Gas turbine OTC temperature calculation method and system based on MBR thermocouple
By generating the adversarial network based on the condition of the MBR thermocouple for temperature calibration, the problem of low sensing accuracy of gas turbines in high-temperature and high-noise environments is solved, and high-precision detection of OTC temperature and improved system stability are achieved.
Patent Information
- Application Number
- CN202411117214.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-15
AI Technical Summary
The existing gas turbine temperature monitoring system is difficult to ensure the sensing and detection accuracy in high-temperature and high-noise environments, resulting in low OTC operation accuracy and poor system stability.
Using the OTC temperature calculation method of gas turbine based on MBR thermocouple, by obtaining real-time smoke exhaust temperature and historical OTC temperature sequence and environmental parameter timing data, the conditions are used to generate an adversarial network for temperature calibration, dynamically adjust the calibration strategy to compensate for changes in environmental conditions and reduce electromagnetics Interference noise impact.
It improves the OTC temperature detection accuracy, provides more uniform temperature data, reduces electromagnetic interference noise error, realizes accurate temperature monitoring and calibration under complex working conditions, and supports accurate combustion control of the gas turbine control system.
Smart Images

Figure CN118626753B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of data analysis and processing, and in particular to a gas turbine OTC temperature calculation method and system based on MBR thermocouples. Background Art
[0002] A gas turbine is a heat engine that converts the heat energy released by fuel combustion into mechanical energy. Its efficient energy conversion capability makes it an important equipment in modern industry and transportation. Gas turbines mainly use temperature monitoring to implement OTC (Outlet Temperature Control). The stability and accuracy of OTC temperature can directly affect the overall control level of the gas turbine control system, such as the A-SMC system (Advanced CombustionMargin Control System for Gas Turbines).
[0003] At present, the temperature monitoring of gas turbines mainly relies on thermocouple sensors. Thermocouple sensors are temperature sensors that use the temperature difference between the contact points of two different materials to generate electromotive force. They have a simple structure, fast response, and a wide measurement range. Most gas turbine temperature monitoring systems rely on traditional temperature sensors, including common K-type, J-type and other ordinary thermocouples, which are often arranged between the combustion chamber and the turbine. Their working environment is complex, especially when the turbine is running at high speed, and electromagnetic noise is often frequent, which can easily cause large measurement errors.
[0004] Although the temperature detection accuracy can be improved to a certain extent by integrating the sensing parameters of multiple thermocouples, the sensing detection accuracy of thermocouples cannot be effectively guaranteed in a high temperature and high noise environment. They need to be effectively calibrated to meet the needs of high-precision OTC operations of the gas turbine control system.
[0005] In response to the above problems, the industry has not yet proposed a better technical solution. Summary of the invention
[0006] The present application provides a gas turbine OTC temperature calculation method, system, storage medium, computer program product and electronic device based on MBR thermocouple, which are used to at least solve the problem in the current related technology that the OTC temperature is directly set by the sensing parameters of the thermocouple without data calibration, resulting in low OTC operation accuracy and poor system stability.
[0007] In a first aspect, an embodiment of the present application provides a gas turbine OTC temperature calculation method based on an MBR thermocouple, comprising: obtaining an anti-surge valve opening and closing signal; in the case where it is determined that the anti-surge valve opening and closing signal is a closed state signal, collecting real-time exhaust gas temperature based on a first thermocouple array; the first thermocouple array includes a plurality of first thermocouples distributed in an area adjacent to the exhaust port, the first thermocouples use band response thermocouples, and each of the first thermocouples has a unique corresponding measurement wavelength; obtaining a historical OTC temperature sequence and environmental parameter time series data; the historical OTC temperature sequence includes a first time series of continuous OTC temperature sequences. The historical OTC temperature outputted in the second step, the environmental parameter time series data includes a plurality of environmental sensing parameters collected for the first thermocouple array in a continuous second time step, and the environmental sensing parameters include ambient temperature, ambient pressure, ambient humidity and electromagnetic interference intensity; the real-time smoke exhaust temperature, the historical OTC temperature sequence and the environmental parameter time series data are input into a temperature calibration model to output the corresponding real-time OTC temperature; the temperature calibration model adopts a conditional generative adversarial network, and the condition vector of the conditional generative adversarial network is determined according to the environmental parameter time series data and the historical OTC temperature sequence.
[0008] In a second aspect, an embodiment of the present application provides an OTC temperature calculation system for a gas turbine based on an MBR thermocouple, comprising: an anti-surge valve signal acquisition unit, for acquiring an anti-surge valve opening and closing signal; an MBR temperature acquisition unit, for acquiring real-time exhaust gas temperature based on a first thermocouple array when it is determined that the anti-surge valve opening and closing signal is a closed state signal; the first thermocouple array comprises a plurality of first thermocouples distributed in an area adjacent to the exhaust port, the first thermocouples adopt band response thermocouples, and each of the first thermocouples has a unique corresponding measurement wavelength; a reference timing acquisition unit, for acquiring historical OTC temperature sequences and environmental parameter timing data; the historical OTC temperature The sequence includes historical OTC temperatures output at continuous first time steps, the environmental parameter time series data includes multiple environmental sensing parameters collected for the first thermocouple array at continuous second time steps, and the environmental sensing parameters include ambient temperature, ambient pressure, ambient humidity and electromagnetic interference intensity; a real-time temperature output unit is used to input the real-time smoke exhaust temperature, the historical OTC temperature sequence and the environmental parameter time series data into a temperature calibration model to output the corresponding real-time OTC temperature; the temperature calibration model adopts a conditional generative adversarial network, and the condition vector of the conditional generative adversarial network is determined according to the environmental parameter time series data and the historical OTC temperature sequence.
[0009] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the steps of the gas turbine OTC temperature calculation method based on MBR thermocouples according to any embodiment of the present application.
[0010] In a fourth aspect, an embodiment of the present application provides a storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, the steps of the gas turbine OTC temperature calculation method based on MBR thermocouples of any embodiment of the present application are implemented.
[0011] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the gas turbine OTC temperature calculation method based on MBR thermocouples of any embodiment of the present application.
