A back pressure steam turbine test bench system and intelligent control method

Abstract

The present invention relates to the field of steam turbine technology, and specifically to a back-pressure steam turbine test bench system and an intelligent control method, comprising an energy supply system, a steam generation system, a steam turbine, a steam storage system, and a multi-stage mechanical steam recompression system, wherein the energy supply system provides electrical energy to the steam generation system, the steam generation system is used to provide steam thermal energy to the steam turbine, the steam storage system is arranged between the steam generation system and the steam turbine, and a multi-stage mechanical steam recompression system is arranged between the steam turbine and the steam storage system, and the back-pressure steam of the steam turbine is compressed, recovered, and stored in the steam storage system by intelligently controlling the multi-stage mechanical steam recompression system. The present invention realizes the recycling of back-pressure steam of a back-pressure steam turbine, saves energy, ensures the safety of the steam generation system and its pipelines, and improves life and economic benefits.

Description

Technical Field

[0001] The invention relates to the technical field of steam turbines, and in particular to a back-pressure steam turbine test bench system and an intelligent control method. Background Art

[0002] Steam turbines are machines that run at high speeds under high temperature and pressure, and their rotating parts are subjected to huge centrifugal forces. Steam turbine commissioning is an important step in ensuring high-quality commissioning of steam turbines, which is to test the quality and performance of the equipment's design, manufacturing, and installation through testing of the steam turbine. In order to ensure the quality of new steam turbine products or to meet the special requirements of users, steam turbine no-load mechanical operation tests must be carried out.

[0003] A large amount of steam will be discharged from the back-pressure steam turbine after the test run on the test bench. Since the test run of the steam turbine is discontinuous and the exhaust pressure values ​​corresponding to different steam turbines are different, it is difficult to recycle and reuse the exhaust steam. At present, the back-pressure steam discharged by each steam turbine manufacturer after the test run of the steam turbine they produce is usually discharged directly, resulting in energy waste and reduced economic benefits. Taking the test bench of a steam turbine plant as an example, more than 300 units are tested annually, wasting 7,000 tons of steam, causing an economic loss of about 2.5 million yuan. At the same time, direct steam discharge affects the appearance of the plant and the on-site standardization construction; the water vapor generated by the steam discharge may have an adverse effect on surrounding equipment and production.

[0004] Since the centrifugal force of the rotating parts of the steam turbine is proportional to the square of the speed, when the speed of the steam turbine exceeds 20% of the rated speed, the centrifugal force is close to 1.5 times the rated speed. At this time, the loose parts may become loose, and once the centrifugal force on the steam turbine rotor exceeds the strength limit allowed by the material, the parts will be damaged, such as blade breakage, dynamic and static friction, broken shaft and other serious accidents. Therefore, steam turbines are equipped with overspeed protection devices, which can activate the emergency shutoff device when the speed of the steam turbine exceeds the rated speed by 10%-12%, unload the safety oil, quickly close all the steam inlet valves of the steam turbine, cut off the gas source entering the steam turbine, and shut down the steam turbine. Through the steam turbine overspeed protection test, the accuracy, reliability and sensitivity of each overspeed protection action of the steam turbine can be checked.

[0005] When the test bench conducts an overspeed protection test on the back-pressure steam turbine, the emergency shut-off device of the test bench will be activated to quickly cut off the steam supply to the back-pressure steam turbine. Since all the steam inlet valves of the steam turbine are suddenly closed, the boiler and pipeline of the steam supply device of the test bench will be subjected to the reverse impact of steam. Depending on the production capacity, the test bench in the factory usually undertakes overspeed tests of hundreds or even thousands of steam turbine units every year. The reverse impact of steam will affect the life of the boiler and pipeline, affect the safety of the test bench system, and even cause safety accidents. Summary of the invention

[0006] In order to solve the technical problems: how to avoid the direct discharge of back-pressure steam from back-pressure steam turbine, causing energy waste and reducing economic benefits; how to avoid steam reverse impact on boilers and pipelines, ensure the safety of boilers and their pipelines, and increase their service life.

