DEH servo control system and method

Abstract

The embodiment of the invention provides a DEH servo control system and method, and belongs to the technical field of servo control. The system comprises an integrated control module used for collecting and processing the position of a steam turbine valve, a servo valve control signal and a rotating speed signal, executing a control strategy and outputting a servo instruction; the feedback sensing module is used for collecting actual displacement signals of the valve and micro-vibration signals in the operation process; the fusion identification module comprises a frequency domain feature extraction module and a vibration classification discrimination model; the fuzzy control module is used for generating a servo control adjusting signal according to the valve deviation and the deviation change rate; and the reconstruction switching module is connected with the fusion identification module and is used for switching to a redundant sensing path or starting a prediction control mechanism when the feedback signal is identified to be abnormal. According to the scheme, the dynamic switching of the control paths can be realized, and the continuous operation of the system is guaranteed, so that the fault identification capability of the DEH system and the robustness in the operation process are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of servo control, and particularly to a DEH servo control system and a DEH servo control method. Background Art

[0002] The steam turbine digital electro-hydraulic control system (Digital Electro-Hydraulic Control System, abbreviated as DEH system) is a key thermal automation control core device in thermal power plants, mainly responsible for the start-stop control, speed regulation and load matching of steam turbines. Its operation stability and response accuracy are directly related to the safety and reliability of the whole machine. The DEH system usually consists of modules such as a control logic module, a servo execution loop, a speed detection module and a feedback device. Among them, the accuracy of the servo control loop and its position feedback signal is the key link to ensure the stable operation of the valve.

[0003] In the existing DEH system, a linear variable differential transformer (LVDT) is generally used as the position feedback device, and the servo valve is controlled by a servo card to achieve the opening control of the steam turbine valve. However, in the actual operation process, the LVDT feedback signal is subject to factors such as mechanical vibration, electrical interference and its own life, and is prone to problems such as signal drift, frequent fluctuations or short-term failure, resulting in valve misoperation or even system protection triggering. In severe cases, it can cause the unit to trip unexpectedly. At the same time, the existing control strategies mostly rely on the PID loop with fixed parameters, lack the adaptive ability to sudden signal abnormalities, and the identification of LVDT faults is mostly a lagging process, lacking feedforward identification and prediction ability.

[0004] Therefore, the core problems faced by the existing technology are: (1) lack of a timely identification and processing mechanism for the fault trend of LVDT, and the system is not sensitive to sensor abnormalities; (2) the control strategy is rigid and cannot be adaptively adjusted according to the operating state, and the overall robustness of the system to disturbances and abnormalities is insufficient. In view of the above problems, there is an urgent need to propose a DEH servo control system with dynamic identification ability and control strategy adaptability to improve its operation stability and safety. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a DEH servo control system and method to at least solve the problems in the prior art that the linear variable differential transformer lacks a fault warning mechanism during operation, resulting in the failure of the feedback signal not being recognized in time, and the traditional control strategy is insufficient in adapting to the dynamic disturbance environment.

[0006] To achieve the above object, a first aspect of the present invention provides a DEH servo control system, which includes: an integrated control module for collecting and processing turbine valve position, servo valve control signal and speed signal, executing a control strategy and outputting a servo command; a feedback sensing module including a linear variable differential transformer and an acceleration sensor for collecting the actual valve displacement signal and the micro-vibration signal during operation; a fusion recognition module including a frequency-domain feature extraction module and a vibration classification discrimination model, where the frequency-domain feature extraction module performs a fast Fourier transform on the micro-vibration signal, and the vibration classification discrimination model is used to identify potential abnormalities of the linear variable differential transformer; a fuzzy control module including a fuzzy rule base and an inference engine, where the fuzzy control module generates a servo control adjustment signal according to the valve deviation and the deviation change rate; a reconstruction switching module connected to the fusion recognition module for switching to a redundant sensing path or enabling a predictive control mechanism when an abnormality of the feedback signal is identified.

[0007] Optionally, the fuzzy rule base in the fuzzy control module includes: multiple sets of fuzzy control rules set for different load states of the steam turbine, each set of rules taking the valve deviation and the deviation change rate as input variables and the servo valve adjustment coefficient as the output quantity, and the inference engine uses the maximum membership degree method for fuzzy inference to generate a continuous control output signal.

[0008] Optionally, the vibration classification discrimination model in the fusion recognition module includes a multi-dimensional feature space constructed based on the K-nearest neighbor algorithm; the multi-dimensional feature space takes the harmonic amplitude, peak frequency, and energy density extracted by the fast Fourier transform as input parameters, calculates the Euclidean distance by comparing with historical fault samples, and realizes the classification and recognition of the abnormal trend of the linear variable differential transformer; the fusion recognition module further includes an anomaly detection sub-module; the anomaly detection sub-module constructs an anomaly judgment model with a multi-tree structure using the isolation forest algorithm, and is used to perform an isolation score based on the frequency-domain features of the acceleration signal provided by the feedback sensing module, and outputs an anomaly trigger signal to the reconstruction switching module when the score exceeds a set threshold.

