Steam turbine operating pressure optimization method based on data completion

Abstract

A steam turbine operating pressure optimization method based on data completion, which relates to the field of steam turbine operating pressure optimization. It solves the problem that the existing steam turbine operating pressure optimization methods use historical data for optimization, and the optimization conclusion highly depends on whether the samples are complete. If the samples are incomplete, the theoretical optimal value of the operating pressure cannot be obtained. The present invention includes the following steps: S1. Calculate the unit heat consumption rate Hr corresponding to each group of historical operating data; S2. Construct a single-layer perceptron model; S3. Take the average value of the main steam pressure as the initial value of the main steam pressure x' to be corrected; use the single-layer perceptron model to calculate the unit heat consumption rate Hr' to be corrected; S4. Use lr to correct x', and obtain the corrected main steam pressure value x''; S5. Judge whether Hr'' is less than Hr'. If the result is yes, let x' = x'', and execute S4; if the result is no, execute S6; S6. Judge whether lr has been inverted. If the result is yes, output x'' and use it as the optimal operating pressure; if the result is no, invert lr so that lr = -lr, and execute S4.

Description

Technical Field

[0001] The invention relates to the field of steam turbine operating pressure optimization. Background Art

[0002] In order to save energy and reduce emissions, thermal power units need to cooperate with more new energy sources to be connected to the grid, which requires the steam turbine to maintain a safe and stable operating state. The optimization of operating pressure is the most direct way to improve the economic efficiency of steam turbine operation. Carrying out the "sliding pressure optimization" task is a common solution for optimizing the operating pressure of steam turbines. However, whether it is the traditional method of conducting proprietary tests or the method of directly using historical data to find the best, it is essentially necessary to operate the unit to the specified operating conditions, and then fine-tune the main steam pressure, and then compare the unit's heat rate index to find the optimal operating pressure. However, the sliding pressure optimization method of conducting proprietary tests has high time and operating costs and cannot adapt to the actual situation of a wide range of variable operating conditions of the unit; the sliding pressure optimization conclusion obtained by the method of finding the best using historical data is highly dependent on whether the sample is complete. In other words, if the historical data is indeed or the test process is incomplete, it will directly lead to the absence of the theoretical optimal value of the operating pressure in the sample.

[0003] Some scholars have tried to use the "micro-increase" method to solve the problem, but this method only describes the unit's operating status from a theoretical level and cannot take into account the operating details of the steam turbine. If the given operating pressure value does not conform to the actual operating conditions of the unit, it may undermine the safety and stability of the steam turbine. It is necessary to start from the idea of ​​"micro-increase" and study a solution that takes into account the operating status of the steam turbine.

[0004] Multilayer perceptron is a common algorithm for machine learning, and is often used to fit complex physical quantities. There are many cases where multilayer perceptron and its derivative algorithms (such as multilayer neural network) are applied to steam turbine performance modeling. The core idea is to use massive historical data to establish a data-driven model, which also has the problem of "highly dependent on whether the sample is complete". If the model is not reasonably configured due to hyperparameters during training, it is very easy to cause overfitting, resulting in almost no generalization ability of the model, which is obviously not conducive to the selection of the optimal pressure of the steam turbine.

[0005] In summary, it is necessary to study a method that can quickly, efficiently and stably calculate the optimal operating pressure by adding a small correction value without affecting the stability of the unit. Summary of the invention

[0006] The purpose of the present invention is to solve the problem that the existing method for optimizing the operating pressure of a steam turbine uses historical data for optimization, and the optimization conclusion is highly dependent on whether the sample is complete. If the sample is incomplete, the theoretical optimal value of the operating pressure cannot be obtained. The present invention provides a method for optimizing the operating pressure of a steam turbine based on data completion.