[0012] The gas turbine OTC temperature calculation method based on MBR thermocouple provided in this application can produce at least the following technical effects:
[0013] (1) For the sampled real-time exhaust temperature, the influence of environmental parameters on OTC temperature is learned based on cGAN (Conditional Generative Adversarial Network) by fusing the historical OTC temperature sequence and the environmental parameter time series data. The conditional vector of cGAN is defined by the environmental parameter time series data and the historical OTC temperature sequence, so that the generator can accurately refer to the time series dependency between environmental parameters and OTC temperature when generating OTC temperature data, and dynamically adjust the calibration strategy for temperature measurement accordingly, effectively compensating for the measurement error caused by changes in environmental conditions, thereby improving the detection accuracy of OTC temperature.
[0014] (2) In terms of signal sampling, compared with installing thermocouples in the area between the combustion chamber and the turbine or at the combustion chamber outlet, the present technical solution uses the first thermocouple array adjacent to the exhaust port for sampling, which can provide more uniform temperature data, and use it to calculate the OTC temperature in the stable operation stage after the anti-surge valve is closed, effectively reducing the impact of electromagnetic interference noise generated by the high-speed rotation of the turbine on the temperature sensor detection parameters. In addition, by taking the electromagnetic interference intensity as an environmental parameter and incorporating it into the conditional vector of the conditional generative adversarial network, the error caused by electromagnetic noise can be further reduced. In addition, by arranging an MBR (Multi-Band Response) thermocouple array, each band response thermocouple can capture infrared radiation signals of different wavelengths, reducing the deviation caused by single wavelength measurement, thereby providing multi-spectrum temperature measurement data, which can more accurately capture subtle changes in exhaust temperature.
[0015] Through this technical solution, the multi-spectral sampling of the MBR thermocouple array, the cGAN model and the multi-dimensional environmental data are used for time-series dynamic fusion to achieve accurate OTC temperature monitoring and calibration, which can provide more refined temperature data support for the control system of the gas turbine and lay a foundation for the subsequent control system to implement precise combustion control and regulation. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0017] Figure 1 A flowchart showing an example of a method for calculating the OTC temperature of a gas turbine based on an MBR thermocouple according to an embodiment of the present application is shown;
[0018] Figure 2 A schematic diagram of the structural connection of an example of a gas turbine suitable for applying the gas turbine OTC temperature calculation method of the embodiment of the present application is shown;
[0019] Figure 3 A schematic diagram showing a structural connection of an example of a temperature calibration model according to an embodiment of the present application is shown;
[0020] Figure 4 An operation flow chart of an example of collecting real-time exhaust temperature based on a first thermocouple array according to an embodiment of the present application is shown;
[0021] Figure 5A structural block diagram of an example of a gas turbine OTC temperature calculation system based on MBR thermocouples according to an embodiment of the present application is shown;
[0022] Figure 6 It is a schematic structural diagram of an embodiment of an electronic device of the present application. DETAILED DESCRIPTION
[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0024] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved shall comply with the provisions of relevant laws and regulations and shall not violate public order and good morals.
[0025] Figure 1 A flowchart of an example of a method for calculating the OTC temperature of a gas turbine based on an MBR thermocouple according to an embodiment of the present application is shown.
[0026] Regarding the executor of the method of the embodiment of the present application, it can be any controller or processor with computing or processing capabilities. Through accurate MBR thermocouple temperature sampling, comprehensive utilization of multimodal environmental parameters and application of the cGAN model, the measurement accuracy of the gas turbine OTC temperature and the dynamic adaptability of the system are significantly improved.
[0027] In some examples, it can be an OTC temperature calculation platform, and can be integrated and configured in an electronic device or terminal through software, hardware, or a combination of software and hardware, and the type of terminal or electronic device can be diverse, such as a mobile phone, tablet computer, or desktop computer, etc.
[0028] like Figure 1 As shown, in step S110, the anti-surge valve opening and closing signal is obtained.
[0029] It should be noted that during the operation of the gas turbine, the anti-surge valve is used to prevent the gas turbine from surging under certain operating conditions. The status signal of the anti-surge valve can be obtained by a sensor installed on the valve. When the valve is closed, the sensor outputs a corresponding closed state signal. By detecting the status signal of the anti-surge valve, the working state of the gas turbine can be accurately determined, such as a stable operating state or other state.
[0030] In step S120, when it is determined that the anti-surge valve opening and closing signal is a closed state signal, the real-time exhaust gas temperature is collected based on the first thermocouple array.
[0031] Figure 2 FIG. 1 is a schematic diagram showing a structural connection of an example of a gas turbine suitable for applying the gas turbine OTC temperature calculation method of the embodiment of the present application. Figure 2 As shown, according to the direction of air flow, the gas turbine includes a compressor, a combustion chamber, a turbine and a diffuser, and finally discharges the gas through the exhaust port, for example, the exhaust will flow to the boiler, etc. Here, the compressor is responsible for compressing the air and sending it into the combustion chamber. The air is mixed and burned with the fuel in the combustion chamber. The high-temperature and high-pressure gas generated by the combustion drives the turbine to rotate and output mechanical energy. The diffuser can slow down and expand the flowing gas and reduce the air flow speed. It should be noted that in some current non-public or potential research technologies, a thermocouple is set between the combustion chamber and the turbine, for example, by setting a thermocouple in the second area plane to sample the combustion chamber exhaust temperature as the OTC temperature, but in the high-speed operation stage of the turbine, for example, the FSNL (Full Speed No Load) stage, the electromagnetic interference generated by the turbine will affect the temperature sampling accuracy.
[0032] According to the embodiment of the present application, a plurality of first thermocouples are arranged in a first area adjacent to the exhaust port to form a first thermocouple array, each first thermocouple is a band-responsive thermocouple, and each first thermocouple has a unique corresponding measurement wavelength, thereby setting an MBR thermocouple array in the first area plane.