[0009] Optionally, the acceleration sensor in the feedback sensing module is fixed to the outer shell of the linear variable differential transformer through an adhesive structure for real-time collection of the micro-vibration signal of the valve drive mechanism; the acceleration sensor is connected to the analog signal acquisition channel of the integrated control module, and the analog signal acquisition channel has an anti-electromagnetic interference function.

[0010] Optionally, the integrated control module includes: a main control logic processing module, a signal acquisition module, and a communication interface module. The signal acquisition module is respectively connected to the feedback sensing module and the fuzzy control module. The communication interface module exchanges data with the DCS system through the industrial Ethernet protocol or the Modbus protocol to implement the reporting of servo control signals and the issuance of control instructions.

[0011] Optionally, the reconstruction switching module includes a status monitoring sub-module and a control logic switching module. The status monitoring sub-module receives the abnormal trigger signal from the fusion recognition module and is used to judge the validity of the feedback signal in real time. When the control logic switching module determines that the signal is abnormal, it switches to a pre-set redundant sensing path or adopts a model-based predictive control strategy. The predictive control strategy is based on the dynamic response model established between the servo actuator and the linear variable differential transformer, and adaptively updates the model parameters by the least squares method. When the feedback signal is lost or delayed, it predicts and compensates the valve displacement state and outputs a control reference value to the integrated control module.

[0012] Optionally, after receiving the signal from the feedback sensing module, the fuzzy control module adjusts the enabling range of the fuzzy rule base according to the rotational speed change trend. The adjustment strategy for the enabling range of the fuzzy rule base is: calculating a control dynamic compensation factor based on the correlation between the rotational speed derivative and the valve opening deviation, and performing real-time gain adjustment on the fuzzy rule result.

[0013] Optionally, the linear variable differential transformer of the feedback sensing module is connected to the redundant linear variable differential transformer through a high-speed sampling switching module. The switching period of the high-speed sampling switching module is less than a preset switching period threshold, and is used to perform sampling synchronization, signal equalization, and drift correction on each feedback signal.

[0014] In a second aspect of the present invention, a DEH servo control method is provided. The method is implemented based on the above DEH servo control system, and the method includes: collecting real-time displacement signals and vibration signals during operation of the steam turbine valve based on a feedback sensing module; wherein, the displacement signal is obtained by a linear variable differential transformer, and the vibration signal is obtained by an acceleration sensor installed on the valve drive structure; performing frequency domain transformation on the vibration signal based on a fusion recognition module, extracting spectral characteristic parameters, and inputting the extracted spectral characteristic parameters into a vibration classification and discrimination module constructed based on a machine learning model to identify whether there is an abnormal trend in the operating state of the linear variable differential transformer; receiving the real-time displacement signal and the target valve opening set value based on a fuzzy control module, calculating the valve deviation and the deviation change rate, and generating a servo control adjustment instruction based on a fuzzy rule base and an inference engine; judging whether it is necessary to switch the control channel based on a reconstruction switching module according to the recognition result of whether there is an abnormal trend in the operating state of the linear variable differential transformer. If an abnormal trend is recognized, redundant feedback path switching or a predictive control strategy is executed. Receiving the servo control adjustment instruction and the output of the reconstructed control path based on an integrated control module, comprehensively judging and then generating a final control instruction to be output to a servo actuator to drive the steam turbine valve to actuate.

[0015] On the other hand, the present invention provides a computer-readable storage medium, on which instructions are stored, and when the instructions are run on a computer, the computer is caused to execute the above DEH servo control method.

[0016] Through the above technical solutions, the solution of the present invention realizes the centralized collection of key operating parameters of the steam turbine and the output of control instructions through an integrated control module, and combines the displacement and vibration information obtained by the feedback sensing module to improve the system sensing accuracy. The fusion recognition module can perform frequency domain feature extraction and intelligent discrimination on the tiny abnormalities in the LVDT feedback to realize early fault trend recognition. The fuzzy control module performs adaptive control adjustment based on the dynamic change of the deviation to improve the flexibility of the control strategy. After detecting feedback abnormalities, the reconstruction switching module can realize the dynamic switching of the control path to ensure the continuous operation of the system, thereby enhancing the fault recognition ability and robustness of the DEH system during operation.

