A smart control system for a steam turbine anti-shaft seizure heating device

Abstract

This invention discloses an intelligent control system for a turbine anti-shaft seizure heating device, applicable to various operating conditions such as shutdown insulation, turning gear, cold preheating, hot restart, and trip recovery. It implements coordinated heating for multiple zones including bearing housing / bearing seat / bearing bush, lubricating oil, and optional shaft seals. The system includes multi-source data acquisition and reliable verification, generating representative temperatures, weights, data quality, change rates, and temperature difference derivatives. It features a state machine and risk assessment output modes and hotspots; constraint and setpoint scheduling outputs target temperatures, ramps, temperature difference windows, and soft power limits; when the temperature difference exceeds the window, it raises the cold side / presses the hot side; the total power budget, under the upper limit constraint, ensures a minimum for key zones and allocates power according to weight / hotspot / efficiency, feeding back restricted information to the closed loop to combat saturation / degradation; the execution module completes power regulation and implements hard / soft interlocks and diagnostics, with interlock / diagnostic write-back triggering upstream recalculation; it outputs start-up permission and cause codes and records them for transmission to the DCS.

Description

Technical Field

[0001] This invention relates to the field of turbine auxiliary equipment control and thermal management, and in particular to a hierarchical closed-loop intelligent control system for anti-seize heating devices during turbine shutdown, turning gear operation, startup and fault recovery stages. It belongs to the cross-technology of industrial process control, thermal control and equipment health management. Background Technology

[0002] During operation such as shutdown cooling, prolonged turning gear operation, cold start, and hot restart after tripping, there are significant differences in thermal inertia among the rotor, bearing housing, bearing bushes, sealing cavity, and lubrication system of a steam turbine. If the temperature distribution in key parts is uneven or the rate of temperature rise / fall is uncontrolled, it can easily lead to: abnormal bearing clearance, difficulty in establishing an oil film, and increased turning gear resistance; increased tendency of rotor thermal bending, resulting in rubbing and excessive vibration; and in severe cases, the risk of "shaft seizure / jamming" (usually manifested as difficulty in starting, unstable turning gear operation, and abnormal temperature rise of bearing bushes).

[0003] Therefore, anti-seize bearing heating devices are often configured on site, such as electric heating of bearing housing / bearing seat, electric heating of lubricating oil tank / oil circuit, electric heating of sealed steam supply, etc., and are equipped with temperature control cabinets to realize heating control and protection. The process is as follows: after collecting the temperature signal, it directly enters the single-loop temperature control (such as PID) or simple switch control, and is loosely coupled with the start / stop / power limiting strategy of the heating circuit.

[0004] Current turbine anti-shaft seizure related electric heating / heat tracing systems are mostly based on "single-point temperature control, single-loop PID, and simple start / stop / alarm," treating the "collected temperature value" as inherently reliable. Furthermore, when there is an anomaly at a certain measuring point, a sudden noise change, or inconsistency across multiple measuring points, these values ​​may still directly participate in the closed-loop calculation, leading to: Single objective: Focusing on a specific temperature point, ignoring the constraints of temperature difference / gradient and heating ramp (rate of change) between bearing housing / bearing seat / bearing bush and oil system; Non-closed-loop linkage of operating conditions: The switching of operating conditions such as shutdown insulation, turning gear, cold preheating, hot restart, and trip recovery lacks stable and smooth control and de-jitter maintenance, resulting in control strategy jitter or inconsistency; Uncontrollable scenarios with limited total power: When the total power of plant power or heating is limited, the lack of "critical area backup + coordinated distribution" can easily lead to local overheating, widening of temperature difference, or temperature loss in critical areas. Lack of risk and diagnostic drive: The lack of an interpretable bearing seizure risk index and hot spot location, and the lack of online diagnosis and self-calibration for heating circuit degradation, insulation aging, sensor drift, etc., often leads to erroneous control, false alarms or maintenance blind spots. Interlocking lacks a write-back and recalculation mechanism: Many interlocking systems only disconnect, without having a complete closed-loop recalculation mechanism for writing back the interlocking state to the upstream, triggering constraint convergence and redistribution. The control chain is not closed, making it difficult to guarantee that "the upstream will never generate an output that breaks through the hard boundary".