[0033] It should be noted that when the anti-surge valve is in the closed state, the gas turbine has reached a stable operating state from startup, such as the FSNL stage. At this time, calling the MBR thermocouple array sampling temperature information of the first area plane can effectively reduce the electromagnetic noise level, and the plane temperature part will be more uniform, achieving high-quality temperature sampling information. Furthermore, by using the MBR thermocouple array, the exhaust temperature change information at multiple wavelengths can be obtained simultaneously, realizing the refined measurement of the temperature field of the entire first area plane.
[0034] In step S130, historical OTC temperature series and environmental parameter time series data are obtained.
[0035] Here, the historical OTC temperature sequence includes the historical OTC temperature output at the continuous first time step, and the environmental parameter time series data includes multiple environmental sensor parameters collected for the first thermocouple array at the continuous second time step, and the environmental sensor parameters include ambient temperature, ambient pressure, ambient humidity and electromagnetic interference intensity.
[0036] It should be noted that changes in ambient temperature will directly affect the measurement accuracy of thermocouples. The output potential (or, thermoelectromotive force) of a thermocouple is determined by the temperature difference between the measuring junction and the reference junction. If the ambient temperature changes, the temperature of the reference junction will also change, thus affecting the output signal of the thermocouple. Changes in ambient pressure will also affect the thermoelectromotive force generation process of the thermocouple. Specifically, pressure changes will change the density and thermal conductivity of the air, thereby affecting the response speed and heat exchange process of the thermocouple. Changes in ambient humidity will affect the surface state of the thermocouple and the generation of thermoelectromotive force. For example, when the humidity is high, the water vapor content in the air increases, which may cause condensation on the surface of the thermocouple and change the thermal conductivity of the thermocouple. The electromagnetic interference intensity refers to the level of interference caused by the electromagnetic field in the environment to the thermocouple signal, especially the interference to the transmission of the sensing signal, which may cause fluctuations or offsets in the thermocouple signal, resulting in inaccurate temperature measurement.
[0037] In some embodiments, environmental sensors such as temperature sensors, pressure sensors, humidity sensors, and electromagnetic interference intensity sensors are arranged near the exhaust port of the gas turbine or the first thermocouple array, and the data of each sensor is transmitted to the main control system through a data transmission network for recording and storage.
[0038] More specifically, by extracting historical OTC temperature sequences and environmental parameter time series data from the historical data storage module of the main control system, including the OTC temperature and environmental sensing parameters continuously recorded in adjacent time periods, it is used to indicate the changing pattern of OTC temperature under recent environmental conditions.
[0039] In step S140, the real-time exhaust temperature, the historical OTC temperature sequence and the environmental parameter time series data are input into the temperature calibration model to output the corresponding real-time OTC temperature.
[0040] Here, the temperature calibration model uses a conditional generative adversarial network, which includes a generator and a discriminator. The generator attempts to generate realistic temperature calibration results, while the discriminator is used to distinguish the generated results from the real results. The discriminator generally only plays a role in the training phase and is not used in the application reasoning phase of the network.
[0041] It should be noted that the conditional vector of the conditional generative adversarial network is determined based on the environmental parameter time series data and the historical OTC temperature sequence. In this way, the model can accurately calibrate the sensed temperature signal while taking into account the recent changes in environmental parameters and the historical output OTC temperature. It supports reducing the measurement errors caused by various environmental interferences under complex working conditions, provides more reliable temperature data calibration services, and guarantees the precise control of gas turbines.
[0042] Figure 3A schematic diagram of the structural connection of an example of a temperature calibration model according to an embodiment of the present application is shown.
[0043] like Figure 3 As shown, the temperature calibration model 300 includes an input layer 310, a conditional feature embedding module 320, a timing analysis module 330, a convolution module 340 and an output layer 350.
[0044] The input layer 310 is used to receive real-time exhaust temperature, historical OTC temperature series and environmental parameter time series data.
[0045] In some embodiments, the input layer 310 standardizes and normalizes the data and processes them through a unified input layer, thereby ensuring that data from different sources and scales can be processed consistently by the system and reducing model instability caused by differences in data range and scale.
[0046] The conditional feature embedding module 320 is used to embed the historical OTC temperature series and environmental parameter time series data into a high-dimensional feature space to determine the corresponding initial condition feature vector.
[0047] In some embodiments, the conditional feature embedding module 320 converts the historical OTC temperature sequence and environmental parameter time series data into an embedding vector in a high-dimensional feature space. Specifically, a deep neural network or a pre-trained embedding model can be used to extract features from the historical OTC temperature sequence and environmental parameter time series data, map the original time series data into a high-dimensional space, capture its inherent complex relationships and features, and enable the model to better understand the impact of environmental conditions on temperature changes, so that the generated initial condition feature vector contains the time series pattern information of the historical temperature and environmental parameters, which is more conducive to achieving refined calibration.
[0048] The timing analysis module 330 is used to extract the time dependency relationship in the initial condition feature vector to obtain a corresponding timing dependency condition vector.
[0049] Here, the timing analysis module 330 can use various timing analysis models, such as recurrent neural network (RNN), long short-term memory network (LSTM) or gated recurrent unit (GRU) and other timing models to process timing data, so as to enrich the association pattern information between the environment and the OTC temperature in the conditional vector. As a result, the model can effectively grasp the change law of temperature data over time, especially when there are periodic changes or delay effects in the temperature data, and can support more refined prediction and calibration effects through time-dependent features.
[0050] The convolution module 340 is used to fuse and convolve the time series dependent condition vector and the real-time smoke exhaust temperature to obtain corresponding fused convolution features.
[0051] In some embodiments, the convolution module 340 may fuse the time-dependent condition vector with the real-time smoke exhaust temperature, and perform convolution processing on the fused data. Specifically, a one-dimensional or two-dimensional convolutional neural network (CNN) may be used to process the fused features, and the local patterns and features in the fused features may be extracted through the convolution operation, thereby obtaining the corresponding fused convolution features. More preferably, the convolution layer may use convolution kernels of different sizes to capture the local and global relationships between features. Thus, through the fusion processing of the convolution module, the current sensor sampling information and the feature time-series correlation pattern of the adjacent time period can be considered simultaneously, thereby improving the ability to capture complex nonlinear relationships.