[0017] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent specific embodiment part. Description of the Drawings

[0018] The drawings are used to provide a further understanding of the embodiments of the present invention, and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation to the embodiments of the present invention. In the drawings: Figure 1It is the system structure diagram of the DEH servo control system provided by an embodiment of the present invention; Figure 2 It is the step flow chart of the DEH servo control method provided by an embodiment of the present invention. Specific embodiments

[0019] The following will describe the specific embodiments of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining and illustrating the present invention, and are not used to limit the present invention.

[0020] Figure 1 It is the system structure diagram of the DEH servo control system provided by an embodiment of the present invention. As Figure 1 shown, the embodiment of the present invention provides a DEH servo control system, and the system includes: an integrated control module, configured to collect and process the steam turbine valve position, servo valve control signal and speed signal, execute a control strategy and output a servo command; a feedback sensing module, including a linear variable differential transformer and an acceleration sensor, configured to collect the actual valve displacement signal and the micro-vibration signal during operation; a fusion recognition module, including a frequency domain feature extraction module and a vibration classification discrimination model, the frequency domain feature extraction module performs a fast Fourier transform on the micro-vibration signal, and the vibration classification discrimination model is used to identify potential abnormalities of the linear variable differential transformer; a fuzzy control module, including a fuzzy rule base and an inference engine, the fuzzy control module generates a servo control adjustment signal according to the valve deviation and the deviation change rate; a reconstruction switching module, connected to the fusion recognition module, configured to switch to a redundant sensing path or enable a predictive control mechanism when an abnormal feedback signal is identified.

[0021] Preferably, the fuzzy rule base in the fuzzy control module includes: multiple groups of fuzzy control rules set for different load states of the steam turbine, each group of rules uses the valve deviation and the deviation change rate as input variables, and the servo valve adjustment coefficient as the output quantity, and the inference engine uses the maximum membership degree method for fuzzy inference to generate a continuous control output signal.

[0022] In the embodiment of the present invention, the fuzzy rule base in the fuzzy control module includes: multiple sets of fuzzy control rules are preset for the valve control characteristics of the steam turbine under different operating states such as start-stop, low load, medium load, and full load. Each set of control rules uses the deviation value between the actual valve opening and the target set opening, and the derivative of the deviation value with respect to time (i.e., the deviation change rate) as input variables, and the control adjustment coefficient of the servo valve as the output variable. Specifically, the deviation value can be divided into membership levels such as large deviation, medium deviation, small deviation, and zero deviation, and the deviation change rate can be divided into multiple dynamic levels such as fast increase, medium increase, slow increase, zero change, slow decrease, medium decrease, and fast decrease. The output variable is divided into membership levels such as strong adjustment, medium adjustment, weak adjustment, and zero adjustment according to the strength of the adjusted voltage or current signal amplitude.

[0023] The inference engine performs fuzzy inference processing based on the membership degrees of the input variables and the "IF-THEN" type control rules defined in the fuzzy rule base. The inference engine adopts the Max Membership Principle, that is, it compares the output results of multiple fuzzy rules that meet the conditions and selects the output result with the highest membership degree as the basis for the current control instruction. Subsequently, this fuzzy output result is converted into a continuous value through a defuzzification module (such as using the centroid method) to generate a continuous control signal that can be directly received by the servo control system, and finally transmitted to the servo drive module to control the action amplitude and rate of the main steam valve or governing valve of the steam turbine. This fuzzy control module can not only automatically select an appropriate control rule group according to the current operating state, but also dynamically adjust the control output according to the deviation change trend. Compared with the defects of the traditional PID controller, such as high sensitivity to parameter changes and prominent response delay problems, this fuzzy control module has stronger self-adaptability and flexible control capabilities.

[0024] Based on the solution of the present invention, this fuzzy control module realizes the refined adjustment ability of the servo control signal by introducing multiple sets of fuzzy control rules matching the load state, combining the maximum membership degree inference mechanism and the continuous output strategy, thereby effectively improving the valve response stability of the steam turbine under dynamic conditions and the overall adjustment accuracy of the servo control system, and significantly enhancing the control robustness of the DEH system in a complex operating environment.

[0025] Preferably, the vibration classification and discrimination model in the fusion recognition module includes a multi-dimensional feature space constructed based on the K-nearest neighbor algorithm; the multi-dimensional feature space uses the harmonic amplitude, peak frequency, and energy density extracted by the fast Fourier transform as input parameters, and calculates the Euclidean distance by comparing with historical fault samples to achieve the classification and recognition of the abnormal trend of the linear variable differential transformer; the fusion recognition module further includes an anomaly detection sub-module; the anomaly detection sub-module uses the isolation forest algorithm to construct an anomaly judgment model with a multi-tree structure, which is used to perform an isolation score based on the frequency domain characteristics of the acceleration signal provided by the feedback perception module, and when the score exceeds the set threshold, an anomaly trigger signal is output to the reconstruction switching module.

