Tubular air preheater for atmospheric and vacuum heating furnace and control method of tubular air preheater

Abstract

The invention discloses a tubular air preheater for an atmospheric and vacuum heating furnace and a control method of the tubular air preheater, and relates to the field of preheaters. The control method comprises the following steps of periodically collecting various operation data of the tubular air preheater to form real-time input vectors and real-time output vectors, estimating a real-time state matrix of the tubular air preheater according to the plurality of real-time input vectors and the plurality of real-time output vectors, and calculating the real-time state matrix of the tubular air preheater according to the real-time state matrix. And the real-time state matrix is compared with a standard state matrix corresponding to each running state type, whether the real-time state matrix is matched with the standard state matrix or not is judged, and if yes, the running state of the tubular air preheater is the running state type corresponding to the standard state matrix. The running state of the tubular air preheater is estimated according to the easily measured state changes of the flue gas and the air before and after passing through the tubular air preheater, and the running state of the tubular air preheater is monitored at low cost.

Description

Technical Field

[0001] The present invention relates to the technical field of preheaters, and particularly to a tubular air preheater for an atmospheric and vacuum heating furnace and a control method thereof. Background Art

[0002] In the atmospheric and vacuum distillation system of petroleum refining, an air preheater is a key energy-saving device, mainly used to recover the waste heat in the flue gas discharged from the heating furnace and preheat the combustion-supporting air, thereby improving the combustion efficiency and reducing the energy consumption. Practice has proved that heating the air from room temperature to 200°C to 300°C through the waste heat of the flue gas can reduce the fuel consumption by about 5% to 10%.

[0003] Air preheaters are divided into two types. One is a rotary air preheater, which is mostly used in power plants and other occasions. The other is a tubular air preheater, whose structure is relatively simple. The flue gas heats the pipes through a heat exchange section with multiple pipes, and the cold air of the air is heated through the pipes. Since the flue gas needs to pass through multiple narrow pipes, and the flue gas contains a large amount of dust, after long-term use, the tubular air preheater is prone to partial pipe blockage due to ash accumulation or scaling, affecting the heat exchange efficiency. At the same time, the flue gas also contains a large amount of sulfides, which will form acid mist to varying degrees after passing through the tubular air preheater. After combining with the dust, it will stay on the pipes and corrode the pipes. When the corrosion is serious, the pipes will be penetrated, causing hazards such as flue gas leakage.

[0004] Since the tubular air preheater is in a high-temperature closed environment, its operating state is difficult to directly measure. Especially the internal space of the pipes is narrow and covered with a large amount of accumulated ash, and the corrosion state is even more difficult to directly measure. Currently, in the prior art, an analytical mathematical model of the tubular air preheater is mostly established to predict the ash accumulation and corrosion state of the tubular air preheater. This not only requires measuring a large amount of data, such as the mole coefficient of dry flue gas products, the mole coefficient of air volume, fuel composition parameters, etc., with a large workload and high cost, but also the states of the flue gas and air will change continuously over time, affecting the accuracy of the prediction results. Summary of the Invention

[0005] In order to solve the technical problem that it is difficult to establish an accurate mathematical model of the air preheater and thus it is impossible to accurately monitor the operating state of the air preheater, the present invention provides a tubular air preheater for an atmospheric and vacuum heating furnace and a control method thereof. Among them, the control method includes the following steps: Periodically collect various operating data of the tubular air preheater to form a real-time input vector and a real-time output vector; Estimate the real-time state matrix of the tubular air preheater according to a plurality of the real-time input vectors and a plurality of the real-time output vectors; Compare the real-time state matrix with the standard state matrix corresponding to each operating state type respectively; Determine whether the real-time state matrix matches the standard state matrix. If so, the operating state of the tubular air preheater is the operating state corresponding to the standard state matrix.

[0006] Estimate the real-time state matrix of the tubular air preheater through the easily measurable state changes of the flue gas and air before and after passing through the tubular air preheater, compare the real-time state matrix with the standard state matrix, judge the operating state of the tubular air preheater, realize the low-cost monitoring of the operating state of the tubular air preheater, so as to remind the maintenance personnel to maintain the tubular air preheater in time, reasonably arrange the maintenance time, and avoid affecting the preheating of the air entering the combustion furnace due to excessive ash accumulation or corrosion of the tubular air preheater.