[0052] The output layer 350 is used to generate real-time OTC temperature based on the fused convolutional features.
[0053] In some embodiments, the output layer 350 may use a fully connected layer to map the high-dimensional fused convolutional features to the target output, thereby obtaining the corresponding real-time OTC temperature, thereby achieving calibration of the current temperature sensing data combined with the environmental time series association pattern, which can effectively improve the accuracy of the detection results.
[0054] Based on the architecture of the temperature calibration model provided in the embodiment of the present application, complex features and patterns are extracted from the input data, and the multidimensionality and timing of the data are taken into account through multi-level calculations such as embedding, timing analysis and convolution processing, so that the model can better adapt to complex working environments and support more refined calibration effects on measured temperature data.
[0055] It should be noted that in the process of practicing this application, one or more modules in cGAN were optimized and designed, and better performance was achieved.
[0056] In some examples of the embodiments of the present application, the conditional feature embedding module 320 processes the historical OTC temperature sequence and environmental parameter time series data through a multimodal attention mechanism to generate an initial condition feature vector.
[0057] Specifically, the multimodal attention mechanism is adopted to dynamically calculate the correlation between the environmental parameters and the historical OTC temperature series at each time step, generate attention weights, and ensure that the model can automatically focus on the information that is most important for the current task during the feature extraction process, and generate a comprehensive feature representation to accurately reflect their importance under different conditions.
[0058] More specifically, multimodal attention calculation is performed in the following way:
[0059] , Formula (1)
[0060] , Formula (2)
[0061] , Formula (3)
[0062] In the formula, Indicates the corresponding The environmental parameter vector for each time step; Represents the historical OTC temperature series; represents the transpose of the trainable weight vector; For the corresponding The unnormalized score of the attention weight at time steps, denoted and relevance; Represents a trainable weight matrix, which is used to transform the The environmental parameter vector of each time step is embedded into the attention calculation space; represents a trainable weight matrix used to embed the historical OTC temperature series into the attention calculation space; represents the result of exponentially calculating and summing the unnormalized scores of all time steps, Indicates the total number of time steps corresponding to the environmental parameter time series data, Indicates The attention weights for time steps; represents the hyperbolic tangent activation function; represents the initial condition eigenvector; Represents an embedding operation to and Combine to generate feature representation.
[0063] Here, by introducing the attention weight, the model can dynamically adjust the importance of each input feature. Different environmental parameters and historical temperature data may have different effects on the final temperature calibration. Through the attention mechanism, the model can highlight those features that are more important under current conditions, thereby filtering out redundant information and improving the accuracy of feature extraction.
[0064] It should be noted that the attention mechanism not only considers the data of a single time step, but also comprehensively considers the environmental parameters and historical data of multiple time steps. Through the weighted average of attention weights, the model can effectively integrate multimodal information (including environmental and historical data) to form a more comprehensive initial condition feature vector , which helps the model better understand and utilize data from different sources and improve the overall prediction performance.
[0065] In some examples of the embodiments of the present application, the timing analysis module 330 adopts a bidirectional LSTM layer to process the attention condition vector through forward and backward processing to obtain a corresponding timing dependency condition vector.
[0066] Here, the bidirectional LSTM layer simultaneously processes the forward and backward information of the time series data, which can capture the complex time dependencies in the sequence data. Since temperature changes are not only affected by the current environment and historical data, but may also be affected by the feedback effects of future conditions, bidirectional processing enables the model to more comprehensively understand the time series characteristics in the temperature data, enhancing the ability to model complex dynamic changes.
[0067] More specifically, the timing dependency condition vector is extracted in the following way:
[0068] , Formula (4)
[0069] , Formula (5)
[0070] , Formula (6)
[0071] In the formula, Indicates that the forward LSTM The hidden state of time steps; represents long short-term memory network processing, represents the initial condition eigenvector Corresponding to The conditional eigenvector of time steps, Indicates that the forward LSTM The hidden state of time steps, Indicates that the reverse LSTM The hidden state of time steps, Indicates that the reverse LSTM The hidden state of time steps; Indicates The bidirectional feature representation of time steps combines the hidden states of the forward and backward LSTMs.
[0072] Through this embodiment, the combined feature representation is used , retaining information from two time directions, making the feature expression more comprehensive. The model can not only learn the short-term dependencies of the data, but also capture long-term trends and patterns, and can more accurately predict future temperature changes. In addition, by combining forward and backward information, the model shows better robustness when facing incomplete or partially missing data. For example, under certain environmental conditions, some historical data may be missing or incomplete, and the bidirectional LSTM can still supplement the missing feature expression through backward information.
[0073] In some examples of the embodiments of the present application, the convolution module 340 uses a deep convolution layer to gradually process the time-dependent condition vector and the real-time smoke exhaust temperature through multiple convolution layers to obtain deep fusion convolution features.
[0074] It should be noted that, through the bidirectional feature representation generated by the time series analysis module 330, the convolution module 340 can further extract local features, which are not limited to the data at a single time point, but also include local patterns and trends of the time series data. As a result, the convolution operation can capture the subtle changes and detailed features in the temperature data, which is more conducive to achieving refined temperature calibration.
[0075] More specifically, the timing analysis module can extract deep fusion convolutional features in the following ways:
[0076] , Formula (7)
[0077] , Formula (8)
[0078] In the formula, represents the fused feature vector, represents the vector concatenation operation, Indicates real-time exhaust temperature data. Represents the temporal dependency conditional vector output by the bidirectional LSTM layer; It is The output of the convolutional layer, which represents the feature representation processed by the convolution operation and activation function; represents the LeakyReLU activation function, represents the convolution operation, and Respectively represent The weight matrix and bias term of the convolutional layer.