[0026] In the embodiment of the present invention, the fusion recognition module includes a vibration classification and discrimination model for fault trend recognition and an anomaly detection sub-module for abnormal mutation detection. Among them, the vibration classification and discrimination model uses a multi-dimensional feature space constructed based on the K-Nearest Neighbor (KNN) algorithm. This multi-dimensional feature space performs a fast Fourier transform (FFT) on the micro-vibration signal collected by the acceleration sensor in the feedback perception module to extract frequency domain characteristic parameters including harmonic amplitude, main frequency peak position, signal energy density, etc. These parameters together constitute the input dimension of the model. In the training stage of the model, multiple groups of characteristic samples in the normal state and fault state of the known LVDT are introduced to form a labeled historical sample library. During operation, the Euclidean distance is calculated between the real-time collected characteristic parameters and various samples in the sample library, and based on this, the belonging category of the current sample in the feature space is judged, so as to realize the early recognition and classification of the tiny abnormal trend of the linear variable differential transformer (LVDT).

[0027] To enhance the system's ability to recognize sudden anomalies, the fusion recognition module further includes an anomaly detection sub-module. This sub-module uses the Isolation Forest algorithm to construct an unsupervised anomaly detection model composed of multiple randomly divided trees, which is used to analyze the isolation path length of newly collected data samples. Through multiple rounds of random feature selection and recursive splitting operations, an anomaly score is given to the input frequency domain characteristics. When the average path length of the characteristic samples at a certain moment in multiple trees is much lower than the normal range of historical samples, it means that the sample has a high degree of "isolation", that is, it is judged as an abnormal point. Once the anomaly score exceeds the set threshold, the anomaly detection sub-module will automatically send an anomaly trigger signal to the reconstruction switching module, indicating that there may be a risk of serious feedback distortion in the main system or the LVDT is about to fail.

[0028] Based on the solution of the present invention, by collaboratively constructing and deploying the K-nearest neighbor classification model and the isolation forest anomaly detection model, on the one hand, it can achieve the early classification and recognition of the LVDT fault trend, improving the system's ability to identify minute anomalies; on the other hand, it can quickly locate and respond to sudden vibration anomalies, forming a complete recognition mechanism for early warning and in-process intervention, thereby enhancing the fault tolerance and operational stability of the entire DEH servo control system under signal fault conditions.

[0029] Preferably, the acceleration sensor in the feedback sensing module is fixed to the outer shell of the linear variable differential transformer through an adhesive structure for real-time acquisition of the micro-vibration signals of the valve drive mechanism; the acceleration sensor is connected to the analog signal acquisition channel of the integrated control module, and the analog signal acquisition channel has an anti-electromagnetic interference function.

[0030] In the embodiment of the present invention, the acceleration sensor in the feedback sensing module is fixedly installed on the outer side of the outer shell of the linear variable differential transformer through an industrial-grade adhesive structure. This adhesive structure uses a two-component high-viscosity insulating material, which has high temperature resistance, anti-vibration, and long-term attachment stability, and can adapt to the working conditions requirements such as high temperature, high humidity, and high electromagnetic interference in the steam turbine on-site operation environment. The acceleration sensor is arranged close to the connection area between the LVDT and the valve actuator to achieve high-sensitivity acquisition of dynamic responses such as micro-vibrations, transient impacts, and periodic disturbances during the valve execution process. The sensor has multi-axis measurement capabilities and has a milligram-level acceleration resolution and a high-frequency bandwidth response range. The output end of the acceleration sensor is connected to the analog signal acquisition channel in the integrated control module through an anti-interference signal cable. The acquisition channel is internally provided with a hardware-level filtering module, including an analog level amplification, band-pass filtering, and common-mode suppression circuit, ensuring good anti-electromagnetic interference ability during the signal acquisition process in the industrial environment and enabling stable transmission and distortion-free processing of the signal.

[0031] Preferably, the integrated control module includes: a main control logic processing module, a signal acquisition module, and a communication interface module. The signal acquisition module is respectively connected to the feedback sensing module and the fuzzy control module. The communication interface module exchanges data with the DCS system through an industrial Ethernet protocol or a Modbus protocol to achieve the reporting of servo control signals and the issuance of control instructions.

[0032] In the embodiment of the present invention, the integrated control module serves as the core control platform of the DEH servo control system and includes three functional sub-modules: a main control logic processing module, a signal acquisition module, and a communication interface module. The main control logic processing module integrates a high-performance embedded processor or an industrial-grade PLC, and is built-in with a control strategy execution module, a data cache module, and an instruction scheduling module, which is used to receive real-time operation data from each module, execute the fuzzy control algorithm and the abnormal judgment logic, and generate corresponding servo valve control instructions.