[0007] Specifically, each component of the real-time input vector respectively represents: the temperature of the flue gas inlet side, the flow rate of the flue gas inlet side, the pressure of the flue gas inlet side, the cold air temperature, the cold air humidity, the cold air flow rate, the cold air pressure.

[0008] Specifically, each component of the real-time output vector respectively represents: the temperature of the flue gas outlet side, the flow rate of the flue gas outlet side, the pressure of the flue gas outlet side, the hot air temperature, the hot air humidity, the hot air flow rate, the hot air pressure.

[0009] Further, the real-time state matrix is obtained through the following steps: Construct a linear mathematical model of the tubular air preheater, and estimate the primary state matrix according to the real-time input vector and the real-time output vector; Substitute the primary state matrix into the linear mathematical model, and use the linear mathematical model and the real-time input vector to estimate the predicted output vector; Calculate the deviation vector between the predicted output vector and the real-time output vector, and construct a corrected linear mathematical model of the tubular air preheater according to the deviation vector; Use the corrected linear mathematical model, the real-time input vector and the real-time output vector to estimate the real-time state matrix.

[0010] By calculating the deviation vector and correcting the mathematical model of the tubular air preheater with the deviation vector, and then correcting the real-time state matrix, the accuracy of the real-time state matrix is improved.

[0011] Specifically, the standard state matrix is obtained through the following steps: Obtain historical operation data, and classify the historical operation data into multiple operation state types; For each of the operation state types, select the historical operation data that has not participated in the calculation, and construct a historical input vector and a historical output vector; Estimate the standard state matrix according to the historical input vector and the historical output vector.

[0012] Further, construct the standard state matrix through the following steps: Construct a linear mathematical model of the tubular air preheater, and estimate the initial state matrix and the minimum mean square error according to the historical input vector and the historical output vector; Judge whether the minimum mean square error is less than a first threshold. If so, the initial state matrix is the standard state matrix. If not, construct a nonlinear mathematical model of the tubular air preheater, and estimate the standard state matrix according to the historical input vector and the historical output vector.

[0013] When the standard state matrix calculated by using the linear mathematical model fluctuates greatly, switch to the nonlinear mathematical model to re-evaluate the standard state matrix, so as to improve the accuracy of the standard state matrix.

[0014] Specifically, construct the standard state matrix through the following steps: Construct a linear mathematical model of the tubular air preheater, and estimate the initial state matrix according to the historical input vector and the historical output vector; Substitute the initial state matrix into the linear mathematical model, and estimate the predicted output vector by using the mathematical model and the historical input vector; Calculate the deviation vector between the predicted output vector and the historical output vector, and construct a corrected linear mathematical model of the tubular air preheater according to the deviation vector; Estimate the standard state matrix by using the corrected linear mathematical model, the historical input vector and the historical output vector.

[0015] Specifically, judge whether the real-time state matrix matches the standard state matrix through the following steps: Subtract the real-time state matrix from the standard state matrix to obtain a difference matrix, and calculate the F-norm of the difference matrix; Compare the F-norm of the difference matrix with a second threshold. If the F-norm of the difference matrix is less than the second threshold, the real-time state matrix matches the standard state matrix.

[0016] Specifically, judge whether the real-time state matrix matches the standard state matrix through the following steps: Calculate the similarity between the real-time state matrix and the standard state matrix; Judge whether the similarity is greater than a third threshold. If so, the real-time state matrix matches the standard state matrix.

[0017] Further, after determining whether the real-time state matrix matches the standard state matrix, the similarity is output.

[0018] The present invention also provides a tubular air preheater for an atmospheric and vacuum heating furnace, which is controlled by the above control method.

[0019] The technical effects and advantages of the present invention: Estimate the real-time state matrix of the tubular air preheater through the easily measurable state changes of the flue gas and air before and after passing through the tubular air preheater, compare the real-time state matrix with the standard state matrix, judge the operating state of the tubular air preheater, and realize the low-cost monitoring of the operating state of the tubular air preheater, so as to remind the maintenance personnel to maintain the tubular air preheater in time, reasonably arrange the maintenance time, and avoid the influence of excessive ash accumulation or corrosion of the tubular air preheater on the preheating of the air entering the tubular heating furnace. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the overall flow of the control method provided by the present invention.

[0021] Figure 2 It is a schematic diagram of the mathematical model of the tubular air preheater constructed by the present invention.

[0022] Figure 3 It is a flowchart of the first acquisition method of the standard state matrix in the present invention.