[0079] Through this embodiment, the time-dependent conditional vector of the bidirectional LSTM layer is spliced with the real-time smoke exhaust temperature data to obtain a fused feature vector. The model can comprehensively consider historical data and real-time status, ensuring that the model can simultaneously utilize current real-time information and historical trends, thereby improving prediction accuracy. In addition, through multiple layers of convolutional layers, the model can deeply extract local patterns in feature representations, such as temperature change trends and local fluctuations at specific time steps, which helps the model identify subtle changes in temperature data and improve the accuracy of temperature calibration. Furthermore, the LeakyReLU activation function is used after the convolution operation. By retaining the gradient of negative values, the model can handle nonlinear features and prevent the "dead ReLU" problem, ensuring that the model can learn more features during training.
[0080] The output layer 350 is used to process the deep fused convolutional features to output the corresponding real-time OTC temperature.
[0081] More specifically, global average pooling and fully connected feature maps are applied on the feature maps output by the convolutional module:
[0082] , Formula (9)
[0083] , Formula (10)
[0084] In the formula, is the feature representation of the final output of the convolution module, n is the total number of convolutional layers in the convolutional module; represents the feature vector after pooling, represents the global average pooling operation; Indicates the real-time OTC temperature output. and Represent the weight matrix and bias term of the output layer respectively.
[0085] Here, through global average pooling, all information of the feature map is summarized into a global feature vector, which effectively reduces the feature dimension while retaining the overall information. This not only reduces the number of parameters and prevents overfitting, but also improves the model's ability to capture global information. Then, the final OTC temperature data is generated through full connection layer mapping, allowing the model to accurately predict the real-time OTC temperature.
[0086] Regarding the training of the temperature calibration model, it can adopt or refer to the training details of various cGANs, and can adopt some common loss functions of cGANs. In some examples of the embodiments of the present application, the discriminator can adopt the commonly used basic discriminator loss, which is as follows:
[0087] , Formula (11)
[0088] In the formula, represents the discriminator loss, which is used to distinguish whether the input data is real data or generated data.
[0089] As a further preferred embodiment, the generator loss of the temperature calibration model can be integrated with the spectrum consistency loss and the sample adaptation loss based on the original basic generator loss. and the base discriminator loss , which can ensure that the generated data is considered as real data by the discriminator under conditional features, and distinguish between real data and generated data. Through spectrum consistency loss, it can ensure that the generated data is consistent with the real data in frequency domain characteristics, especially taking into account the periodicity and fluctuation characteristics of temperature data. Through sample adaptive loss, the importance of different samples is emphasized, so that the model can better focus on key data points and optimize the temperature calibration effect.
[0090] More specifically, the loss function of the generator of the temperature calibration model is for:
[0091] , Formula (12)
[0092] , Formula (13)
[0093] , Formula (14)
[0094] , Formula (15)
[0095] , Formula (16)
[0096] , Formula (17)
[0097] , Formula (18)
[0098] In the formula, represents the base generator loss term, represents the spectrum consistency loss term, represents the sample adaptive loss term, Respectively represent the weight coefficients of the corresponding loss terms; Describes the total number of samples in the data sample set, Indicates OTC temperature data generated by each sample; For the The conditional feature vector corresponding to the samples; represents the discriminator function, Indicates that the given conditional feature is Time Evaluation The probability of being the true data; represents the Fourier transform operator, represents the Euclidean norm; Indicates The adaptive weight of a sample reflects the importance of the sample relative to other samples in the calibration process; and Respectively The true OTC temperature and the generated OTC temperature of the samples; It is The characteristic difference of a sample indicates the difference between the sample and the mean of the data sample set; Indicates The characteristics of the samples are different; It is The representativeness of the samples indicates how close the generated data is to the real data; It represents the sum of the products of the characteristic differences and representativeness of all samples; is the mean of all real OTC temperature data, representing the central tendency of the data sample set; It is The representativeness of the samples indicates how close the generated data is to the real data.
[0099] Through the embodiment of the present application, the constraints of conditional features are incorporated into the loss function, and the generator can generate temperature data that conforms to specific scenarios and environments, so that the model can adapt to different environments and historical states more flexibly, thereby achieving more personalized and accurate temperature calibration. It should be noted that there are often periodic or fluctuating characteristics in temperature data, and these characteristics are more obvious in the frequency domain. By introducing spectral consistency loss and calibrating data in the frequency domain, the model can better capture and retain these characteristics, avoid the deviation of generated data from real data in frequency domain characteristics, and improve the accuracy and credibility of generated temperature data in scenarios involving periodic temperature changes. In addition, by assigning different weights to different samples and highlighting the influence of key samples in the loss function, the model can dynamically adjust the weights according to the feature differences and representativeness of each sample, and can dynamically adjust the contribution of different samples to the final temperature calculation, so that samples with prominent features get more attention in model training, so that the generated real-time OTC temperature is closer to the real temperature value, and the accuracy of calibration using the generated temperature is further optimized.
[0100] Regarding the above step S110, on the other hand, in some examples of the embodiments of the present application, after obtaining the anti-surge valve opening and closing signal, when it is determined that the anti-surge valve opening and closing signal is an open state signal, the OTC temperature is calculated based on the temperature signal matrix collected by the second thermocouple array, and the second thermocouple array includes a plurality of second thermocouples distributed in the second area between the combustion chamber and the turbine. Figure 2 In the example, the second thermocouple array can be located in the second regional plane.
[0101] More specifically, when the engine is started to the FSNL stage, the anti-surge valve is in the open state. At this time, the thermocouple deployed in the second area plane can be called to sample the temperature signal, and the temperature signal is used to calculate the OTC temperature. Therefore, during the startup stage and under low-load conditions, the thermocouple in the second area plane is used to calculate the temperature signal, and the basic data of the OTC temperature is provided by monitoring the flue gas temperature at the combustion chamber outlet, so as to prevent the surge phenomenon during the startup and ensure the safe startup of the equipment under low-load conditions.
[0102] Through the embodiments of the present application, different operating stages of the gas turbine are detected based on the anti-surge valve opening and closing signals, and the OTC temperature is calculated accordingly using temperature sampling signals on different planes, thereby providing a strong guarantee for the stable operation of the gas turbine under different loads and operating conditions, and also promoting the intelligent and refined development of gas turbine control technology.
[0103] Figure 4 An operational flow chart of an example of collecting real-time exhaust temperature based on a first thermocouple array according to an embodiment of the present application is shown.