[0007] Steam turbine operating pressure optimization method based on data completion, the operating pressure optimization method includes the following steps:

[0008] S1. Collect N groups of historical operating data under a certain working condition through the DCS control system, and calculate the unit heat consumption rate Hr corresponding to each group of historical operating data; N is an integer;

[0009] S2. Use the main steam pressure x in the N groups of historical operating data and the N unit heat consumption rates Hr to construct a single-layer perceptron model;

[0010] S3. Take the average value of the main steam pressure x in the N groups of historical operating data to obtain the average value of the main steam pressure and use it as the initial value of the main steam pressure x' to be corrected; input the main steam pressure x' to be corrected into the single-layer perceptron model, and the single-layer perceptron model calculates the unit heat consumption rate Hr' to be corrected;

[0011] S4. Use the given pressure correction value lr to correct the main steam pressure x' to be corrected to obtain the corrected main steam pressure value x''; then input the corrected main steam pressure value x'' into the single-layer perceptron model, and the single-layer perceptron model calculates the corrected unit heat consumption rate Hr'';

[0012] S5. Judge whether Hr'' is less than Hr', if the result is yes, update x', let x' = x'', and execute step S4; if the result is no, execute step S6;

[0013] S6. Judge whether lr has been inverted, if the result is yes, output the corrected main steam pressure value x'' and use it as the optimal operating pressure; if the result is no, update lr, invert lr, so that lr = -lr, and execute step S4.

[0014] Preferably, in S4, the implementation method of using the given pressure correction value lr to correct the main steam pressure x' to be corrected to obtain the corrected main steam pressure value x'' is: x'' = x' + lr.

[0015] Preferably, each group of historical operating data includes the main steam pressure x, the main steam temperature T ms , the main steam flow rate F ms , the reheat steam pressure P hrh , the reheat steam temperature T hrh , the main feed water flow rate F fw , the main feed water temperature T fw , the extraction steam temperature T of the first stage 1 , the water temperature T at the inlet of the first high-pressure heater fi1 , the normal drain temperature T of the first high-pressure heater d1 , the water temperature T at the outlet of the first high-pressure heaterfo1 、The extraction steam pressure P of the second high-pressure heater 2 、The temperature T of the second-stage extraction steam 2 、The water temperature T at the inlet of the second high-pressure heater fi2 、The normal drain temperature T of the second high-pressure heater d2 、The water temperature T at the outlet of the second high-pressure heater fo2 、The temperature T of the desuperheating water for the superheated steam shsp 、The flow rate F of the desuperheating water for the superheated steam shsp 、The temperature T of the desuperheating water for the reheated steam rhsp 、The flow rate F of the desuperheating water for the reheated steam rhsp and the unit power P.

[0016] Preferably, in S1, the calculation method of the unit heat consumption rate Hr corresponding to each set of historical operation data is as follows:

[0017] S11. Obtain h fo1 、h fi1 、h 1 、h d1 、h fo2 、h fi2 、h 2 、h d2 、H ms 、H fw 、H hrh 、H crh 、H shsp and H rhsp ;

[0018] Wherein, h fo1 is the enthalpy of the water at the outlet of the first high-pressure heater of the steam turbine, h fi1 is the enthalpy of the water at the inlet of the first high-pressure heater of the steam turbine, h 1 is the extraction steam enthalpy of the first high-pressure heater of the steam turbine, h d1 is the normal drain enthalpy of the first high-pressure heater of the steam turbine, h fo2 is the enthalpy of the water at the outlet of the second high-pressure heater of the steam turbine, h fi2 is the enthalpy of the water at the inlet of the second high-pressure heater of the steam turbine, h 2 is the extraction steam enthalpy of the second high-pressure heater of the steam turbine, h d2 is the normal drain enthalpy of the second high-pressure heater of the steam turbine, H ms is the main steam enthalpy, H fw is the main feed water enthalpy, H hrh is the reheated steam enthalpy, H crh is the enthalpy of the cold reheat steam, H shsp is the enthalpy of the desuperheating water for the superheated steam, H rhsp is the enthalpy of the desuperheating water for the reheated steam;

[0019] S12. According to the main feed water flow rate F in each set of historical operation data fw , the main steam flow rate F ms , and h obtained in step S11 fo1 , h fi1 , h 1 , h d1 , h fo2 , h fi2 , h d2 and h 2 , calculate the reheater steam flow rate F hrh ;

[0020] F hrh = F ms - F 1 - F 2 Formula 1;

[0021] In Formula 1,

[0022] wherein, F 1 and F 2 respectively represent the first and second intermediate variables;

[0023] S13. According to the reheater steam flow rate F obtained in step S12 hrh , the calculation method of the unit heat consumption rate Hr corresponding to each set of historical operation data is:

[0024]

[0025] wherein, F crh = F hrh , F crh is the reheater cold section steam flow rate, F shsp is the desuperheating water flow rate of the superheated steam in the historical operation data, F rhsp is the desuperheating water flow rate of the reheater steam in the historical operation data.