[0007] The present invention provides the following technical solutions:

[0008] On the one hand, the present invention provides a back-pressure steam turbine test bench system, comprising an energy supply system, a steam generation system, a steam turbine, a steam storage system, and a multi-stage mechanical steam recompression system, wherein the energy supply system is connected to the steam generation system, the steam generation system is connected to the steam turbine, the energy supply system provides electrical energy to the steam generation system, the steam generation system is used to provide steam thermal energy to the steam turbine, the steam storage system is arranged between the steam generation system and the steam turbine, a multi-stage mechanical steam recompression system is arranged between the steam turbine and the steam storage system, and the multi-stage mechanical steam recompression system is used to compress and recover the back-pressure steam of the steam turbine and store it in the steam storage system.

[0009] Furthermore, the energy supply system is a photovoltaic power generation system.

[0010] Furthermore, the steam generating system is an electric boiler.

[0011] Furthermore, the steam storage system is a steam storage tank.

[0012] Furthermore, the multi-stage mechanical vapor recompression system comprises a plurality of MVR evaporators of different levels, and the MVR evaporators of different levels are connected in sequence through steam pipes, wherein the input end of the first-stage MVR evaporator is connected to the back pressure output end of the steam turbine, and the output end of the last-stage MVR evaporator is connected to the input end of the steam storage tank.

[0013] Furthermore, a bypass steam pipeline leading to the steam storage tank is also provided on the steam pipeline between two adjacent MVR evaporators, and a switch valve is correspondingly provided on the bypass steam pipeline.

[0014] Furthermore, the back-pressure steam turbine test bench system also includes a drain and steam exhaust system, and the drain and steam exhaust system bypass is arranged on the steam pipeline between the steam turbine and the first-stage MVR evaporator.

[0015] On the other hand, the present invention provides an intelligent control method for a back-pressure steam turbine test bench system.

[0016] Step S1: Obtain the speed, power, steam inlet pressure, steam inlet temperature, steam inlet flow rate, and exhaust pressure of different steam turbines to form an original database, perform data cleaning on the original database, process missing values ​​and abnormal values, scale the data features, reduce the dimensions of the original database, and form a feature database;

[0017] Step S2: using the processed feature data as the input of the Transformer recurrent neural network model, training the Transformer recurrent neural network model;

[0018] Step S3: input the influencing factors of the exhaust pressure of the steam turbine under test into the Transformer recurrent neural network model in step S2 to obtain the exhaust pressure P0 of the steam turbine at the final setting time;

[0019] Step S4: Compare the exhaust steam pressure P0 of the test turbine with the steam pressure value P at the output end of each level of MVR device. i (where 0<i≤n) comparison, when P i ≤P0<P i+1 When the steam is compressed by the corresponding MVR stages, the PID controller is used to control the i-th MVR switch valve to move in advance, so that the switch is opened in advance, and the steam enters the steam storage system from the switch valve after being compressed by the corresponding MVR stages.

[0020] Furthermore, the specific method in step S2 is to adopt a method of predicting a single variable with multiple variables, that is, setting the speed, power, steam inlet pressure, steam inlet temperature, and steam inlet flow rate of the steam turbine as influencing factors of the exhaust pressure, and obtaining the exhaust pressure from the original database as the input of the encoder.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] (1) The back-pressure steam turbine test bench system of the present invention is equipped with a multi-stage mechanical steam recompression system and a steam storage system, so that the back-pressure steam of the steam turbine is no longer discharged directly into the atmosphere. The exhaust pressure of different steam turbines is different. The multi-stage mechanical steam recompression system intelligently controls the number of compression stages, compresses the steam to a set pressure value, and then enters the steam storage system, and then supplies it to the steam turbine as a power source for the steam turbine. Therefore, the present invention realizes the intelligent recycling of back-pressure steam, saves energy, and improves economic benefits; at the same time, it avoids direct steam discharge that affects the beauty of the plant area and the on-site standardization construction, and causes adverse effects on surrounding equipment and production.

[0023] (2) Since the steam is recycled, it will no longer reversely impact the electric boiler and its pipelines, thereby increasing the life of the electric boiler and its pipelines, avoiding safety accidents of the test bench system caused by the safety of the boiler and pipelines, and increasing the life of the test bench. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a structural block diagram of the back-pressure steam turbine test bench system of the present invention;

[0025] Figure 2 It is a schematic diagram of the Transformer recurrent neural network model used in the control method of the present invention. DETAILED DESCRIPTION

[0026] The technical solution of the present invention will be clearly described below in conjunction with the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention, and all other embodiments obtained by ordinary technicians in the field without making creative work are within the protection scope of the present invention.