[0104] like Figure 4 As shown, in step S410, for the real-time temperature signal collected by each first thermocouple, the real-time temperature signal is Fourier transformed to obtain a corresponding temperature spectrum, and the corresponding spectrum peak frequency and spectrum energy are extracted from the temperature spectrum.
[0105] In some embodiments, the temperature signal of each thermocouple is preprocessed, including denoising and standardization, and a bandpass filter is used to remove noise in non-target frequency bands to ensure the purity of the data. Then, the time domain signal is converted into a frequency domain signal through Fourier transform, and the frequency characteristics of the signal are analyzed. The main frequency is determined by the maximum amplitude point of the spectrum, and the main frequency components of the signal are identified. The total energy of the spectrum in the selected frequency band is calculated to measure the overall energy level of the signal.
[0106] More specifically, the spectrum peak frequency and spectrum energy can be calculated as follows:
[0107] , Formula (19)
[0108] , formula (20)
[0109] , Formula (21)
[0110] In the formula, Indicates the first thermocouple At time step The real-time temperature signal collected, Indicates the first thermocouple Real-time temperature signal After Fourier transformation, the frequency The temperature spectrum under Indicates the first thermocouple The peak frequency of the temperature spectrum, express The amplitude of Indicates the first thermocouple The spectrum energy, and They are the lower and upper frequency limits of the spectrum analysis respectively.
[0111] Through this embodiment, the temperature characteristics of different wavelengths are extracted by Fourier transform and spectrum analysis, providing the information basis of frequency domain. Through multi-wavelength spectrum analysis, the main frequency components of the exhaust temperature signal can be accurately captured, reflecting the temperature change characteristics of different components, which helps to distinguish the temperature characteristics under different wavelengths.
[0112] In step S420, the temperature signal of each first thermocouple is subjected to wavelet transformation to be decomposed into multi-scale signal components, and information fusion is performed according to the wavelet coefficients of each scale to obtain a multi-scale fusion temperature to determine the corresponding real-time smoke exhaust temperature.
[0113] In some embodiments, the temperature signal of each thermocouple is subjected to wavelet transformation and decomposed into multi-scale signal components. The wavelet coefficients of each scale are integrated to perform multi-scale information fusion, and the information is weighted with the corresponding spectrum peak frequency and spectrum energy to obtain comprehensive temperature data. Thus, the different frequency components of the temperature signal can be decomposed by wavelet transformation, and the change characteristics of different scales can be captured. Based on the fusion of multi-scale features, the accuracy of signal measurement analysis can be improved.
[0114] It should be noted that when fusing signals from different thermocouples, the signal quality coefficient and wavelength importance coefficient can be combined to calculate the adaptive weight of each thermocouple to ensure that the data fusion is more accurate and reliable and to implement an adaptive weight mechanism. This can effectively reduce the interference of low-quality signals and improve the overall measurement accuracy when the signal quality is uneven.
[0115] More specifically, wavelet transform decomposition and adaptive multi-scale signal fusion are performed in the following ways:
[0116] , Formula (22)
[0117] , Formula (23)
[0118] , Formula (24)
[0119] In the formula, Indicates the first thermocouple At time step The wavelet transform temperature, Indicates that at time step Multi-scale fusion temperature; It is a scale The weight coefficient of K represents the total number of scale levels of wavelet transform, is in scale On Wavelet coefficients of the temperature signal of a thermocouple; and is the weight coefficient of spectrum energy and spectrum peak frequency; H is the total number of first thermocouples in the first thermocouple array; Indicates the first thermocouple Adaptive fusion weights of Indicates the first thermocouple The signal-to-noise ratio of the measured signal is Indicates the first thermocouple The contribution of the measurement wavelength, It represents the weighted sum of the signal-to-noise ratio index and contribution of all the first thermocouples.
[0120] Through this embodiment, the temperature signal is analyzed and fused at multiple scales through wavelet transform, and the temperature features at different scales can be extracted and combined to provide richer temperature information. Based on the multi-scale analysis, the fast and slow temperature change features can be captured, and the comprehensiveness of the data can be improved. Furthermore, according to the signal quality and the importance of the wavelength, the weight of each thermocouple is dynamically adjusted to ensure that high-quality data is given priority in the data fusion process, and the adaptive mechanism is used to support the provision of stable measurement results under different working conditions and environmental conditions.
[0121] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of actions combined, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.
[0122] Figure 5 A structural block diagram of an example of a gas turbine OTC temperature calculation system based on MBR thermocouples according to an embodiment of the present application is shown.
[0123] like Figure 5 As shown, the gas turbine OTC temperature calculation system 500 based on MBR thermocouple includes an anti-surge valve signal acquisition unit 510, an MBR temperature acquisition unit 520, a reference timing acquisition unit 530 and a real-time temperature output unit 540.
[0124] The anti-surge valve signal acquisition unit 510 is used to acquire the anti-surge valve opening and closing signal.
[0125] The MBR temperature acquisition unit 520 is used to collect real-time exhaust temperature based on the first thermocouple array when it is determined that the anti-surge valve opening and closing signal is a closed state signal; the first thermocouple array includes a plurality of first thermocouples distributed in an area adjacent to the exhaust port, the first thermocouples are band response thermocouples, and each of the first thermocouples has a unique corresponding measurement wavelength.
[0126] The reference timing acquisition unit 530 is used to acquire a historical OTC temperature sequence and environmental parameter timing data; the historical OTC temperature sequence includes the historical OTC temperature output at a continuous first time step, and the environmental parameter timing data includes a plurality of environmental sensing parameters collected for the first thermocouple array at a continuous second time step, and the environmental sensing parameters include ambient temperature, ambient pressure, ambient humidity and electromagnetic interference intensity.
[0127] The real-time temperature output unit 540 is used to input the real-time smoke exhaust temperature, the historical OTC temperature sequence and the environmental parameter time series data into the temperature calibration model to output the corresponding real-time OTC temperature; the temperature calibration model adopts a conditional generative adversarial network, and the conditional vector of the conditional generative adversarial network is determined based on the environmental parameter time series data and the historical OTC temperature sequence.