[0033] The signal acquisition module is configured with multiple high-precision analog signal input interfaces, which are respectively connected to the feedback sensing module and the fuzzy control module, and is used to collect the displacement signal output by the linear variable differential transformer, the micro-vibration signal output by the acceleration sensor, and the control adjustment coefficient calculated by the fuzzy control module. The acquisition module internally integrates isolation amplification, anti-interference filtering, and high-speed analog-to-digital conversion circuits to ensure the high-fidelity input and processing stability of the signals. The communication interface module is configured with dual-redundant industrial communication ports and supports multiple industrial protocol standards such as Modbus-TCP, Modbus-RTU, and industrial Ethernet (such as Profinet or EtherNet / IP). This communication interface conducts two-way data exchange with the DCS system: transmits the servo control status, fault identification results, and feedback signal analysis data upward, and receives system control commands such as target valve set values and control strategy switching instructions downward, ensuring real-time coordination between the DEH system and the plant-level control system.

[0034] Preferably, the reconstruction switching module includes a status monitoring sub-module and a control logic switching module; the status monitoring sub-module receives the abnormal trigger signal from the fusion recognition module and is used to judge the validity of the feedback signal in real time. When the control logic switching module determines that the signal is abnormal, it switches to a pre-set redundant sensing path or adopts a model-based predictive control strategy; the predictive control strategy is based on the dynamic response model established between the servo actuator and the linear variable differential transformer, adaptively updates the model parameters through the least squares method, predicts and compensates the valve displacement state when the feedback signal is lost or delayed, and outputs a control reference value to the integrated control module.

[0035] In the embodiment of the present invention, the reconstruction switching module is used to maintain the continuous control ability of the DEH servo control system in the case of abnormal or invalid feedback signals, and it includes two parts: a status monitoring sub-module and a control logic switching module. The status monitoring sub-module maintains a communication connection with the fusion recognition module, receives the abnormal trigger signal output by it in real time, and comprehensively judges the validity of the current feedback signal by combining parameters such as the signal stability of the linear variable differential transformer, the vibration characteristic judgment result, and the data sampling integrity. Once it is determined that a certain feedback signal has an abnormal trend, distortion, or disconnection, the result will be transmitted to the control logic switching module.

[0036] The control logic switching module has a dual control channel selection mechanism: when the system is configured with redundant linear variable differential transformer signal paths, it preferentially performs the automatic switching of the main redundant signal channel; if there is no available redundant signal or the redundant signal is also abnormal, it automatically activates the predictive control strategy. The predictive control strategy is based on the historical control response data between the servo actuator and the linear variable differential transformer to establish a dynamic response model. This model uses the least squares method as the parameter identification means to adaptively update the model coefficients in real time, so that the model can predict the actual displacement state of the valve at the current moment in the case of feedback signal loss, delay or severe fluctuation. The predicted value is output as a control reference signal to the integrated control module for generating subsequent control instructions.

[0037] Preferably, after receiving the signal from the feedback sensing module, the fuzzy control module adjusts the enabled range of the fuzzy rule base according to the change trend of the rotational speed; the adjustment strategy for the enabled range of the fuzzy rule base is: calculating the control dynamic compensation factor based on the correlation between the rotational speed derivative and the valve opening deviation, and performing real-time gain adjustment on the fuzzy rule result.

[0038] In the embodiment of the present invention, after receiving the real-time signal from the feedback sensing module, the fuzzy control module not only uses the valve opening deviation and its change rate as the basic input variables, but also further introduces the current rotational speed of the steam turbine and its derivative (i.e., acceleration) as the basis for dynamic adjustment. A fuzzy rule base enabling adjustment mechanism is set in the fuzzy control module, which can automatically adjust the range of activated control rules according to the current operating state to achieve the dynamic coupling of the operating state and the control strategy.

[0039] Specifically, the fuzzy control module continuously tracks the change trend of the steam turbine rotational speed, extracts the first derivative of the rotational speed, and is used to judge the dynamic working condition state of the current system, such as steady-state operation, acceleration, deceleration or sudden load switching. The system performs online fitting and calculation on the correlation between the rotational speed derivative and the valve opening deviation to obtain a dynamic compensation factor representing the control correction amplitude. This compensation factor is used as the adjustment parameter for the enabled weight of the fuzzy rule base, which determines whether a certain type of rule (such as high-response, high-output rules) should be temporarily strengthened or suppressed, so as to perform refined real-time adjustment on the gain of the fuzzy inference result. For example, in the case of a rapid decrease in rotational speed at high speed, the compensation factor may increase the weight of the "quick response" type fuzzy rules to ensure that the control system has sufficient adjustment sensitivity; during steady-state operation, the weight of the high-response rules is automatically reduced to avoid overshoot or causing system oscillation.

[0040] Preferably, the linear variable differential transformer of the feedback sensing module is connected to the redundant linear variable differential transformer through a high-speed sampling and switching module; the switching period of the high-speed sampling and switching module is less than a preset switching period threshold, and is used for sampling synchronization, signal equalization and drift correction of each feedback signal.