[0027] like Figure 1 As shown, the present invention provides a back-pressure steam turbine test bench system, including an energy supply system, a steam generation system, and a steam turbine, wherein the output end of the energy supply system is connected to the input end of the steam generation system, and the output end of the steam generation system is connected to the input end of the steam turbine.

[0028] The energy supply system is used to generate electrical energy and transmit the electrical energy to the steam generation system to provide power for the steam generation system;

[0029] A steam generation system, used to convert the electrical energy provided by the energy supply system into steam heat energy and transmit it to the steam turbine;

[0030] Steam turbine is used to convert the steam thermal energy delivered by the steam generation system into mechanical energy to perform external work.

[0031] The back-pressure steam turbine test bench system also includes a steam storage system, which is arranged between the output end of the steam generation system and the input end of the steam turbine. The steam storage system is used to store the thermal energy steam generated by the steam generation system in the steam storage system and then transport it to the steam turbine.

[0032] The back-pressure steam turbine test bench system also includes a multi-stage mechanical steam recompression system, which is arranged between the steam turbine back-pressure output end and the steam storage system input end. The multi-stage mechanical steam recompression system is used to compress the steam turbine back-pressure steam step by step and then recycle it into the steam storage system.

[0033] The back-pressure steam turbine test bench system also includes a drain exhaust system, the drain exhaust system input end is connected to the steam turbine back pressure output end, the drain exhaust system bypass is arranged on the steam pipeline between the steam turbine and the multi-stage mechanical steam recompression system, and a first switch valve is arranged at the drain exhaust system input end. By controlling the first switch valve of the drain exhaust system, it is possible to choose to discharge the back pressure steam into the atmosphere or to allow the back pressure steam to enter the multi-stage mechanical steam recompression system.

[0034] Therefore, the back-pressure steam turbine test bench system of this embodiment can compress the steam originally discharged directly into the atmosphere at the back-pressure output end of the steam turbine step by step into high-pressure steam for intelligent recovery and reuse by the steam turbine, thereby saving energy, improving economic benefits, and at the same time reducing the adverse effects on the surrounding factory environment, surrounding equipment and production.

[0035] The energy supply system adopts a photovoltaic power generation system, which converts light energy into electrical energy and outputs it to the steam generation system. Photovoltaic energy is green, environmentally friendly and low-carbon energy. The steam generation system is an electric boiler, which is electrically connected to the photovoltaic power generation system. The electric boiler converts the electrical energy provided by the photovoltaic power generation system into thermal energy and outputs steam thermal energy to the outside.

[0036] The electric boiler and steam turbine are connected through a steam pipe. The steam heat energy output by the electric boiler enters the test turbine for the test bench to conduct no-load mechanical operation test of the steam turbine.

[0037] The steam storage system is a steam storage tank, which is arranged on the steam pipe between the electric boiler and the steam turbine. The steam storage tank is used to store the steam heat energy output by the electric boiler on the one hand, and to recover the steam output by the back pressure end of the steam turbine on the other hand.

[0038] The multi-stage mechanical steam recompression system is composed of multiple MVR evaporators of different levels. The steam output from the turbine back pressure is compressed and recovered by multiple MVR evaporators and enters the steam storage tank. Multiple MVR evaporators of different levels mechanically compress the steam, and the steam pressure output at each level is different.

[0039] The multiple MVR evaporators of different levels include a first-stage MVR evaporator, a second-stage MVR evaporator, a third-stage MVR evaporator, etc., which are connected in sequence. Taking the three-stage MVR evaporator as an example, the input end of the first-stage MVR evaporator is connected to the output end of the back pressure turbine, the output end of the first-stage MVR evaporator is connected to the input end of the second-stage MVR evaporator, the output end of the second-stage MVR evaporator is connected to the input end of the third-stage MVR evaporator, and the output end of the third-stage MVR evaporator is connected to the input end of the steam storage tank; in addition, the output end of the first-stage MVR evaporator is also connected to the input end of the steam storage tank through a steam pipeline, and a second switch valve is arranged in the pipeline; the output end of the second-stage MVR evaporator is also connected to the input end of the steam storage tank through a steam pipeline, and a third switch valve is arranged in the pipeline.

[0040] Therefore, the MVR technology is used to recompress the steam and re-enter the steam storage tank to avoid direct discharge of steam and cause energy waste; the back-pressure steam is recovered by the steam storage tank to ensure continuous steam supply during the test run while reducing the impact on the electric boiler and pipeline during overspeed shutdown, avoiding safety accidents of the test bench system caused by the safety of the electric boiler and pipeline, and increasing the life of the electric boiler and steam pipeline.