[0026] Preferably, in S11, according to each set of historical operation data, the implementation manner of obtaining h fo1 , h fi1 , h 1 , h d1 , h fo2 , h fi2 , h 2 and h d2 is:

[0027] respectively according to T in the historical operation data fo1 , T fi1 , T 1 , T d1 , T fo2 , Tfi2 and T 2 and T d2 Look up the enthalpy - entropy chart of water and steam, and correspondingly obtain h fo1 , h fi1 , h 1 , h d1 , h fo2 , h fi2 , h 2 and h d2 ;

[0028] Among them, T 1 is the extraction steam temperature of the first stage, T fi1 is the water inlet temperature of the first high - pressure heater, T d1 is the normal drain temperature of the first high - pressure heater, T fo1 is the water outlet temperature of the first high - pressure heater, T 2 is the extraction steam temperature of the second stage, T fi2 is the water inlet temperature of the second high - pressure heater, T d2 is the normal drain temperature of the second high - pressure heater, T fo2 is the water outlet temperature of the second high - pressure heater.

[0029] Preferably, in S11, according to each set of historical operation data, the implementation methods of obtaining H ms , H fw , H hrh , H crh , H shsp and H rhsp are as follows:

[0030] According to the main steam pressure x and the main steam temperature T ms in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the main steam enthalpy H ms ;

[0031] According to the main feed - water temperature T fw in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the main feed - water enthalpy H fw ;

[0032] According to the reheat steam pressure P hrh and the reheat steam temperature T hrh in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the reheat steam enthalpy H hrh ;

[0033] According to the extraction steam pressure P 2 of the second high - pressure heater and the extraction steam temperature T 2 of the second stage in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the reheat cold - section steam enthalpy H crh ;

[0034] According to the temperature T of the desuperheating water of the superheated steam in the historical operation data shsp , look up the enthalpy-entropy chart of water and steam to obtain the enthalpy H of the desuperheating water of the superheated steam shsp ;

[0035] According to the temperature T of the desuperheating water of the reheated steam in the historical operation data rhsp , look up the enthalpy-entropy chart of water and steam to obtain the enthalpy H of the desuperheating water of the reheated steam rhsp .

[0036] Preferably, in S11, F crh = F 2 , T fi1 = T fo2 .

[0037] Preferably, in step S2, the expression of the single-layer perceptron model is:

[0038]

[0039] where a is the weight, b is the threshold, and e is the natural constant.

[0040] Preferably, in step S4, the value range of lr is from 0.001 to 0.1.

[0041] Preferably, in step S2, the implementation method of constructing a single-layer perceptron model by using the main steam pressure x in N groups of historical operation data and the heat rate Hr of N units is as follows:

[0042] Take the main steam pressure x in each group of historical operation data as the input of the single-layer perceptron model, and take the heat rate Hr of the unit corresponding to this group of historical operation data as the output of the single-layer perceptron model to train the single-layer perceptron model. During the training process, use the gradient descent method to determine the model parameters a and b, so as to complete the construction of the single-layer perceptron model; where a is the weight and b is the threshold.