[0023] Figure 4 It is a flowchart of the second acquisition method of the standard state matrix in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] Refer to Figure 1 , the control method for the tubular air preheater for an atmospheric and vacuum heating furnace provided by the present invention includes the following steps: Periodically collect various operation data of the tubular air preheater to form a real-time input vector and a real-time output vector; Estimate the real-time state matrix of the tubular air preheater according to multiple real-time input vectors and multiple real-time output vectors; Compare the real-time state matrix with the standard state matrix corresponding to each operating state type respectively; Determine whether the real-time state matrix matches the standard state matrix. If so, the operating state of the tubular air preheater is the operating state type corresponding to the standard state matrix.

[0026] Reference Figure 2 , for the system of the tubular air preheater (which can also be simply referred to as the air preheater), although its working state changes with time on a relatively long time scale, at a certain definite moment, when the input variables of the tubular air preheater are determined, its output variables are also determined, and its operating state is also determined. Therefore, its operating state can be inferred from the input variables and output variables.

[0027] Specifically, the mathematical model of the tubular air preheater can be the following formula:

[0028] In the formula, is the real-time output vector of the tubular air preheater, and each component represents an output variable. The output variables that are easy to measure during the operation of the tubular air preheater are: the temperature on the flue gas outlet side , the flow rate on the flue gas outlet side , the pressure on the flue gas outlet side , the hot air temperature , the hot air humidity , the hot air flow rate , the hot air pressure , then the real-time output vector is:

[0029] is the real-time input vector of the tubular air preheater, and each component represents an input variable. The input variables of the tubular air preheater mainly include the state of the flue gas entering the tubular air preheater and the state of the cold air. The input variables that are easy to measure among these two are: the temperature on the flue gas inlet side , the flow rate on the flue gas inlet side , the pressure on the flue gas inlet side , the cold air temperature , the cold air humidity , the cold air flow rate , the cold air pressure , then the real-time input vector is

[0030] The real-time input vector and the real-time output vector do not have to have the same dimension. When the dimension of one of the vectors is less, constant terms can be used to supplement the dimension.

[0031] is the real-time state matrix of the tubular air preheater, based on the real-time input vector created above and the real-time output vector . The real-time state matrix is a 7x7 matrix, containing 7x7 elements. It is impossible to calculate the real-time state matrix only through a set of real-time input vectors and real-time output vectors . Therefore, multiple sets of real-time input vectors and real-time output vectors are needed. The real-time state matrix is estimated by data fitting.

[0032] Specifically, assuming that the real-time input vector and the real-time output vector are linearly related, multiple linear regression analysis can be used to estimate the real-time state matrix . Assuming that the tubular air preheater has collected data of n sets of real-time input vectors and real-time output vectors , the real-time input data matrix and the real-time output data matrix Y are constructed as follows: ,

[0033] The real-time state matrix can be estimated by the following formula:

[0034] The above formula is a simple estimation method for the real-time state matrix .

[0035] When considering the complexity of the actual working state of the tubular air preheater and the deviation generated during the measurement of the real-time input vector and the real-time output vector, a deviation correction vector can be introduced when constructing the mathematical model of the tubular air preheater:

[0036] Correspondingly, the estimation method of the real-time state matrix also needs to be adjusted accordingly.

[0037] In the actual application process, non-linear regression can also be used to estimate the real-time state matrix , that is, assuming that the relationship between the input vector and the output vector is non-linear, a non-linear mathematical model of the tubular air preheater is established. This will result in a larger scale of the real-time state matrix and require more powerful data processing capabilities.

[0038] Specifically, the real-time state matrix can be estimated when every n groups of real-time input vectors and real-time output vectors are collected. It can also be estimated that every time a new real-time input vector and real-time output vector are collected after the first n groups of real-time input vectors and real-time output vectors are collected. In this case, the real-time state matrix is estimated based on the n groups of real-time input vectors and real-time output vectors with the latest time.

[0039] The processing method of the historical operation data of the tubular air preheater is the same as above. For the historical operation data of the tubular air preheater that has been collected during daily operation, since the operation state type of the tubular air preheater can be confirmed during routine maintenance of the tubular air preheater, the historical operation data can be classified based on the determined operation state type. For example, it can be classified into normal operation state and blockage state, or the severity of ash accumulation can be graded, and the historical operation data can be classified according to the severity of ash accumulation. The corrosion degree of the tubular air preheater can also be comprehensively considered, and the historical operation data can be classified as normal or in need of maintenance. Then, the historical operation data under each operation state type is respectively fitted to obtain the standard state matrix of the tubular air preheater under this operation state type.