[0041] In an embodiment of the present invention, a high-speed sampling and switching module is provided between the primary linear variable differential transformer and the redundant linear variable differential transformer in the feedback sensing module, and is used for realizing parallel monitoring and fast switching of multiple feedback channels during the operation of the control system. The high-speed sampling and switching module includes a signal sampling module, a data caching module, a switching control module and a signal processing module, and has the data channel processing ability of high precision and high speed. The sampling and switching module adopts a dual-channel input structure, can simultaneously receive the analog displacement feedback signals of the primary and backup LVDTs, and performs parallel sampling at a period of microseconds or sub-milliseconds. The actual switching period is less than the switching period threshold set by the system, such as 1 ms or less, to ensure that the influence of the switching action on the response of the control system can be ignored.

[0042] In order to avoid data deviation caused by factors such as sensor installation difference, temperature drift, cable length or signal mismatch between different channels, the high-speed sampling and switching module is internally provided with a signal synchronization processing mechanism to perform amplitude calibration, timing alignment and dynamic equalization adjustment on the primary and backup signals. The signal processing module further includes a drift correction logic, and adopts algorithms such as moving average and baseline correction to filter out the slow-varying errors in the signal, and improve the consistency and switchability of the feedback signal. When an abnormality in the primary LVDT signal is detected, the system can issue a channel switching instruction based on the real-time evaluation result by the switching control module, and quickly switch from the primary channel to the redundant channel without affecting the data continuity of the fuzzy control module and the integrated control module.

[0043] Figure 2 It is a flowchart of the DEH servo control method provided by an embodiment of the present invention. As Figure 2 shown, an embodiment of the present invention provides a DEH servo control method, and the method includes: Step S10: Collect the real-time displacement signal of the steam turbine valve and the vibration signal during the operation process based on the feedback sensing module; wherein, the displacement signal is obtained by a linear variable differential transformer, and the vibration signal is obtained by an acceleration sensor installed on the valve drive structure.

[0044] Specifically, the linear variable differential transformer (LVDT) perceives the change in valve opening in real time through mechanical connection with the servo actuator, and converts the mechanical displacement into a voltage signal for output. This voltage signal is sent to the integrated control module through the analog acquisition interface to reflect the current valve displacement state. At the same time, the acceleration sensor is fixedly installed on the outer shell of the valve drive mechanism by an adhesive structure. The sensor supports multi-axis measurement and can collect the minute mechanical vibrations generated during the valve operation, including components such as periodic resonance, transient shock, and random disturbance. The acceleration signal is synchronously sent to the control system after being processed by the anti-interference filtering circuit for subsequent fault identification and dynamic characteristic analysis. By establishing a multi-source perception channel, high-precision and multi-dimensional real-time perception of the valve state is achieved, providing raw data support for subsequent vibration anomaly identification and control strategy optimization.

[0045] Step S20: Based on the fusion recognition module, perform frequency-domain transformation on the vibration signal, extract spectral feature parameters, and input the extracted spectral feature parameters into the vibration classification and discrimination module constructed based on the machine learning model to identify whether there is an abnormal trend in the operating state of the linear variable differential transformer.

[0046] Specifically, the vibration signal first enters the frequency-domain processing module in the fusion recognition module. The system uses the fast Fourier transform (FFT) algorithm to decompose the spectrum of the original acceleration signal and extract multiple frequency-domain characteristic quantities, including harmonic amplitude, peak frequency, frequency distribution density, energy concentration, etc. These characteristic parameters are then normalized and form an input vector, which is sent into the vibration classification and discrimination model. This model is constructed based on the K-nearest neighbor algorithm. A large number of normal and fault samples are introduced in the training stage, and sample classification is performed by constructing an Euclidean distance feature space. During operation, the real-time feature vector is compared with the sample library. If the current signal feature is close to the historical fault sample in terms of spatial distance, the model outputs a result indicating "abnormal trend exists". This result will be used as a trigger condition for the control switching logic and input into the next module to achieve feedforward recognition of the impending abnormality of the LVDT, intervene in the control strategy in advance, and prevent misoperation of the valve.

[0047] Step S30: The fuzzy control module receives the real-time displacement signal and the target valve opening setting value, calculates the valve deviation and the deviation change rate, and generates a servo control adjustment instruction based on the fuzzy rule base and the inference engine.