[0128] In some embodiments, an embodiment of the present application provides a non-volatile computer-readable storage medium, in which one or more programs including execution instructions are stored, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute the steps of any of the above-mentioned gas turbine OTC temperature calculation methods based on MBR thermocouples in the present application.
[0129] In some embodiments, the embodiments of the present application also provide a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes any step of the above-mentioned gas turbine OTC temperature calculation method based on MBR thermocouple.
[0130] In some embodiments, the embodiments of the present application also provide an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the steps of the gas turbine OTC temperature calculation method based on MBR thermocouples.
[0131] Figure 6 FIG. 1 is a schematic diagram of the hardware structure of an electronic device for executing a gas turbine OTC temperature calculation method based on an MBR thermocouple provided in another embodiment of the present application, such as Figure 6 As shown, the device includes:
[0132] One or more processors 610 and memory 620, Figure 6 A processor 610 is taken as an example.
[0133] The device for executing the gas turbine OTC temperature calculation method based on the MBR thermocouple may further include: an input device 630 and an output device 640 .
[0134] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.
[0135] The memory 620, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as the program instructions / modules corresponding to the gas turbine OTC temperature calculation method based on MBR thermocouples in the embodiment of the present application. The processor 610 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 620, that is, the gas turbine OTC temperature calculation method based on MBR thermocouples in the above method embodiment is realized.
[0136] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely arranged relative to the processor 610, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0137] The input device 630 can receive input digital or character information and generate signals related to user settings and function control of the electronic device. The output device 640 can include a display device such as a display screen.
[0138] The one or more modules are stored in the memory 620, and when executed by the one or more processors 610, the gas turbine OTC temperature calculation method based on MBR thermocouple in any of the above method embodiments is executed.
[0139] The above-mentioned product can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present application.
[0140] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:
[0141] (1) Mobile communication equipment: This type of equipment is characterized by having mobile communication functions and its main purpose is to provide voice and data communications. This type of terminal includes: smart phones, multimedia phones, functional phones, and low-end phones.
[0142] (2) Ultra-mobile personal computer devices: These devices belong to the category of personal computers, have computing and processing functions, and generally also have mobile Internet access features. These terminals include: PDA, MID and UMPC devices, etc.
[0143] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players, handheld game consoles, e-books, smart toys, and portable car navigation devices.
[0144] (4) Other onboard electronic devices with data interaction functions, such as on-board devices installed in vehicles.
[0145] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0146] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the relevant technology, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.
[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A gas turbine OTC temperature calculation method based on MBR thermocouple, It is characterized in that The method includes: Obtain the anti-surge valve opening and closing signal; In the case where it is determined that the anti-surge valve opening and closing signal is a closed state signal, real-time exhaust gas temperature is collected based on a first thermocouple array; the first thermocouple array includes a plurality of first thermocouples distributed in a first area adjacent to the exhaust port, the first thermocouples are band response thermocouples, and each of the first thermocouples has a unique corresponding measurement wavelength; Acquire a historical OTC temperature sequence and environmental parameter time series data; the historical OTC temperature sequence includes historical OTC temperatures output at continuous first time steps, the environmental parameter time series data includes a plurality of environmental sensing parameters collected for the first thermocouple array at continuous second time steps, the environmental sensing parameters include environmental temperature, environmental pressure, environmental humidity and electromagnetic interference intensity; The real-time smoke exhaust temperature, the historical OTC temperature sequence and the environmental parameter time series data are input into a temperature calibration model to output the corresponding real-time OTC temperature; the temperature calibration model adopts a conditional generative adversarial network, and the condition vector of the conditional generative adversarial network is determined according to the environmental parameter time series data and the historical OTC temperature sequence; The temperature calibration model comprises an input layer, a conditional feature embedding module, a timing analysis module, a convolution module and an output layer; The input layer is used to receive real-time exhaust temperature, historical OTC temperature series and environmental parameter time series data; The conditional feature embedding module is used to embed the historical OTC temperature series and environmental parameter time series data into a high-dimensional feature space to determine the corresponding initial condition feature vector; The timing analysis module is used to extract the time dependency in the initial condition feature vector to obtain a corresponding timing dependency condition vector; The convolution module is used to fuse and convolve the time series dependent condition vector and the real-time smoke exhaust temperature to obtain corresponding fused convolution features; The output layer is used to generate real-time OTC temperature based on the fused convolutional features.
2. The method according to claim 1, It is characterized in that The conditional feature embedding module is used to process the historical OTC temperature series and environmental parameter time series data through a multimodal attention mechanism to generate an initial condition feature vector: , , , In the formula, Indicates the corresponding The environmental parameter vector for each time step; Represents the historical OTC temperature series; represents the transpose of the trainable weight vector; For the corresponding The unnormalized score of the attention weight at time steps, denoted and relevance; Represents a trainable weight matrix, which is used to transform the The environmental parameter vector of each time step is embedded into the attention calculation space; represents a trainable weight matrix used to embed the historical OTC temperature series into the attention calculation space; represents the result of exponentially calculating and summing the unnormalized scores of all time steps, Indicates the total number of time steps corresponding to the environmental parameter time series data, Indicates The attention weights for time steps; represents the hyperbolic tangent activation function; represents the initial condition eigenvector; Represents an embedding operation to and Combine to generate feature representation.
3. The method according to claim 2, It is characterized in that The timing analysis module uses a bidirectional LSTM layer to process the initial condition feature vector in the forward and backward directions to obtain the corresponding timing dependency condition vector: , , , In the formula, Indicates that the forward LSTM The hidden state of time steps; represents long short-term memory network processing, represents the initial condition eigenvector Corresponding to The conditional eigenvector of time steps, Indicates that the forward LSTM The hidden state of time steps, Indicates that the reverse LSTM The hidden state of time steps, Indicates that the reverse LSTM The hidden state of time steps; Indicates The bidirectional feature representation of time steps combines the hidden states of the forward and backward LSTMs.