[0005] Therefore, there is an urgent need for an intelligent control system for the turbine anti-shaft seizure heating device to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent control system for a turbine anti-seize heating device. This system addresses various operating conditions, including shutdown insulation, turning gear operation, cold preheating, hot restart, and fault / trip recovery. It implements coordinated control over the heating zones related to the bearing housing / bearing seat / bearing bush, the lubricating oil heating zone, and optional zones. Through a combination of strategies including "temperature difference window management + temperature rise / fall ramp constraints + total power budget allocation," an adaptive and traceable closed-loop control system is formed to reduce the risk of temperature difference expansion and thermal shock, improve start-up reliability, and ensure that the thermal boundary conditions of key zones are prioritized in interlocking or fault situations.

[0007] To achieve the above objectives, this invention does not directly use the raw acquired data for control. Instead, it uses the raw data frames output by the multi-source acquisition module only as input data containers. Specifically, the raw data frames contain acquisition time information, raw temperature data sets, raw electrical parameter data sets, and raw status data sets. The key improvement lies in setting up a reliable measurement and consistency verification module to convert the raw data frames into reliable data frames, serving as the unique data standard across the entire link. The reliable data frame includes at least: Representative temperature data group: obtained by fusing multiple measurement points in the same partition, or representative temperatures after removing / reducing the weight of abnormal measurement points; Weighted data sets available for measurement points: quantify the availability of data from each measurement point / zone for subsequent control; Data quality indicators: identifying open circuits, short circuits, out-of-bounds errors, mutations, inconsistencies, etc. Temperature change rate data set: used for slope constraints and risk assessment; Temperature difference derived quantity information: used for temperature difference window calculation and management strategy input.

[0008] Therefore, the RawFrame is no longer the entry point for "direct drive control," but rather completes data quality gating, representative quantity unification, weighted availability expression, and temperature difference / rate of change derivation through the "trusted frame," forming a unified standard that runs through the state machine, risk, constraint package, power budget, and execution interlocking. This step is the key "connecting button" in the control chain of this invention, significantly different from the "acquisition as control" of ordinary systems.

[0009] Based on trusted frames, this invention further implements a collaborative mechanism of "closed-loop fallback + upstream recalculation" by executing power regulation / interlocking / diagnosis modules: While outputting the actual heating power, the execution module generates interlock status information / interlock cause coding information and health diagnosis result information; When the interlock is triggered or an anomaly is diagnosed, the above information is written back to the temperature difference constraint setpoint generation and constraint scheduling module and the total power budget and collaborative allocation module, triggering upstream recalculation of constraint convergence and power reallocation; The upstream recalculated constraint and setpoint package, power budget results are sent out again for execution, so that the system can maintain stable control under hard boundary conditions, with critical partition protection, priority given to temperature difference management, and power-limited feedback to resist saturation.

[0010] Therefore, this invention is not simply about "adding an interlock" or "adding an alarm," but rather about using the interlock / diagnosis results as effective feedback in the control chain to form a closed-loop reconfiguration mechanism of "execution → write-back → recalculation → re-execution," ensuring that upstream will not continuously generate outputs that break through hard boundaries, thereby achieving system-level collaborative innovation from data to strategy to security safeguards.

[0011] Specifically, an intelligent control system for a turbine anti-shaft seizure heating device includes at least the following modules and forms a closed loop through information exchange: Multi-source acquisition module: Acquires temperature, status and electrical parameters, and outputs raw data frames; the raw data frames include at least acquisition time information, temperature raw data sets, electrical parameter raw data sets and status raw data sets. Trusted Measurement and Consistency Verification Module: Performs validity judgment and consistency verification on the original data frame and outputs a trusted data frame; the trusted data frame includes at least a group of representative temperature data, a group of available weight data for measurement points, data quality flag information, a group of temperature change rate data, and temperature difference derived quantity information. Operating condition state machine and sequential control module: Based on trusted data frames, identify operating conditions and perform segmented switching and / or gating control, output operating condition identification information, which includes at least operating condition mode identifier, mode duration information and event encoding information, and includes debouncing and hold logic; Bearing seizure risk index calculation module: Based on temperature difference, temperature change rate, oil temperature / oil film conditions and operating condition information, calculate and output risk assessment information, which includes at least the total risk index or risk level, sub-risk composition information and hot spot indication information. Temperature difference constraint setpoint generation and constraint scheduling module: Generates and outputs constraint and setpoint package based on working condition identification information, risk assessment information and available weight data group of measuring points. The constraint and setpoint package includes at least the target temperature setpoint information of each heating zone, heating ramp constraint parameter information and / or cooling ramp constraint parameter information, temperature difference window limit information and soft power upper limit information of each heating zone. Total power budget and collaborative allocation module: Under the constraint of total power upper limit information, the module budgets and allocates power, outputs power control command information and priority information for each heating zone, and outputs power limitation feedback information. Zoned closed-loop temperature control module: Based on the target temperature setpoint information and ramp constraint parameter information, it generates power demand information for each zone, and performs anti-integral saturation and / or degradation control according to the power-limited feedback information; Execution and power regulation / interlocking / diagnosis module: converts power control command information into actual heating output and outputs execution feedback information; the execution feedback information includes at least actual output power information, interlocking status information and / or interlocking cause coding information, and health diagnosis result information; Start-up permission and recording communication module: Outputs start-up permission information based on comprehensive risk assessment information, temperature difference window satisfaction status, temperature change rate constraint satisfaction status, and interlock status information. The start-up permission information includes at least: start-up permission / prohibition indication, risk level information, and cause code information. It is also used to perform trend recording and event auditing of operating condition identification information, constraint and set value packages, power control command information, execution feedback information, and alarm events, and send them to the DCS / supervisor system. The online thermal inertia identification module uses actual output power information as a reference to identify the thermal inertia time constant parameters, equivalent thermal efficiency parameters, and thermal response hysteresis parameters of each zone online and outputs model confidence information and preset model confidence threshold. Its output is constrained and generated with setpoints and / or the total power budget is allocated for adaptive adjustment.