[0041] In addition, the exhaust pressure values ​​of different back-pressure steam turbine units are different. For steam turbines with different exhaust pressures, the MVR multi-stage compression technology is used to control the pressure value entering the steam storage tank. Different levels of switch valves correspond to different steam compression values. By opening or closing different switch valves, the different pressure values ​​of the steam entering the steam storage tank can be controlled.

[0042] This embodiment also provides an intelligent control method for a back-pressure steam turbine test bench system.

[0043] The recurrent neural network model in machine learning, such as the Transformer recurrent neural network model, is applied to the prediction of turbine exhaust pressure and the action setting of the switch valve of the MVR device to complete the intelligent control of steam compression.

[0044] The specific method is as follows:

[0045] Step S1: Obtain the speed, power, steam inlet pressure, steam inlet temperature, steam inlet flow rate, and exhaust pressure of different steam turbines to form an original database, perform data cleaning on the original database, process missing values ​​and abnormal values, scale the data features, reduce the dimensions of the original database, and form a feature database;

[0046] Step S2: using the processed feature data as the input of the Transformer recurrent neural network model, training the Transformer recurrent neural network model, specifically adopting a method of multivariate prediction of a single variable, setting the turbine speed, power, steam inlet pressure, steam inlet temperature, and steam inlet flow rate as the influencing factors of the exhaust pressure, and obtaining the exhaust pressure from the original database as the input of the encoder;

[0047] The Transformer recurrent neural network model is a deep learning model that uses a self-attention mechanism, which calculates the weight of each part of the input data and sums them up. Figure 2 As shown in the figure. The Transformer recurrent neural network model mainly consists of an encoder module and a decoder module. The encoder module of the Transformer recurrent neural network model is composed of several isomorphic encoders. The output of each encoder is the input of the next encoder. For the input of the Transformer recurrent neural network model, an embedding layer is used to map it to a low-dimensional space and add position information to the input. The encoder mainly consists of two parts, namely a multi-head attention module and a feedforward neural network. In addition, residual connections and layer normalization methods are used after each module to accelerate the convergence efficiency of the model.

[0048] The specific application of feature data mainly involves the encoder and decoder parts. In the encoder:

[0049] 1) The input sequence will be converted into a vector representation in a high-dimensional space. The initial features of the input sequence are obtained here, which is called embedding.

[0050] 2) Add positional encoding to these embeddings to obtain feature data of the input sequence.

[0051] 3) The feature data is input to the self-attention layer, which calculates the attention relevance scores of all elements in the sequence and outputs the weighted feature representation. The features here contain the contextual information of the entire sequence based on the previous step.

[0052] 4) The feature data is input into the feedforward neural network, and after two linear transformations and a nonlinear activation function, the features at each position are further transformed to capture complex data patterns. The first linear transformation transforms the sequence to a higher dimension, the activation function increases the nonlinear ability, and the second linear transformation transforms the data from the high-dimensional space to the original embedding dimension. Subsequently, residual connection and layer normalization are performed to avoid the problem of gradient vanishing in deep networks and stabilize the training process.

[0053] The process in the decoder is similar to that in the encoder. Feature data is the data after processing the original input. The purpose is to reduce the dimension of the original data and retain only the most important data information. Extracting the most representative data improves the model's understanding and generalization ability of the data.

[0054] Step S3: Input the influencing factors of the steam turbine under test into the Transformer recurrent neural network model in step S2 to obtain the exhaust pressure P0 of the steam turbine at the final setting time. The Transformer recurrent neural network model processes multiple input variable sequences based on the attention mechanism to capture the time dependency between multiple sequences, thereby improving the accuracy of the prediction.

[0055] Step S4: Compare the exhaust steam pressure P0 of the test turbine with the steam pressure value P at the output end of each level of MVR device. i (where 0<i≤n) comparison. When P i ≤P0<P i+1 When the steam is compressed by the corresponding MVR stages, the PID controller is used to control the i-th MVR switch valve to move in advance, so that the switch is opened in advance, and the steam enters the steam storage system from the switch valve after being compressed by the corresponding MVR stages.