[0043] The beneficial effects brought by the present invention are:

[0044] 1) Compared with the pressure optimization method that conducts special tests, the present invention continuously corrects the main steam pressure by using historical data, without the need to conduct special tests, saving time costs, labor and production costs, etc.;

[0045] 2) Compared with the traditional pressure optimization method based on a large amount of historical data, the present invention can continuously correct and complement the main steam pressure with very little historical operation data, so as to obtain the optimal operating pressure corresponding to the corresponding working conditions. The method of the present invention does not need to rely on a large number of samples and can obtain the optimal operating pressure based on a small number of samples. In specific applications, the optimal operating pressure is used to guide sliding pressure operation. In addition, the traditional pressure optimization method based on a large amount of historical data highly depends on the completeness of samples. If the optimal pressure does not exist in the historical data, it will affect the conclusion. The present invention solves the problem that the optimal pressure does not exist in the operation data by continuously correcting the main steam pressure with a given pressure correction value lr.

[0046] 3) In addition, compared with the traditional pressure optimization method based on "incremental power" heat consumption difference analysis, the present invention can more effectively maintain the stable operation of the steam turbine by constructing a single-layer perceptron model to fit the pressure-heat consumption rate relationship, bringing stable economic benefits and having great feasibility. By adding a small correction value, the present invention can quickly, efficiently and stably calculate the optimal operating pressure under the corresponding working conditions without affecting the stability of the unit. Brief Description of the Drawings

[0047] Figure 1 is a flowchart of the steam turbine operating pressure optimization method based on data completion according to the present invention. Detailed Embodiments

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to 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 the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0049] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0050] Embodiment 1:

[0051] Refer to Figure 1 To illustrate Embodiment 1, the steam turbine operating pressure optimization method based on data completion described in Embodiment 1 includes the following steps:

[0052] S1. Collect N groups of historical operation data under a certain working condition through the DCS control system, and calculate the unit heat consumption rate Hr corresponding to each group of historical operation data; N is an integer;

[0053] S2. Construct a single-layer perceptron model using the main steam pressure x in N sets of historical operation data and the heat rate Hr of N units.

[0054] S3. Take the mean value of the main steam pressure x in N sets of historical operation data to obtain the average value of the main steam pressure , and use it as the initial value of the main steam pressure x' to be corrected; input the main steam pressure x' to be corrected into the single-layer perceptron model, and the single-layer perceptron model calculates the heat rate Hr' of the unit to be corrected.

[0055] S4. Use the given pressure correction value lr to correct the main steam pressure x' to be corrected to obtain the corrected main steam pressure value x''; then input the corrected main steam pressure value x'' into the single-layer perceptron model, and the single-layer perceptron model calculates the corrected heat rate Hr'' of the unit.

[0056] S5. Judge whether Hr'' is less than Hr'. If the result is yes, update x', let x' = x'', and execute step S4; if the result is no, execute step S6.

[0057] S6. Judge whether lr has been reversed. If the result is yes, output the corrected main steam pressure value x'' and use it as the optimal operating pressure; if the result is no, update lr, reverse lr, so that lr = -lr, and execute step S4.

[0058] In this embodiment, for the steam turbine operating pressure optimization method based on data completion of the present invention, the main steam pressure is continuously corrected using historical data, without the need to carry out special tests, saving time costs, labor, production costs, etc.; it does not rely on a large number of historical samples, and the optimal operating pressure can be obtained based on a small number of samples; by adding a given pressure correction value lr, that is, a small correction value, it can quickly, efficiently and stably calculate the optimal operating pressure under corresponding working conditions on the basis of not affecting the stability of the unit; by judging whether lr has been reversed, the iteration can be normally exited to avoid infinite loops.

[0059] In specific applications, the value range of N is: not less than 10 sets of data, without an upper limit; the value range of the given pressure correction value lr is from 0.001 to 0.01, and its optimal value is 0.001; set the value of lr to 0.001 and use it as the variable for each revision, and continuously correct the main steam pressure with a small value. By adding a small correction value, it can quickly, efficiently and stably calculate the optimal operating pressure under corresponding working conditions on the basis of not affecting the stability of the unit.

[0060] Further, in S4, the implementation of correcting the main steam pressure x' to be corrected by using the given pressure correction value lr to obtain the corrected main steam pressure value x'' is: x'' = x' + lr.

