parameters are determined according to Table 1 above. After obtaining the acid gas parameters, the key operating parameter values corresponding to the selected evaluation index when it is optimal are calculated by the optimization algorithm model B. The key operating parameter values corresponding to the optimal energy efficiency evaluation index are called key operating parameter guidance values. The optimal evaluation index value is also the evaluation index benchmark value when the energy consumption is the lowest/energy efficiency level is the highest. Because the key operating parameters in different evaluation indexes are the same, different evaluation indexes may obtain different key operating parameter guidance values for different key operating parameters when they are optimal. When the key operating parameter guidance values are used to control and adjust the operation of the natural gas purification device (mainly the tail gas treatment unit, and may also include the upstream sulfur recovery unit), one of them can be selected according to the actual situation or field experience, or the average of different key operating parameter guidance values can be calculated, or weights can be set for different evaluation indexes, and the key operating parameter guidance values corresponding to the evaluation index can be selected according to the weights for the control and adjustment of the natural gas purification device.

As shown in FIG10 , in this embodiment, a genetic algorithm NSGA-Ⅱ in an artificial intelligence algorithm is used to establish a multi-objective optimization algorithm model, the optimization objectives are unit fuel gas consumption and unit medium-pressure steam production, and the variables are key operating parameters including acid gas parameters. FIG11 is a Pareto frontier solution of unit fuel gas consumption and unit medium-pressure steam production obtained by solving the NSGA-Ⅱ optimization algorithm model under certain acid gas parameters. During the calculation process of the NSGA-Ⅱ algorithm, the population number is 500, the iteration number is 200 generations, and Matlab is used for programming and solving the calculation.

Each point in the figure represents a set of optimized solutions, and the left and right ends represent the two sets of solutions with the highest medium-pressure steam production and the lowest fuel gas consumption. Because the Pareto front solution is a non-inferior set solution, it is necessary to select one of the Pareto front solutions based on the actual focus on the site, so as to determine the benchmark values of unit fuel gas consumption and unit medium-pressure steam production, and the optimization guidance values of the corresponding key operating parameters can also be determined. It is also possible to combine the two objectives and obtain the unit comprehensive energy consumption benchmark value based on all Pareto front solution sets.

4. Substitute the acid gas parameters of the actual on-site operating conditions into the optimization algorithm model B to obtain the evaluation index benchmark value and the corresponding adjustable key operating parameter optimization guidance value.

Under actual working conditions, the difference between the actual operating value of the key operating parameter and the guidance value of the key operating parameter calculated by the optimization algorithm model B is the on-site tuning space, which provides direct support for the optimized operation of the burner in the natural gas purification process, that is, the hydrogenation feed burner and the tail gas incinerator are controlled according to the guidance value, so that each key operating parameter is adjusted towards the guidance value, thereby improving the energy efficiency level of the burner in the natural gas purification process.

Table 2 shows the unit comprehensive energy consumption benchmark values and the optimization guidance values of the corresponding key operating parameters obtained under certain acid gas parameters. All the values given in the table are dimensionless quantities, and the obtained unit comprehensive energy consumption benchmark values and the corresponding operating parameter optimization guidance values are defined as "1", while the actual unit comprehensive energy consumption and actual operating parameters are given dimensionless values relative to the benchmark conditions. This is only used to demonstrate the evaluation and optimization methods, and the difference can also be used in actual field applications.

Table 2 Benchmark value and optimization guidance value of unit comprehensive energy consumption under certain acid gas parameters

The control method of the burner in the natural gas purification process of the present invention establishes a multivariate evaluation index for the burner in the natural gas purification plant and determines the acid gas parameters and key operating parameters of the equipment related thereto; the evaluation index is calculated based on the actual on-site operation data; a prediction model of the evaluation index, acid gas parameters and key operating parameters is established by using a neural network; an optimization algorithm model between the acid gas parameters and the key operating parameters when the evaluation index is optimal is established by using a genetic algorithm; and the benchmark value (optimal value or the lowest energy consumption value) of the burner evaluation index under different acid gas parameters can be calculated. After comparing the difference between the benchmark value of the evaluation index and the actual on-site value, the analysis shows that the calculation result is consistent with the theoretical analysis and on-site practice. While determining the benchmark value of the evaluation index, the optimization algorithm model can also determine the optimization guidance value of the key operating parameters.