[0040] When processing a large amount of historical operation data, if the linear hypothesis cannot fit all the data well, the linear hypothesis can be changed to a non-linear hypothesis to establish a non-linear mathematical model of the tubular air preheater, thereby improving the calculation accuracy of the state matrix.

[0041] Reference Figure 3 , when obtaining the standard state matrix , it can be assumed that there is a linear relationship between the historical input vector and the historical output vector first, and a linear mathematical model of the tubular air preheater is established accordingly. Then, multiple historical operation data that have not participated in the fitting are selected, including historical input data and historical output data, to construct a historical input vector and a historical output vector. The initial state matrix and the minimum mean square error are estimated using the multiple linear regression method based on the historical input vector and the historical output vector; It is judged whether the minimum mean square error is less than the first threshold. If so, the initial state matrix is designated as the standard state matrix , if not, assume that the historical input vector and the historical output vector are in a non - linear relationship, and use the polynomial regression method to calculate the standard state matrix .

[0042] The selection of the first threshold determines the fluctuation of the standard state matrix , that is, the sensitivity of the judgment of the operating state of the tubular air preheater, which can be adjusted according to the actual situation during the actual application process.

[0043] Furthermore, referring to Figure 4 , the following steps can also be used to obtain the standard state matrix : After establishing the linear mathematical model of the tubular air preheater, select multiple historical operating data that have not participated in fitting, construct the historical input vector and the historical output vector, and estimate the initial state matrix using the multiple linear regression method according to the historical input vector and the historical output vector , substitute the initial state matrix into the linear mathematical model of the tubular air preheater, use the mathematical model and the historical input vector to estimate the predicted output vector, and calculate the deviation vector between the predicted output vector and the historical output vector , and construct the corrected linear mathematical model of the tubular air preheater:

[0044] Use the corrected linear mathematical model of the tubular air preheater, the historical input vector and the historical output vector to estimate the standard state matrix .

[0045] By using the deviation vector to correct the initial state matrix obtained in one calculation , the accuracy of the standard state matrix can be further improved.

[0046] The estimation method of the real - time state matrix is the same as the estimation method of the standard state matrix, that is, the least mean square error can be used to determine whether the estimated standard state matrix is accurate enough, or the linear mathematical model can be corrected by calculating the deviation vector, and the standard state matrix is estimated for the second time, thereby improving the accuracy of the standard state matrix . The specific steps are as follows: Construct the linear mathematical model of the tubular air preheater, and estimate the primary state matrix according to the real - time input vector and the real - time output vector ; Substitute the primary state matrix into the linear mathematical model, and use the linear mathematical model and the real - time input vector Estimate the predicted output vector; Calculate the deviation vector between the predicted output vector and the real-time output vector and construct a modified linear mathematical model of the tubular air preheater based on the deviation vector; Use the modified linear mathematical model, the real-time input vector and the real-time output vector to estimate the standard state matrix.

[0047] Real-time state matrix There are various methods to compare with the standard state matrix For example, the real-time state matrix can be judged through the following steps and the standard state matrix whether they match: Subtract the real-time state matrix from the standard state matrix to obtain a difference matrix, and calculate the F-norm of the difference matrix; Compare the F-norm of the difference matrix with a second threshold. If the F-norm of the difference matrix is less than the second threshold, then the real-time state matrix matches the standard state matrix .

[0048] Furthermore, in the actual application process, simply prompting the maintenance personnel about the operating state of the tubular air preheater is not convenient for the maintenance personnel to make decisions. Considering that any method may have misjudgment when judging the operating state of the tubular air preheater, therefore, when prompting the maintenance personnel about the operating state of the tubular air preheater, the proximity between the real-time state matrix and the standard state matrix can also be provided. For example, the similarity between the two matrices is calculated using the following formula :

[0049] where is the F-norm of the real-time state matrix , is the F-norm of the standard state matrix .

[0050] After judging whether the real-time state matrix matches the standard state matrix , regardless of the result, the similarity is output so that the maintenance personnel can evaluate the judgment result and make a decision on whether to stop the machine for maintenance.

[0051] The similarity calculated by the above formula can also be directly used to judge the real-time state matrix and the standard state matrix Whether it matches, for example, when the similarity is greater than the third threshold, it is determined that the real-time state matrix matches the standard state matrix The size of the third threshold can be adjusted according to the need for sensitivity.