[0012] Based on the above, when an interlocking trigger or health diagnosis abnormality occurs in the execution and power regulation / interlocking / diagnosis module, the interlocking status information and / or health diagnosis result information will be fed back to the temperature difference constraint setpoint generation and constraint scheduling module and the total power budget and collaborative allocation module to trigger upstream recalculation of constraint convergence and power reallocation, ensuring that the control output does not exceed the hard boundary.

[0013] Preferably, the system operates in a layered cycle: a fast loop centered on the execution and power regulation / interlocking / diagnosis module; a control loop consisting of multi-source acquisition, reliable verification, total power budget, and execution modules; a strategy loop consisting of the operating state machine, risk assessment, and constraint scheduling; a slow identification loop consisting of online thermal inertial identification; and a background recording loop consisting of start-up permission and recording communication. Through the above layered cycle, an integrated closed loop is achieved that prioritizes safety, ensures stable control, facilitates gradual strategy changes, and ensures traceable records.

[0014] Preferably, the reliable measurement module identifies open circuits, short circuits, out-of-bounds conditions, abrupt changes, and consistency anomalies to form data quality indicators, and generates representative temperatures and usable weights for measurement points based on these indicators. Subsequent control prioritizes the use of representative temperatures and weights.

[0015] Preferably, the operating state machine provides the hold time and outputs the mode duration information. The operating state machine and the sequential control module are used to stably maintain the segmented switching / gating process of the heating zone, avoiding control jumps caused by mode jitter.

[0016] Preferably, the model confidence information output by the online thermal inertia identification module corresponds to the preset model confidence threshold. When the model confidence is lower than the preset threshold, the object characteristic parameters (thermal inertia time constant parameter, equivalent thermal efficiency parameter, and thermal response hysteresis parameter) of the previous time step are not updated or are frozen to avoid abnormal excitation or abnormal data causing model drift.

[0017] Preferably, the risk assessment includes at least sub-risks such as temperature difference, rate of change, oil temperature / oil film and operating conditions, and outputs hot spot zones and dominant factors for strategy scheduling.

[0018] Preferably, when the temperature difference constraint setpoint generation and constraint scheduling module detects that the temperature difference exceeds the temperature difference window limit, it executes a temperature difference management priority strategy. This strategy involves increasing the target temperature setpoint information and / or the weight corresponding to the available weight data group of the measuring point in the relatively cold side zone, and limiting the temperature rise slope constraint parameter information and / or soft power upper limit information in the relatively hot side zone, in order to suppress the expansion of the temperature difference.

[0019] Preferably, the total power budget and collaborative allocation module performs a minimum allocation to key partitions under the constraint of the total power upper limit information, and performs weighted allocation based on the available weighted data group of measurement points, hot spot indication information and / or equivalent thermal efficiency parameters, and can perform peak shaving or time-sharing polling.

[0020] Preferably, the power-limited feedback information includes at least the limited flag information and / or the allocation reduction ratio information. The zoned closed-loop temperature control module performs anti-integral saturation and / or degradation control based on the power-limited feedback information to avoid temperature overshoot after the power limitation is lifted.