4. The method according to claim 3, It is characterized in that The convolution module uses a deep convolution layer to gradually process the time-dependent condition vector and the real-time smoke exhaust temperature through multiple convolution layers to obtain a deep fusion convolution feature: , , In the formula, represents the fused feature vector, represents the vector concatenation operation, Indicates real-time exhaust temperature data. Represents the temporal dependency conditional vector output by the bidirectional LSTM layer; It is The output of the convolutional layer, which represents the feature representation processed by the convolution operation and activation function; represents the LeakyReLU activation function, represents the convolution operation, and Respectively represent The weight matrix and bias term of the convolutional layer; The output layer is used to process the deep fusion convolutional features to output the corresponding real-time OTC temperature: , , In the formula, is the feature representation of the final output of the convolution module, n is the total number of convolutional layers in the convolutional module; represents the feature vector after pooling, represents the global average pooling operation; Indicates the real-time OTC temperature output. and Represent the weight matrix and bias term of the output layer respectively.
5. The method according to claim 4, It is characterized in that The loss function of the generator of the temperature calibration model is for: , , , , , , , In the formula, represents the base generator loss term, represents the spectrum consistency loss term, represents the sample adaptive loss term, Respectively represent the weight coefficients of the corresponding loss terms; Describes the total number of samples in the data sample set, Indicates OTC temperature data generated by each sample; For the The conditional feature vector corresponding to the samples; represents the discriminator function, Indicates that the given conditional feature is Time Evaluation The probability of being the true data; represents the Fourier transform operator, represents the Euclidean norm; Indicates The adaptive weight of a sample reflects the importance of the sample relative to other samples in the calibration process; and Respectively The true OTC temperature and the generated OTC temperature of the samples; It is The characteristic difference of a sample indicates the difference between the sample and the mean of the data sample set; Indicates The characteristics of the samples are different; It is The representativeness of the samples indicates how close the generated data is to the real data; It represents the sum of the products of the characteristic differences and representativeness of all samples; is the mean of all true OTC temperature data, representing the central tendency of the data sample set.
6. The method according to any one of claims 1 to 5, It is characterized in that The collecting of real-time exhaust temperature based on the first thermocouple array includes: For the real-time temperature signal collected by each first thermocouple, the real-time temperature signal is subjected to Fourier transform processing to obtain a corresponding temperature spectrum, and the corresponding spectrum peak frequency and spectrum energy are extracted from the temperature spectrum: , , , In the formula, Indicates the first thermocouple At time step The real-time temperature signal collected, Indicates the first thermocouple Real-time temperature signal After Fourier transformation, the frequency The temperature spectrum under Indicates the first thermocouple The peak frequency of the temperature spectrum, express The amplitude of Indicates the first thermocouple The spectrum energy, and are the lower and upper frequency limits of spectrum analysis respectively; The temperature signal of each first thermocouple is subjected to wavelet transform to be decomposed into multi-scale signal components, and information fusion is performed according to the wavelet coefficients of each scale to obtain the multi-scale fusion temperature to determine the corresponding real-time smoke exhaust temperature: , , , In the formula, Indicates the first thermocouple At time step The wavelet transform temperature, Indicates that at time step Multi-scale fusion temperature; It is a scale The weight coefficient of K represents the total number of scale levels of wavelet transform, is in scale On Wavelet coefficients of the temperature signal of a thermocouple; and is the weight coefficient of spectrum energy and spectrum peak frequency; H is the total number of first thermocouples in the first thermocouple array; Indicates the first thermocouple Adaptive fusion weights of Indicates the first thermocouple The signal-to-noise ratio of the measured signal is Indicates the first thermocouple The contribution of the measurement wavelength, It represents the weighted sum of the signal-to-noise ratio index and contribution of all the first thermocouples.
7. The method according to any one of claims 1 to 5, It is characterized in that After obtaining the anti-surge valve opening and closing signal, the method further includes: When it is determined that the anti-surge valve opening and closing signal is an open state signal, the OTC temperature is calculated based on the temperature signal matrix collected by the second thermocouple array; the second thermocouple array includes a plurality of second thermocouples distributed in a second area between the combustion chamber and the turbine.
8. A gas turbine OTC temperature calculation system based on MBR thermocouple, It is characterized in that The system includes: An anti-surge valve signal acquisition unit, used to acquire an anti-surge valve opening and closing signal; The MBR temperature acquisition unit is used to acquire the real-time exhaust gas temperature based on the first thermocouple array when it is determined that the anti-surge valve opening and closing signal is a closed state signal; the first thermocouple array includes a plurality of first thermocouples distributed in an area adjacent to the exhaust port, the first thermocouples are band response thermocouples, and each of the first thermocouples has a unique corresponding measurement wavelength; A reference timing acquisition unit is used to acquire a historical OTC temperature sequence and environmental parameter timing data; the historical OTC temperature sequence includes the historical OTC temperature output at a continuous first time step, and the environmental parameter timing data includes a plurality of environmental sensing parameters collected for the first thermocouple array at a continuous second time step, and the environmental sensing parameters include environmental temperature, environmental pressure, environmental humidity and electromagnetic interference intensity; A real-time temperature output unit, used for inputting the real-time exhaust temperature, the historical OTC temperature sequence and the environmental parameter time series data into a temperature calibration model to output the corresponding real-time OTC temperature; the temperature calibration model adopts a conditional generative adversarial network, and the condition vector of the conditional generative adversarial network is determined according to the environmental parameter time series data and the historical OTC temperature sequence; The temperature calibration model comprises an input layer, a conditional feature embedding module, a timing analysis module, a convolution module and an output layer; The input layer is used to receive real-time exhaust temperature, historical OTC temperature series and environmental parameter time series data; The conditional feature embedding module is used to embed the historical OTC temperature series and environmental parameter time series data into a high-dimensional feature space to determine the corresponding initial condition feature vector; The timing analysis module is used to extract the time dependency in the initial condition feature vector to obtain a corresponding timing dependency condition vector; The convolution module is used to fuse and convolve the time series dependent condition vector and the real-time smoke exhaust temperature to obtain corresponding fused convolution features; The output layer is used to generate real-time OTC temperature based on the fused convolutional features.