The method of the present invention not only considers the impact of differences in acid gas parameters on the actual operating level, but also considers the impact of controllable key operating parameter values on the actual operating energy consumption level under the same acid gas parameters. It not only realizes the scientific and intuitive evaluation of the operating level and energy-saving space of the combustion furnace under different working conditions, but also provides easy-to-implement operating parameter tuning guidance for its optimized operation.

Device Example:

This embodiment provides a control device for a burner in a natural gas purification process, as shown in FIG12 , including a memory, a processor and an internal bus. The processor and the memory communicate with each other through the internal bus.

The processor may be a processing device such as a microprocessor MCU, a programmable logic device FPGA, etc.

The memory can be various memories that use electrical energy to store information, such as RAM, ROM, etc.; various memories that use magnetic energy to store information, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memories, USB flash drives, etc.; various memories that use optical methods to store information, such as CDs, DVDs, etc. Of course, there are other types of memory, such as quantum memory, graphene memory, etc.

The processor can call the logic instructions in the memory to implement a method for controlling a burner in a natural gas purification process. The method is described in detail in the method embodiment and will not be described again here.

Claims (4)

1. A method for controlling a combustion furnace in a natural gas purification process, characterized in that it comprises the following steps: 1) Collect the acid gas parameters currently entering the sulfur recovery unit, and use the intelligent algorithm B to find the key operating parameters corresponding to the optimal energy efficiency conditions under the acid gas parameter conditions; the acid gas parameters include: acid gas flow rate, hydrogen sulfide content, carbon dioxide content, water content, methane content, hydroxyl sulfur content; the key operating parameters include: acid gas flow rate, tail gas H ₂ S/SO ₂ , reactor inlet temperature, hydrogenation feed burner air ratio and hydrogenation feed burner outlet temperature; The energy efficiency conditions include one or more of the ratio of the fuel gas consumption of the tail gas treatment unit to the acid gas treatment amount entering the sulfur recovery unit, the ratio of the steam generated by the tail gas treatment unit to the acid gas treatment amount entering the sulfur recovery unit, and the ratio of the comprehensive energy consumption of the tail gas treatment unit to the acid gas treatment amount entering the sulfur recovery unit; the comprehensive energy consumption of the tail gas treatment unit is obtained by weighting the sum of the fuel gas consumption of the tail gas treatment unit and the steam generated by the tail gas treatment unit according to their respective energy conversion coefficients; 2) controlling the corresponding natural gas purification device according to the key operating parameters; The optimization process of the intelligent algorithm B in step 1) includes: firstly initializing the population, the population includes the key operating parameters; then calculating the fitness of the population, the fitness is the energy efficiency condition, and the energy efficiency condition is calculated by the prediction model A; finally, changing the population and iterating multiple times to finally select the key operating parameters corresponding to the population with the best energy efficiency condition as the optimization result; The prediction model A is a machine learning model, which is trained by the historical data of the acid gas parameters, key operating parameters and corresponding energy efficiency conditions; The natural gas purification process includes: the raw natural gas is treated with MDEA lean amine liquid in the deacidification unit, and then dehydrated in the dehydration unit before being transported out; the acid gas generated by the deacidification unit enters the sulfur recovery unit, is mixed with air and enters the reactor to react to recover the sulfur element as liquid sulfur, and is processed by the sulfur forming unit to produce industrial sulfur. The reaction gas in this process is passed through the waste heat boiler to generate medium-pressure saturated steam; the tail gas of the sulfur recovery unit is processed by the tail gas treatment unit, the acid water generated in the purification process is sent to the acid water stripping unit, and the acid gas generated by the stripping is sent to the tail gas treatment unit for treatment; the tail gas treatment unit includes a combustion furnace, and the combustion furnace includes a hydrogenation feed combustion furnace and a tail gas incinerator that consumes fuel gas, and a tail gas incinerator waste heat boiler that generates medium-pressure steam. 2. The method for controlling a burner in a natural gas purification process according to claim 1, wherein the prediction model A is a neural network model. 3. The method for controlling a burner in a natural gas purification process according to claim 1, characterized in that the intelligent algorithm B is a genetic algorithm. 4. A control device for a burner in a natural gas purification process, characterized in that it comprises a controller and a memory, wherein the controller executes instructions stored in the memory to implement the control method for a burner in a natural gas purification process as claimed in any one of claims 1 to 3.