[0021] Preferably, the execution and power regulation / interlocking / diagnostic module includes hard interlocking and soft interlocking, where the trigger threshold of hard interlocking is higher than that of soft constraint and cut-off or limiting is performed first; The startup permission and recording communication module outputs the startup permission information based on a configurable threshold system. The configurable threshold system includes at least one or more of the following: upper temperature limit threshold and / or lower temperature limit threshold, temperature difference window limit threshold, temperature change rate threshold, available weight threshold for measurement points, model confidence threshold, upper total power limit threshold, and de-jitter hold time threshold. It also performs trend recording, event recording, and audit recording of operating condition identification information, risk assessment information, constraint and setpoint package, power control command information, interlock status information, and health diagnosis result information.

[0022] Preferably, an interlock write-back triggered recalculation state machine control mechanism is established between the execution and power regulation / interlock / diagnosis module, the constraint / setpoint scheduling module, the power budget and collaborative allocation module, and the start-up permission and recording communication module. The state machine includes at least a normal control state, a local degradation state, a power degradation state, a global protection state, and a recovery observation state. When the execution and power regulation / interlock / diagnosis module detects zone over-temperature, power anomaly, or electrical fault, it generates a feedback data packet containing interlock level, interlock type, diagnostic results, and actual power information and writes it back to the supervision and collaboration layer, triggering the constraint / setpoint scheduling module to adjust the corresponding zone according to the feedback data packet. The temperature setpoint, temperature difference window, temperature rise / fall ramp, and soft power limit of the zone or the whole are recalculated. At the same time, the power budget and collaborative allocation module is triggered to recalculate the power command and priority of each zone under the total power constraint. If necessary, the state is switched from normal control state to local degradation state, power degradation state, or global protection state. The start permission and record communication module outputs the corresponding start permission or prohibition signal based on the updated risk level. When the interlock is released and no new fault occurs after the recovery observation time, the state machine switches the system state from the recovery observation state back to the normal control state, thereby realizing the closed-loop write-back and constraint recalculation of the interlock result to the upstream control strategy.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows: Through a systematic mechanism of "unified trusted frame caliber + interlocking / diagnostic write-back and recalculation", this invention achieves at least the following: Stable control can be maintained even under conditions of inconsistent measurement points, bad points, and sudden noise changes, avoiding the dragging effect of bad points; During multi-condition switching, the strategy operates continuously based on a trusted frame aperture to reduce jitter and misadjustment. Under conditions of limited total power and fault / interlock triggering, the thermal boundary conditions of key zones are prioritized to be met by writing back and recalculating, thereby reducing the risk of temperature difference expansion and bearing seizure. The entire chain is traceable (data frame - trusted frame - mode - risk - constraint package - budget - execution - write-back - license / record), which facilitates operation, maintenance and auditing. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] This invention aims to provide an intelligent control system for multiple operating conditions and multiple zones, employing a layered closed-loop architecture of "supervision and coordination layer + zone execution layer," achieving the following while satisfying the hard interlocking safety boundary: Synergistic anti-seize shaft heating with controlled temperature difference, controlled slope, and controlled power; "Data quality gating" based on reliable measurement prevents bad pixels from dragging down control. Adaptive strategies based on risk index and hotspot interpretation (tightening temperature difference window, reducing ramp, switching modes, and granting startup permission). Critical area backup and coordinated allocation based on total power budget (including peak shaving, time-sharing / polling). Self-calibration and maintainable closed loop based on online identification and consistency diagnosis; Interlock / diagnostic results can be written back to upstream triggers for recalculation, forming a traceable and complete control system.

[0026] This invention's system is designed for multiple operating conditions (shutdown insulation, turning gear, cold preheating, hot restart, fault / trip recovery, etc.) and controls multiple zones of objects. Heating zones related to bearing housing / bearing seat / bearing bush; Lubricating oil heating zones; Optional features: zoned heating for shaft seal / sealing / journal, etc.

[0027] The system of this invention communicates with DCS, outputs startup permission / prohibition, and has trend / event / audit records.

[0028] The collaborative closed-loop link is as follows: data acquisition → reliability verification → operating condition identification → risk assessment → constraint and setpoint scheduling → zoned closed-loop temperature control → total power budget and collaborative allocation → execution / interlocking / diagnosis → temperature rise and electrical parameter feedback → re-acquisition; Furthermore: risk and diagnosis dynamically adjust strategy parameters; interlocking provides full protection and can be written back to upstream to trigger recalculation, forming a complete closed-loop system.