[0061] Each set of historical operation data includes the main steam pressure x, the main steam temperature T ms , the main steam flow rate F ms , the reheat steam pressure P hrh , the reheat steam temperature T hrh , the main feed water flow rate F fw , the main feed water temperature T fw , the extraction steam temperature T of the first stage 1 , the water temperature T at the inlet of the first high-pressure heater fi1 , the normal drain water temperature T of the first high-pressure heater d1 , the water temperature T at the outlet of the first high-pressure heater fo1 , the extraction steam pressure P of the second high-pressure heater 2 , the extraction steam temperature T of the second stage 2 , the water temperature T at the inlet of the second high-pressure heater fi2 , the normal drain water temperature T of the second high-pressure heater d2 , the water temperature T at the outlet of the second high-pressure heater fo2 , the desuperheating water temperature T of the superheated steam shsp , the desuperheating water flow rate F of the superheated steam shsp , the desuperheating water temperature T of the reheat steam rhsp , the desuperheating water flow rate F of the reheat steam rhsp and the unit power P.

[0062] Specifically, in S1, the calculation method of the unit heat consumption rate Hr corresponding to each set of historical operation data is:

[0063] S11. Obtain h fo1 , h fi1 , h 1 , h d1 , h fo2 , h fi2 , h 2 , h d2 , H ms , H fw , H hrh , H crh , H shsp and H rhsp ;

[0064] Among them, h fo1 is the enthalpy of the water at the outlet of the first high-pressure heater of the steam turbine, h fi1 is the enthalpy of the water at the inlet of the first high-pressure heater of the steam turbine, h 1is the enthalpy of the extraction steam from the first high-pressure heater of the steam turbine, h d1 is the normal drain enthalpy of the first high-pressure heater of the steam turbine, h fo2 is the enthalpy of the outlet water of the second high-pressure heater of the steam turbine, h fi2 is the enthalpy of the inlet water of the second high-pressure heater of the steam turbine, h 2 is the enthalpy of the extraction steam from the second high-pressure heater of the steam turbine, h d2 is the normal drain enthalpy of the second high-pressure heater of the steam turbine, H ms is the main steam enthalpy, H fw is the main feed water enthalpy, H hrh is the reheated steam enthalpy, H crh is the enthalpy of the cold reheat steam, H shsp is the enthalpy of the desuperheating water for the superheated steam, H rhsp is the enthalpy of the desuperheating water for the reheated steam;

[0065] S12. According to the main feed water flow rate F fw , the main steam flow rate F ms , and h fo1 , h fi1 , h 1 , h d1 , h fo2 , h fi2 , h d2 and h 2 obtained in step S11, calculate the reheated steam flow rate F hrh ;

[0066] F hrh = F ms - F 1 - F 2 Formula 1;

[0067] In Formula 1,

[0068] wherein, F 1 and F 2 respectively represent the first and second intermediate variables;

[0069] S13. According to the reheated steam flow rate F hrh obtained in step S12, the calculation method of the unit heat consumption rate Hr corresponding to each set of historical operation data is:

[0070]

[0071] wherein, F crh = F hrh , F crh is the cold reheat steam flow rate, F shsp is the desuperheating water flow rate of the superheated steam in the historical operation datarhsp is the desuperheating water flow rate of the reheater steam in the historical operation data.

[0072] In this preferred embodiment, each set of historical operation data is used to obtain the unit heat rate Hr corresponding to this set of historical operation data.

[0073] In specific applications, in S11, according to each set of historical operation data, h fo1 、h fi1 、h 1 、h d1 、h fo2 、h fi2 、h 2 and h d2 are obtained as follows:

[0074] Respectively, according to T fo1 、T fi1 、T 1 、T d1 、T fo2 、T fi2 、T 2 and T d2 in the historical operation data, look up the enthalpy-entropy chart of water and steam, and correspondingly obtain h fo1 、h fi1 、h 1 、h d1 、h fo2 、h fi2 、h 2 and h d2 ;

[0075] wherein, T 1 is the extraction steam temperature of the first stage, T fi1 is the water inlet temperature of the first high-pressure heater, T d1 is the normal drain temperature of the first high-pressure heater, T fo1 is the water outlet temperature of the first high-pressure heater, T 2 is the extraction steam temperature of the second stage, T fi2 is the water inlet temperature of the second high-pressure heater, T d2 is the normal drain temperature of the second high-pressure heater, T fo2 is the water outlet temperature of the second high-pressure heater.