[0052] The present invention also provides a tubular air preheater for an atmospheric and vacuum heating furnace, which is controlled by the above control method to monitor the operating state of the tubular air preheater at low cost, so as to remind maintenance personnel to maintain the tubular air preheater in time, reasonably arrange the maintenance time, and avoid affecting the preheating of the air entering the tubular heating furnace due to excessive ash accumulation or corrosion of the tubular air preheater.

[0053] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A control method for a tubular air preheater of a normal and low pressure heating furnace, characterized in that: The following steps are involved: Periodically collect various operating data of the tubular air preheater to form a real-time input vector and a real-time output vector; estimating a real-time state matrix of the tubular air preheater according to a plurality of the real-time input vectors and a plurality of the real-time output vectors; Comparing the real-time state matrix with the standard state matrix corresponding to each operating state type respectively; It is determined whether the real-time state matrix matches the standard state matrix. If so, the operating state of the tubular air preheater is the operating state type corresponding to the standard state matrix.

2. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 1, characterized in that: Each component of the real-time input vector represents: smoke inlet side temperature, smoke inlet side flow rate, smoke inlet side pressure, cold air temperature, cold air humidity, cold air flow rate, and cold air pressure.

3. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 1, characterized in that: Each component of the real-time output vector represents: smoke outlet side temperature, smoke outlet side flow rate, smoke outlet side pressure, hot air temperature, hot air humidity, hot air flow rate, and hot air pressure.

4. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 1, characterized in that: The real-time state matrix is ​​obtained by the following steps: Constructing a linear mathematical model of the tubular air preheater, and estimating a state matrix according to the real-time input vector and the real-time output vector; Substituting the primary state matrix into the linear mathematical model, and using the linear mathematical model and the real-time input vector to estimate a predicted output vector; Calculating a deviation vector between the predicted output vector and the real-time output vector, and constructing a corrected linear mathematical model of the tubular air preheater according to the deviation vector; The real-time state matrix is ​​estimated using the modified linear mathematical model, the real-time input vector and the real-time output vector.

5. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 1, characterized in that: The standard state matrix is ​​obtained by the following steps: Acquire historical operation data, and classify the historical operation data into multiple operation status types; For each of the operation status types, select the historical operation data that is not involved in the calculation, and construct a historical input vector and a historical output vector; The standard state matrix corresponding to the operating state type is estimated according to the historical input vector and the historical output vector.

6. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 5, characterized in that: The standard state matrix is ​​constructed by the following steps: Constructing a linear mathematical model of the tubular air preheater, and estimating an initial state matrix and a minimum mean square error according to the historical input vector and the historical output vector; Determine whether the minimum mean square error is less than a first threshold value. If so, the initial state matrix is ​​the standard state matrix. If not, construct a nonlinear mathematical model of the tubular air preheater and estimate the standard state matrix based on the historical input vector and the historical output vector.

7. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 5, characterized in that: The standard state matrix is ​​constructed by the following steps: Constructing a linear mathematical model of the tubular air preheater, and estimating an initial state matrix according to the historical input vector and the historical output vector; Substituting the initial state matrix into the linear mathematical model, and using the mathematical model and the historical input vector to estimate the predicted output vector; Calculating a deviation vector between the predicted output vector and the historical output vector, and constructing a corrected linear mathematical model of the tubular air preheater according to the deviation vector; The standard state matrix is ​​estimated using the modified linear mathematical model, the historical input vector and the historical output vector.

8. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 1, characterized in that: The following steps are used to determine whether the real-time state matrix matches the standard state matrix: Subtracting the real-time state matrix from the standard state matrix to obtain a difference matrix, and calculating the F-norm of the difference matrix; The F-norm of the difference matrix is ​​compared with a second threshold value, and if the F-norm of the difference matrix is ​​less than the second threshold value, the real-time state matrix matches the standard state matrix.

9. The control method for a tubular air preheater for a normal and low pressure heating furnace according to claim 1, characterized in that: The following steps are used to determine whether the real-time state matrix matches the standard state matrix: Calculating the similarity between the real-time state matrix and the standard state matrix; It is determined whether the similarity is greater than a third threshold value, and if so, the real-time state matrix matches the standard state matrix.

10. A tubular air preheater for a normal and low pressure heating furnace, characterized in that: Control is performed using the control method described in any one of claims 1 to 9.

Images