[0029] Example 1 This invention discloses an intelligent control system for a turbine anti-shaft seizure heating device, specifically comprising several modules including a multi-source acquisition module (RawFrame), a trusted measurement and consistency verification module (TrustedFrame), a state machine and sequential control module (ModeToken), an online thermal inertia identification module (ThermalModel), a shaft seizure risk index module (RiskVector), a temperature difference constraint setpoint generation and constraint scheduling module (ConstraintPack, temperature difference priority), a total power budget and collaborative allocation module (PowerCommand), an execution and power regulation / interlocking / diagnosis module (highest priority), and a start-up permission and record communication module (StartPermit and traceability), etc., more specifically as follows.

[0030] A multi-source acquisition module is used to collect the temperature, status, and electrical parameters of the turbine anti-seize shaft heating device and encapsulate them into a raw data frame. It includes a signal access unit and an acquisition and encapsulation unit, and outputs a RawFrame(ts,T_raw[],E_raw[],S_raw[]), where ts is a timestamp; T_raw is the raw temperature array, E_raw is the electrical parameter array, and S_raw is the status array. RawFrame works in conjunction with subsequent layers: it enters TrustedFrame to perform a trust verification, ensuring that all subsequent control is based on trusted data.

[0031] This module is used to perform validity assessment and consistency verification on raw data frames and generate reliable representative values. It includes a validity assessment unit, a representative value fusion unit, and a weight generation unit. Output objects: TrustedFrame(ts,T_hat[],w[],flag_T[],dTdt[],DeltaT_set[]), where T_hat represents the temperature, w is the weight, flag_T is the quality flag, and dTdt is the rate of change; DeltaT_set is a temperature-related derived / reference value that works in conjunction with subsequent modules: Subsequent modules (ModeToken, ThermalModel, RiskVector, ConstraintPack, PowerCommand, execution and power regulation / interlocking / diagnosis modules) preferentially use T_hat and w. When flag_T is abnormal, it triggers degradation / reweighting and provides a basis for interlocking.

[0032] The working condition state machine and sequential control module are used to identify working conditions such as shutdown insulation, turning gear, cold preheating, hot restart and fault / trip recovery and to perform segmented switching / gating. It includes a working condition identification unit, a debouncing and holding unit and a sequential control gating unit. The output object is ModeToken(mode,mode_age,event_code), which works with the subsequent stage: ModeToken drives the constraint and setpoint scheduling of ConstraintPack. Drives the gating, switching sequence, and protection strategy switching of the PowerCommand / Execution and Power Regulation / Interlocking / Diagnostics modules; event_code is entered into the logging system for auditing and tracing.

[0033] The online thermal inertia identification module, used for online identification of zoned thermal inertia and efficiency based on the actual power returned by the execution, includes a parameter identification unit and a confidence evaluation unit, using the actual power P_act returned by the execution and power regulation / interlock / diagnosis module as the reference excitation output: ThermalModel{tau[z],eta[z],delay[z],model_conf[z]}, where tau is the thermal inertia time constant, eta is the equivalent efficiency, delay is the hysteresis, and model_conf is the confidence level; When the confidence level is insufficient (below mc_min, etc.), the model can be frozen and used in conjunction with subsequent stages: ThermalModel enters ConstraintPack / PowerCommand for adaptive tightening / relaxing of constraints, power allocation according to efficiency, peak shaving and time-shaping, etc.

[0034] The binding risk index calculation module is used to quantify and decompose binding risk and provide hotspot explanations to drive strategy scheduling and startup authorization. It includes an index calculation unit and an index fusion unit. Output objects: RiskVector{RISK,RISK_part,hotspot(zone,factor)}, where RISK is the total risk, RISK_part is the sub-risk decomposition, and hotspot gives the hotspot partition and the main factors; The "hub-holding risk" is quantified and explained as an output, serving as the input for control strategy and the basis for startup permission. It works in conjunction with the following stages: RiskVector drives ConstraintPack to tighten W / decrease k; drives ModeToken to trigger mode switching; drives StartPermit to grant startup permission; and hotspot enters PowerCommand for priority scheduling.