[0056] The steam turbine test bench system conducts overspeed protection tests on steam turbines with different characteristic data. The exhaust pressures corresponding to steam turbines with different characteristic data are different. In addition, the test run is discontinuous, and the exhaust is discontinuous. In view of this feature, the present invention uses intelligent control of multi-stage mechanical steam recompression (MVR) technology to effectively utilize exhaust steam with different pressures, intelligently control the pressure entering the steam storage tank, avoid the waste of direct drainage exhaust steam, and provide continuous and stable steam for the test bench.

[0057] Take a steam turbine plant as an example. The plant's test bench works 150 days a year, tests more than 300 steam turbine units, and exhausts about 7,000 tons of steam each year. After the technical improvement of the present invention, the steam turbine test bench is equivalent to reducing the consumption of 500,000 cubic meters of natural gas each year, saving 2.5 million yuan in costs and reducing carbon emissions by 1,082 tons each year.

[0058] The above technical features constitute the best embodiment of the present invention, which has strong adaptability and best implementation effect. Non-essential technical features can be added or reduced according to actual needs to meet the needs of different situations.

[0059] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. An intelligent control method for a back-pressure steam turbine test bench system, characterized in that: The back-pressure steam turbine test bench system includes a multi-stage mechanical steam recompression system, which is a plurality of MVR evaporators of different levels, and the MVR evaporators of different levels are sequentially connected through steam pipes, wherein the input end of the first-stage MVR evaporator is connected to the back-pressure output end of the steam turbine, and the output end of the last-stage MVR evaporator is connected to the input end of the steam storage tank; a bypass steam pipe leading to the steam storage tank is also provided on the steam pipe between two adjacent MVR evaporators, and a switch valve is correspondingly provided on the bypass steam pipe; the intelligent control method includes the following steps: Step S1: Obtain the speed, power, steam inlet pressure, steam inlet temperature, steam inlet flow rate, and exhaust pressure of different steam turbines to form an original database, perform data cleaning on the original database, process missing values ​​and abnormal values, scale the data features, reduce the dimensions of the original database, and form a feature database; Step S2: using the processed feature data as the input of the Transformer recurrent neural network model, training the Transformer recurrent neural network model; Step S3: input the influencing factors of the exhaust pressure of the steam turbine under test into the Transformer recurrent neural network model in step S2 to obtain the exhaust pressure P0 of the steam turbine at the final setting time; Step S4: Compare the exhaust steam pressure P0 of the test turbine with the steam pressure value P at the output end of each level of MVR device. i Compare, i = 1, 2, 3..., when P i ≤P0<P i+1 When the steam is compressed by the corresponding MVR stages, the PID controller is used to control the i-th MVR switch valve to move in advance, so that the switch is opened in advance, and the steam enters the steam storage system from the switch valve after being compressed by the corresponding MVR stages.

2. The intelligent control method according to claim 1, characterized in that: The specific method in step S2 is to adopt a method of predicting a single variable with multiple variables, that is, setting the speed, power, steam inlet pressure, steam inlet temperature, and steam inlet flow rate of the steam turbine as the influencing factors of the exhaust pressure, and obtaining the exhaust pressure from the original database as the input of the encoder.

3. A back pressure steam turbine test bench system, characterized in that: Used to execute the intelligent control method described in any one of claims 1-2, the back-pressure steam turbine test bench system includes an energy supply system, a steam generation system, a steam turbine, a steam storage system, and a multi-stage mechanical steam recompression system, the energy supply system is connected to the steam generation system, the steam generation system is connected to the steam turbine, the energy supply system provides electrical energy to the steam generation system, the steam generation system is used to provide steam thermal energy to the steam turbine, the steam storage system is arranged between the steam generation system and the steam turbine, a multi-stage mechanical steam recompression system is arranged between the steam turbine and the steam storage system, and the multi-stage mechanical steam recompression system is used to compress and recover the back-pressure steam of the steam turbine and store it in the steam storage system.

4. The back-pressure steam turbine test bench system according to claim 3, characterized in that: The energy supply system is a photovoltaic power generation system.

5. The back-pressure steam turbine test bench system according to claim 3, characterized in that: The steam generation system is an electric boiler.

6. The back-pressure steam turbine test bench system according to claim 3, characterized in that: The steam storage system is a steam storage tank.

7. The back-pressure steam turbine test bench system according to claim 6, characterized in that: The back-pressure steam turbine test bench system further comprises a drain and steam exhaust system, wherein a bypass of the drain and steam exhaust system is arranged on a steam pipeline between the steam turbine and the first-stage MVR evaporator.

Images