[0048] Specifically, the fuzzy control module is internally equipped with a deviation calculation module and a rule inference engine, which receives the valve displacement signal and the target opening value set by the control system in real time, and calculates the deviation Δx and its change rate Δx' between the two. According to the set fuzzyfication level, the system divides the deviation input variable into multiple membership functions, such as "large deviation", "medium deviation", "small deviation", etc., and divides the deviation change rate into dynamic levels such as "rapid increase", "slow decrease", "zero change", etc. The fuzzy rule base is constructed based on a large number of working condition experiences and contains multiple fuzzy control rules in the form of "IF-THEN", such as "IF the deviation is medium and the change rate is slow increase THEN the output is medium adjustment". The inference engine uses the maximum membership degree method or the centroid method to perform fusion calculation on the rules to obtain the fuzzy control output result. This control quantity is then defuzzified and converted into a continuous analog control instruction to adjust the response intensity and speed of the servo actuator, thereby realizing precise and flexible dynamic control of the valve.

[0049] Step S40: Based on the reconstruction switching module, it is judged whether it is necessary to switch the control channel according to the recognition result of whether there is an abnormal trend in the operating state of the linear variable differential transformer. If it is recognized that there is an abnormal trend, the redundant feedback path switching or the predictive control strategy is executed.

[0050] Specifically, the reconstruction switching module includes an abnormal state evaluation sub-module and a control path switching logic module, which receives the fault trend recognition result from the fusion recognition module in real time. When the result indicates that there is an abnormal trend or fluctuation distortion in the LVDT feedback signal, the abnormal state evaluation sub-module will trigger the channel switching logic. First, the system judges whether there is a redundant LVDT feedback channel. If so, the high-speed channel switching module will be preferentially executed to switch the control feedback source to the redundant sensor, and the switching period is controlled within milliseconds to ensure the continuity of the feedback signal. If there is no available redundant path, or the redundant signal is also unavailable, the system will call the predictive control strategy. This strategy is based on the historical control response relationship between the servo mechanism and the LVDT, adaptively establishes a dynamic model using the least squares method, predicts and estimates the current valve displacement state, and uses it as a temporary control reference input to maintain the stable operation of the system until the signal channel returns to normal.

[0051] Step S50: Based on the integrated control module, it receives the servo control adjustment instruction and the output of the reconstructed control path, and generates a final control instruction after comprehensive judgment and outputs it to the servo actuator to drive the steam turbine valve to act.

[0052] Specifically, the integrated control module, as the execution core of control decision-making, integrates the main logic processor, signal acquisition module, and instruction output interface, and is responsible for fusing and judging the servo adjustment instructions generated by the fuzzy control module and the prediction or redundant channel feedback signals output by the reconstruction switching module. Under normal working conditions, the fuzzy control output is preferentially adopted; under fault conditions, the input weights are dynamically adjusted according to the control path switching state to generate the final servo control signal. The control instruction is sent to the servo drive module through the industrial bus or analog output interface to accurately control the opening and response rate of the servo valve, so as to stably drive the main steam valve or governing valve of the steam turbine, thereby achieving the purpose of accurately adjusting the unit load and responding to the dispatching command. This process is executed in real-time closed-loop and can upload the execution status, feedback curve, and control data to the superior DCS system for online monitoring and analysis.

[0053] The embodiment of the present invention also provides a computer-readable storage medium, on which instructions are stored, and when running on a computer, the computer is made to execute the above DEH servo control method.

[0054] Those skilled in the art can understand that all or part of the steps in the methods of the above embodiments can be completed by instructing relevant hardware through a program. The program is stored in a storage medium and includes several instructions to make a single-chip microcomputer, chip, or processor execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drive, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk, or optical disc, etc., which can store program codes.

[0055] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the technical concept scope of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all belong to the protection scope of the embodiments of the present invention. In addition, it should be noted that, among the various specific technical features described in the above specific embodiments, they can be combined in any suitable manner without conflict. To avoid unnecessary repetition, the embodiments of the present invention will not separately describe various possible combination methods.

[0056] In addition, any combination can be made between various different embodiments of the present invention, as long as it does not violate the idea of the embodiments of the present invention, and it should also be regarded as the content disclosed in the embodiments of the present invention.

Claims

1. A DEH servo control system, characterized in that: The system comprises: Integrated control module, used to collect and process turbine valve position, servo valve control signal and speed signal, execute control strategy and output servo command; Feedback sensing module, including a linear variable differential transformer and an acceleration sensor, used to collect the actual displacement signal of the valve and the micro-vibration signal during operation; A fusion identification module, including a frequency domain feature extraction module and a vibration classification and discrimination model, wherein the frequency domain feature extraction module performs a fast Fourier transform on the micro-vibration signal, and the vibration classification and discrimination model is used to identify potential abnormalities of the linear variable differential transformer; A fuzzy control module, comprising a fuzzy rule base and an inference engine, wherein the fuzzy control module generates a servo control adjustment signal according to a valve deviation and a deviation change rate; The reconstruction switching module is connected to the fusion identification module and is used to switch to a redundant sensing path or enable a predictive control mechanism when an abnormal feedback signal is identified.