[0035] The temperature difference constraint setting value generation and constraint scheduling module is used to generate temperature difference windows, slope constraints and target setting values ​​for each zone under current working conditions and risk-driven conditions, and to perform adaptive constraint scheduling. It includes a constraint generation unit, a setting value scheduling unit and a smooth transition unit. Based on the model, risk, object characteristics, and weights, a "setpoint and constraint package" is generated, ensuring that temperature difference management takes precedence over slope and conventional temperature control. Output object: ConstraintPack{SP_T[z],ramp_up[z],ramp_dn[z],DeltaT_limit[],P_soft_limit[z]}, where SP_T is the target temperature of the partition, ramp_up / ramp_dn are the ramps for heating and cooling, DeltaT_limit is the temperature difference window / upper limit, and P_soft_limit is the soft power limit of the partition; Constraint Adaptation: Automatically tighten / loosen by pressing ModeToken+RiskVector+ThermalModel+w; Smooth transition: S-shaped / smooth transition is used to avoid the impact of mode switching; When the temperature difference exceeds the window, the "cold side is raised / hot side is suppressed" approach is adopted. Instead of directly outputting power, the temperature difference is managed by adjusting SP_T and ramp / soft limit to guide the closed loop and distribution of the subsequent stage. In coordination with the subsequent stage: ConstraintPack enters PowerCommand to perform budgeting and allocation; enters the execution and power regulation / interlocking / diagnostic module to perform limiting and protection boundaries; and accepts write-back triggers from the execution and power regulation / interlocking / diagnostic module to "constraint convergence and recalculation".

[0036] This module, responsible for ensuring critical partition safety and coordinating power allocation based on weights, hotspots, and efficiency under total power cap constraints, and outputting power commands downstream, includes a power cap management unit, a safety allocation unit, and a weighted allocation unit. Key inputs include: ConstraintPack (containing SP_T, ramp, and soft limit), Risk hotspot, efficiency eta, weight w, and total power cap P_total_max. Output objects include: PowerCommand{P_cmd / duty_cmd / angle_cmd,priority[z]}, provides power commands, duty cycles, or trigger angle commands based on the actuator type, and outputs priority; Under the P_total_max constraint, the critical area is first guaranteed; then, it is allocated according to w / hotspot / efficiency eta, and the "constrained information" is fed back to the downstream closed loop to resist integral saturation / degradation, prevent control instability and overshoot, and cooperate with the downstream: PowerCommand enters the execution and power regulation / interlock / diagnostic module to execute, and accepts the interlock / diagnostic write-back trigger for reallocation from the execution and power regulation / interlock / diagnostic module.

[0037] The partition closed-loop temperature control module is used to convert power commands into expected power demands for each partition and maintain stability when power is limited. It includes a setpoint ramp shaping unit, a partition temperature control calculation unit, and an anti-integral saturation / degradation unit to combine SP_T and ramp parameters to form the power demand for each partition and coordinate with the budget allocation results.

[0038] This module is used to translate power commands into actual heating output, implement hardware / soft interlocking and health diagnostics, and write the interlocking and diagnostic results back upstream to trigger recalculation. It includes a power execution unit, an interlocking protection unit, and a diagnostic unit. Output objects: P_act[z] (actual power / actual output), ILK_state / ILK_code (interlocking state / interlocking code), DiagResult[z] (diagnostic result) and write back to the constraint scheduling module and power budget module; The command is translated into actual power output, and it also undertakes hardware / software interlocking and health diagnosis. It is the core gate of system safety and closed-loop recalculation. When executing the command, SCR / SSR / grading power is executed, and P_cmd is translated into P_out, forming actual power feedback. Interlocking and diagnostic content: Interlocking status and type information ILK_state: No interlock / Soft interlock / Hard interlock / Releasing...; ILK_code: What type of interlock? Over-temperature interlock (T > T_hi in a certain area); Leakage / insulation fault; Open circuit / short circuit; Heater overcurrent / wire breakage; Forced shutdown, etc.; Diagnosis and Health Information DiagResult[z]: Partition health status: heater_degrade (heater degradation); sensor_drift (sensor drift); insulation aging wiring_fault (wiring fault); normal; Add simple quantitative indicators: such as estimating the percentage decrease in efficiency, the degree of mismatch between temperature rise and power, etc. Write-back and recalculation mechanism: Interlocking / diagnosis results are written back to ConstraintPack / PowerCommand to trigger constraint convergence and redistribution, ensuring that upstream outputs do not exceed hard boundaries; When a soft / hard interlock occurs with the power regulation module or a partition fault is diagnosed, the corresponding circuit is not only disconnected or limited locally, but also the interlock status ILK_state, interlock type ILK_code, actual power P_act, power limitation flag and diagnosis result DiagResult are combined into a feedback data packet and written back to the supervision and coordination layer. After the feedback data packet is received by the constraint scheduling module and the power budget module, it triggers the recalculation of the set values, temperature difference windows, temperature rise and fall ramps, and total power allocation for each partition. This causes the interlocked or abnormal partitions to be automatically removed or downgraded, while the remaining healthy partitions regain power quotas according to their criticality and risk hotspots under the total power constraint. At the same time, the start-up permission module recalculates the start-up permission signal and reason code based on the updated risk index and interlocking reasons, realizing the closed-loop write-back of the interlocking results to the upstream strategy, and preventing the upstream from continuing to generate control outputs that exceed the hard boundary.