[0076] In this preferred embodiment, by looking up the enthalpy-entropy chart of water and steam to obtain the enthalpy corresponding to each type of temperature, the operation method is simple and easy to implement.

[0077] Furthermore, in S11, according to each set of historical operation data, H ms 、H fw 、H hrh 、Hcrh , H shsp and H rhsp is implemented as follows:

[0078] Based on the main steam pressure x and the main steam temperature T in the historical operation data ms , look up the enthalpy - entropy chart of water and steam to obtain the main steam enthalpy H ms ;

[0079] Based on the feed - water temperature T in the historical operation data fw , look up the enthalpy - entropy chart of water and steam to obtain the feed - water enthalpy H fw ;

[0080] Based on the reheat steam pressure P hrh and the reheat steam temperature T hrh in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the reheat steam enthalpy H hrh ;

[0081] Based on the extraction steam pressure P 2 of the second high - pressure heater and the second - stage extraction steam temperature T 2 in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the reheat cold - section steam enthalpy H crh ;

[0082] Based on the attemperating water temperature T shsp of the superheated steam in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the attemperating water enthalpy H shsp ;

[0083] Based on the attemperating water temperature T rhsp of the reheat steam in the historical operation data, look up the enthalpy - entropy chart of water and steam to obtain the attemperating water enthalpy H rhsp .

[0084] In this preferred embodiment, by looking up the enthalpy - entropy chart of water and steam to obtain the enthalpy corresponding to each type of temperature, the operation method is simple and easy to implement. Specifically, in S11, F crh = F 2 , T fi1 = T fo2 .

[0085] Specifically, in step S2, the expression of the single - layer perceptron model is:

[0086]

[0087] where a is the weight, b is the threshold, and e is the natural constant.

[0088] Specifically, in step S2, the implementation method of constructing a single-layer perceptron model by using the main steam pressure x in N groups of historical operation data and the heat rate Hr of N units is as follows:

[0089] Taking the main steam pressure x in each group of historical operation data as the input of the single-layer perceptron model, and the heat rate Hr of the unit corresponding to this group of historical operation data as the output of the single-layer perceptron model to train the single-layer perceptron model. During the training process, the gradient descent method is used to determine the model parameters a and b, thereby completing the construction of the single-layer perceptron model; where a is the weight and b is the threshold.

[0090] In this preferred embodiment, by constructing a single-layer perceptron model to fit the main steam pressure-heat rate relationship, the stable operation of the steam turbine can be maintained more effectively; in addition, by restricting an appropriate number of data to train the single-layer perceptron, problems such as model divergence and accuracy decline caused by too large or too small sample size are avoided.

[0091] Principle analysis: The steam turbine operation pressure optimization method based on data completion described in the present invention reconstructs the pressure-heat rate relationship based on the historical operation data of the unit under a certain working condition to obtain a single-layer perceptron model. Each group of historical operation data is used as a sample data. Based on a very small amount of sample data, the pressure is corrected by a given pressure correction value. Specifically, when the main steam pressure x' to be corrected is corrected for the first time, it is judged whether Hr″ is less than Hr′, so as to determine the change trend of the heat rate of the unit under the given pressure correction value. If the heat rate Hr″ of the unit develops in an increasing trend after the first correction, the given pressure correction value lr is reversed, and the main steam pressure is continuously corrected until the heat rate Hr″ of the unit no longer decreases after correction. It is determined whether lr has been reversed. If the result is yes, the corrected main steam pressure value x″ is output and used as the optimal operating pressure; if the heat rate Hr″ of the unit develops in a decreasing trend after the first correction, its pressure is continuously corrected until it no longer decreases. It is determined again whether lr has been reversed. If the result is yes, the corrected main steam pressure value x″ is output and used as the optimal operating pressure.