2. The DEH servo control system according to claim 1, characterized in that: The fuzzy rule base in the fuzzy control module includes: Multiple groups of fuzzy control rules are set for different load states of the steam turbine. Each group of rules takes the valve deviation and the deviation change rate as input variables and the servo valve adjustment coefficient as output. The reasoning engine uses the maximum membership method to perform fuzzy reasoning to generate a continuous control output signal.

3. The DEH servo control system according to claim 1, characterized in that: The vibration classification and discrimination model in the fusion recognition module includes a multi-dimensional feature space constructed based on a K-nearest neighbor algorithm; The multidimensional feature space uses the harmonic amplitude, peak frequency and energy density extracted by the fast Fourier transform as input parameters, and calculates the Euclidean distance by comparing with historical fault samples to achieve classification and identification of abnormal trends of linear variable differential transformers; The fusion recognition module also includes an anomaly detection submodule; The anomaly detection submodule uses the isolation forest algorithm to construct an anomaly judgment model with a multi-tree structure, which is used to perform isolation scoring based on the frequency domain characteristics of the acceleration signal provided by the feedback perception module, and output an anomaly trigger signal to the reconstruction switching module when the score exceeds a set threshold.

4. The DEH servo control system according to claim 1, characterized in that: The acceleration sensor in the feedback sensing module is fixed to the housing of the linear variable differential transformer through an adhesive structure, and is used to collect the micro-vibration signal of the valve drive mechanism in real time; The acceleration sensor is connected to the analog signal acquisition channel of the integrated control module, and the analog signal acquisition channel has an anti-electromagnetic interference function.

5. The DEH servo control system according to claim 1, characterized in that: The integrated control module comprises: The main control logic processing module, the signal acquisition module and the communication interface module, the signal acquisition module is connected to the feedback perception module and the fuzzy control module respectively, and the communication interface module exchanges data with the DCS system through the industrial Ethernet protocol or the Modbus protocol to realize the reporting of servo control signals and the issuance of control instructions.

6. The DEH servo control system according to claim 1, characterized in that: The reconstruction switching module includes a state monitoring submodule and a control logic switching module; The state monitoring submodule receives the abnormal trigger signal of the fusion identification module to judge the validity of the feedback signal in real time. When the control logic switching module judges that the signal is abnormal, it switches to a preset redundant sensing path or adopts a model-based predictive control strategy; The predictive control strategy is based on a dynamic response model established between the servo actuator and the linear variable differential transformer, and adaptively updates the model parameters through the least squares method. When the feedback signal is lost or delayed, the valve displacement state is predicted and compensated, and a control reference value is output to the integrated control module.

7. The DEH servo control system according to claim 1, wherein after receiving the signal from the feedback sensing module, the fuzzy control module adjusts the activation range of the fuzzy rule base according to the rotation speed change trend; The adjustment strategy for the activation scope of the fuzzy rule base is: The control dynamic compensation factor is calculated based on the correlation between the speed derivative and the valve opening deviation, and the fuzzy rule result is adjusted in real time.

8. The DEH servo control system according to claim 1, wherein the linear variable differential transformer of the feedback sensing module is connected to the redundant linear variable differential transformer via a high-speed sampling switching module; The switching period of the high-speed sampling switching module is less than a preset switching period threshold, and is used to perform sampling synchronization, signal equalization and drift correction on each feedback signal.

9. A DEH servo control method, characterized in that: The method is implemented based on the DEH servo control system according to any one of claims 1 to 8, and the method comprises: Based on the feedback sensing module, the real-time displacement signal of the turbine valve and the vibration signal during operation are collected; The displacement signal is obtained by a linear variable differential transformer, and the vibration signal is obtained by an acceleration sensor installed on the valve drive structure; Based on the fusion recognition module, the vibration signal is transformed in the frequency domain to extract the spectrum feature parameters, and the extracted spectrum feature parameters are input into the vibration classification and discrimination module constructed based on the machine learning model to identify whether there is an abnormal trend in the operation state of the linear variable differential transformer; The fuzzy control module receives the real-time displacement signal and the target valve opening setting value, calculates the valve deviation and the deviation change rate, and generates a servo control adjustment instruction based on the fuzzy rule base and the inference engine; Based on the reconstruction switching module, judging whether it is necessary to switch the control channel according to the identification result of whether there is an abnormal trend in the operation state of the linear variable differential transformer, if the abnormal trend is identified, executing the redundant feedback path switching or enabling the predictive control strategy; Based on the integrated control module receiving the servo control adjustment instruction and the output of the reconstructed control path, a final control instruction is generated after comprehensive judgment and output to the servo actuator to drive the turbine valve to move.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the DEH servo control method of claim 9.

Images