[0039] This startup permission and logging communication module is used to output startup permission / prohibition, complete trend, event and audit logs, and send them to DCS. It includes a permission determination unit, a reason coding unit and a communication logging unit. The output objects are StartPermit, RiskLevel and ReasonCode, and it realizes full-chain traceability recording. It works with the subsequent level: StartPermit serves as a key startup condition; ReasonCode provides an explainable reason to support operation decision-making and operation and maintenance closed loop.

[0040] This intelligent control system adopts a hierarchical closed-loop architecture of "supervision and coordination layer + zone execution layer". The supervision and coordination layer includes at least the working condition state machine and sequential control module, the bearing risk index calculation module, the temperature difference constraint setpoint generation and constraint scheduling module, and the total power budget and collaborative allocation module. The zone execution layer includes at least the zone closed-loop temperature control module and the execution and power regulation / interlocking / diagnosis module. It forms a collaborative closed-loop link: data acquisition → reliable verification → operating condition identification → risk assessment → constraint and setpoint scheduling → zoned closed-loop temperature control → total power budget and collaborative allocation → execution / interlocking / diagnosis → temperature rise and electrical parameter feedback → re-acquisition, thereby achieving collaborative anti-seize shaft heating with controlled temperature difference, controlled ramp and controlled power under multiple operating conditions.

[0041] In this system, the reliable measurement and consistency verification module performs open circuit / short circuit / out-of-bounds / mutation / consistency judgment on temperature measurement points and generates a data quality flag_T and a weight w. The temperature T_hat, the rate of change dTdt, and the weight w are used as the priority inputs for subsequent control. When the flag_T indicates an abnormality, the system reduces the weight of the corresponding measurement point or removes it, and provides the abnormal status to the interlocking and degradation strategy to avoid bad points dragging down the control.

[0042] In this system, the operating condition state machine and sequential control module include a de-jitter hold time t_hold and a mode age mode_age, which are used to suppress jitter and maintain stability during operating condition switching. In at least some operating conditions such as cold preheating, hot restart and accident recovery, segmented switching and gating are performed. The segmented switching includes at least the order of lubricating oil zone priority, critical bearing zone priority, and balanced other zones, so as to reduce the risk of thermal shock and temperature difference expansion.

[0043] In this system, the thermal inertia online identification module uses the actual power P_act returned by the execution and power regulation / interlocking / diagnosis module as the reference excitation, identifies the time constant tau, efficiency eta, and hysteresis delay for each partition, and outputs the model confidence level model_conf. When model_conf is lower than the preset confidence threshold mc_min, the model parameters are frozen and the previous reliable model is maintained to avoid model drift caused by low excitation or abnormal data, which would affect constraint scheduling and power allocation.

[0044] In this system, the bearing seizure risk index calculation module decomposes the bearing seizure risk into at least four sub-risks: temperature difference R1, rate of change R2, oil film / oil temperature R3, and operating condition R4, forming the total risk RISK and risk level RiskLevel. At the same time, it outputs hotspot(zone,factor) to indicate the zone that most needs to be addressed and its dominant factor. RiskLevel and hotspot serve as strategy inputs for constraint scheduling and power allocation, driving the tightening of the temperature difference window, the reduction of the slope upper limit, the adjustment of zone weights, and mode switching, thereby achieving risk-driven adaptive control and interpretable governance.

[0045] In this system, the temperature difference constraint setpoint generation and constraint scheduling module adaptively generates and schedules the constraint package ConstraintPack based on ModeToken, RiskVector, ThermalModel and weight w, and adopts an S-shaped or smooth transition method to achieve gradual switching between setpoints and constraints. Among them, "temperature difference management takes precedence over ramp constraints and conventional temperature control". When the temperature difference is detected to exceed DeltaT_limit, the management strategy of "raising the cold side / pressing the hot side" is executed. That is, the SP_T and / or weight of the cold side partition is increased and the ramp_up or soft power limit P_soft_limit of the hot side partition is limited. Power is not directly output to ensure that temperature difference management is achieved through the natural closed loop of the subsequent closed loop and budget allocation.

[0046] In this system, the total power budget and collaborative allocation module executes a collaborative strategy of "critical partition backup + allocation based on weight w / hotspot / efficiency eta" under the constraint of the total power upper limit P_total_max, and can further perform peak shaving, time-sharing, and peak-valley scheduling. At the same time, the "restricted information" (including the allocation reduction ratio or restricted flag) is fed back to the zone closed-loop temperature control module to trigger anti-integral saturation and / or control degradation, thereby avoiding temperature overshoot after the power constraint is released and improving system stability.