[0092] Although the present invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed, as long as they do not deviate from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the different dependent claims and the features described herein can be combined in a manner different from that described in the original claims. It should also be understood that the features described in connection with a single embodiment can be used in other described embodiments.

Claims

1. A method for optimizing the operating pressure of a steam turbine based on data completion, characterized in that, the operating pressure optimization method includes the following steps: S1. Collect N sets of historical operating data under a certain operating condition through the DCS control system, and calculate the unit heat rate Hr corresponding to each set of historical operating data; N is an integer; S2. Use the main steam pressure x in the N sets of historical operating data and the N unit heat rates Hr to construct a single-layer perceptron model; S3. Take the mean value of the main steam pressure x in N groups of historical operation data to obtain the average value of the main steam pressure and use it as the initial value of the main steam pressure x' to be corrected; the main steam pressure x' to be corrected is input into the single-layer perceptron model, and the single-layer perceptron model calculates the heat rate Hr' of the unit to be corrected; S4. Use the given pressure correction value lr to correct the main steam pressure x' to be corrected, and obtain the corrected main steam pressure value x"; then input the corrected main steam pressure value x" into the single-layer perceptron model, and the single-layer perceptron model calculates the corrected unit heat rate Hr"; S5. Judge whether Hr" is less than Hr', if the result is yes, update x', let x' = x", and execute step S4; if the result is no, execute step S6; S6. Judge whether lr has been inverted, if the result is yes, output the corrected main steam pressure value x", and use it as the optimal operating pressure; if the result is no, update lr, invert lr, so that lr = -lr, and execute step S4.

2. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 1, characterized in that, In S4, the implementation method of using the given pressure correction value lr to correct the main steam pressure x' to be corrected to obtain the corrected main steam pressure value x" is: x" = x' + lr.

3. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 1, characterized in that, Each set of historical operation data includes main steam pressure x, main steam temperature T ms , main steam flow rate F ms , reheat steam pressure P hrh , reheat steam temperature T hrh , main feed water flow rate F fw , main feed water temperature T fw , extraction steam temperature T of the first stage 1 , water temperature T at the inlet of the first high-pressure heater fi1 , normal drain water temperature T of the first high-pressure heater d1 , water temperature T at the outlet of the first high-pressure heater fo1 , extraction steam pressure P of the second high-pressure heater 2 , extraction steam temperature T of the second stage 2 , water temperature T at the inlet of the second high-pressure heater fi2 , normal drain water temperature T of the second high-pressure heater d2 , water temperature T at the outlet of the second high-pressure heater fo2 , desuperheating water temperature T of the superheated steam shsp , desuperheating water flow rate F of the superheated steam shsp , desuperheating water temperature T of the reheat steam rhsp , desuperheating water flow rate F of the reheat steam rhsp and unit power P.

4. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 1, characterized in that, In S1, the calculation method of the unit heat rate Hr corresponding to each set of historical operating data is: S11. Obtain h according to each set of historical operation data fo1 h fi1 h 1 h d1 h fo2 h fi2 h 2 h d2 H ms H fw H hrh H crh H shsp and H rhsp ; Among them, h fo1 is the enthalpy of the outlet water of the first high-pressure heater of the steam turbine, h fi1 is the enthalpy of the inlet water of the first high-pressure heater of the steam turbine, h 1 is the enthalpy of the extraction steam of the first high-pressure heater of the steam turbine, h d1 is the enthalpy of the normal drain water of the first high-pressure heater of the steam turbine, h fo2 is the enthalpy of the outlet water of the second high-pressure heater of the steam turbine, h fi2 is the enthalpy of the inlet water of the second high-pressure heater of the steam turbine, h 2 is the enthalpy of the extraction steam of the second high-pressure heater of the steam turbine, h d2 is the enthalpy of the normal drain water of the second high-pressure heater of the steam turbine, H ms is the main steam enthalpy, H fw is the main feed water enthalpy, H hrh is the reheated steam enthalpy, H crh is the enthalpy of the cold reheat steam, H shsp is the enthalpy of the desuperheating water of the superheated steam, H rhsp is the enthalpy of the desuperheating water of the reheated steam; S12. Calculate the reheater steam flow rate F fw based on the main feed water flow rate F ms , the main steam flow rate F fo1 , and the h fi1 , h 1 , h d1 , h fo2 , h fi2 , h d2 , and h 2 obtained in step S11 hrh ;​ F hrh = F ms - F 1 - F 2 Formula 1; In Formula 1, where, F 1 and F 2 represent the first and second intermediate variables, respectively; S13. The reheated steam flow rate F obtained according to step S12 hrh , the calculation method for obtaining the unit heat consumption rate Hr corresponding to each set of historical operation data is as follows: Among them, F crh = F hrh ,F crh is the reheater cold section steam flow rate, F shsp is the desuperheating water flow rate of the superheated steam in the historical operation data, F rhsp is the desuperheating water flow rate of the reheated steam in the historical operation data.

5. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 4, characterized in that, In S11, obtain h according to each set of historical operation data fo1 h fi1 h 1 h d1 h fo2 h fi2 h 2 and h d2 are implemented as follows: Look up the enthalpy-entropy charts of water and water vapor respectively according to T fo1 、T fi1 、T 1 、T d1 、T fo2 、T fi2 、T 2 and T d2 and obtain h fo1 、h fi1 、h 1 、h d1 、h fo2 、h fi2 、h 2 and h d2 ; Among them, T 1 is the temperature of a section of extraction steam, T fi1 is the water temperature at the inlet of the first high-pressure heater, T d1 is the normal drain temperature of the first high-pressure heater, T fo1 is the water temperature at the outlet of the first high-pressure heater, T 2 is the temperature of the second section of extraction steam, T fi2 is the water temperature at the inlet of the second high-pressure heater, T d2 is the normal drain temperature of the second high-pressure heater, T fo2 is the water temperature at the outlet of the second high-pressure heater.

6. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 4, characterized in that, In S11, obtain H according to each set of historical operation data ms and H fw and H hrh and H crh and H shsp and H rhsp are implemented as follows: According to the main steam pressure x and the main steam temperature T in the historical operation data ms , search the enthalpy-entropy chart of water and steam to obtain the main steam enthalpy H ms ; According to the main feed water temperature T in the historical operation data fw , look up the enthalpy-entropy chart of water and steam to obtain the main feed water enthalpy H fw ; According to the reheat steam pressure P in the historical operation data hrh and the reheat steam temperature T hrh , look up the enthalpy-entropy chart of water and steam to obtain the reheat steam enthalpy H hrh ; According to the extraction steam pressure P of the second high-pressure heater in the historical operation data 2 and the extraction steam temperature T of the second stage 2 , look up the enthalpy-entropy chart of water and steam to obtain the enthalpy H of the cold reheat steam crh ; According to the temperature T of the desuperheating water of the superheated steam in the historical operation data shsp , search the enthalpy-entropy chart of water and steam to obtain the enthalpy H of the desuperheating water of the superheated steam shsp ; According to the attemperating water temperature T of the reheated steam in the historical operation data rhsp , search the enthalpy-entropy chart of water and water vapor to obtain the enthalpy H of the attemperating water of the reheated steam rhsp .

7. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 4, characterized in that, In S11, F crh = F 2 , T fi1 = T fo2 .

8. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 1, characterized in that, In step S2, the expression of the single-layer perceptron model is: where a is the weight, b is the threshold, and e is the natural constant.

9. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 1, characterized in that, In step S4, the value range of lr is from 0.001 to 0.

1.

10. The method for optimizing the operating pressure of a steam turbine based on data completion according to claim 1, characterized in that, In step S2, the implementation method of using the main steam pressure x in the N sets of historical operating data and the N unit heat rates Hr to construct a single-layer perceptron model is: Take the main steam pressure x in each set of historical operation data as the input of the single-layer perceptron model, and use the unit heat rate Hr corresponding to this set of historical operation data as the output of the single-layer perceptron model to train the single-layer perceptron model. During the training process, use the gradient descent method to determine the model parameters a and b, so as to complete the construction of the single-layer perceptron model; where a is the weight and b is the threshold.

Images