[0047] In this system, the power regulation / interlocking / diagnosis module is the highest priority module. It uses SCR / SSR / grading and other methods to convert PowerCommand into actual power P_act / P_out and transmit electrical parameters and status back. Its interlocking protection includes at least hard / soft interlocks for over-temperature, leakage, insulation abnormality and circuit fault. Hard interlocks take precedence over soft constraints and are executed to cut off / limit. Its diagnostics include at least electrical fault diagnosis and power-temperature rise consistency / thermal inertia residual diagnosis to identify heater degradation, insulation aging and sensor drift, and output DiagResult to drive self-calibration and maintenance decisions.

[0048] In this system, when the execution and power regulation / interlocking / diagnosis modules trigger interlocking or diagnose abnormalities, they write back ILK_state / ILK_code and DiagResult to the temperature difference constraint setpoint generation and constraint scheduling module and the total power budget and collaborative allocation module. This triggers the upstream recalculation mechanism for constraint convergence and power reallocation, ensuring that the upstream does not generate outputs that exceed hard boundaries and that the system maintains the key partition thermal boundary conditions even after local fault isolation. This forms a complete closed-loop safety strategy of "interlocking fallback - write back recalculation - collaborative protection".

[0049] In this system, the StartPermit is generated by integrating RiskLevel, the temperature difference window DeltaT_limit satisfaction status, the rate of change constraint satisfaction status, and thresholds such as upper / lower temperature limits T_hi / T_lo. The ReasonCode is then output to indicate the primary reason for prohibition or permission. Simultaneously, the trends / events / audits recorded and uploaded to DCS include at least ModeToken, RiskVector, ConstraintPack, PowerCommand, P_act, ILK status and diagnostic results, achieving full-chain traceability and auditability; among them, mode switching, interlock recovery and startup permission all adopt debouncing and hold time t_hold to avoid frequent jitter.

[0050] In a preferred embodiment: the execution and power regulation / interlocking / diagnosis module, the constraint scheduling module, the power budget and collaborative allocation module, and the start-up permission and recording communication module of the present invention, through the introduction of an interlocking write-back trigger recalculation state machine control mechanism, realize multi-level degradation and global protection control of the multi-zone anti-seize shaft heating process. Specifically, the system sets system state variables in the supervision and coordination layer to characterize the current control in different working states such as normal control, local soft degradation, local hard locking, global power degradation, global protection, and recovery observation; wherein, in the normal control state, each zone, under the constraints of ModeToken and RiskVector of the corresponding working condition, executes anti-seize shaft heating control according to the ConstraintPack output by the constraint / setpoint scheduling module and the PowerCommand output by the power budget and collaborative allocation module. The execution and power regulation module periodically collects the actual power P_act, current, voltage, insulation status, and temperature feedback of each heating circuit. When no interlocking event is detected, the system remains in the normal control state.

[0051] When the execution and power regulation / interlocking / diagnosis module detects a slight over-temperature, excessive temperature change rate, or power-temperature rise inconsistency in a certain partition, an L1-level soft interlocking event is generated, which includes the interlocking level, interlocking type ILK_state / ILK_code, partition diagnosis result DiagResult[z], actual power P_act[z], and power limitation flag. This data packet is written back to the supervision and coordination layer via the communication bus, and the system state switches from normal control to partial soft degradation state. The constraint / setpoint scheduling module automatically tightens constraint parameters for the partitions and their associated partitions that experience soft interlocking under local soft degradation conditions. This includes reducing the ramp-up, appropriately tightening the temperature difference window DeltaT_limit, and lowering the soft power limit P_soft_limit. If necessary, it lowers or freezes the temperature setpoint SP_T for that partition and reduces the weight w of abnormal measurement points in TrustedFrame, so that subsequent temperature difference assessment and risk calculation no longer overly rely on abnormal measurement points. The power budget and collaborative allocation module recalculates PowerCommand based on the updated ConstraintPack, reduces the power commands or priorities of the corresponding partitions, and prioritizes the thermal boundary conditions of critical partitions under the total power constraint, thereby completing a "constraint recalculation + power redistribution" based on interlock writeback.

[0052] When a zone experiences a severe overheating, short circuit, insulation breakdown, or other L2-level hard interlocking event, the execution and power regulation / interlocking / diagnosis module immediately cuts off the heating circuit locally. It then packages and writes back the zone's interlocking level, interlocking cause, actual power (zero), and diagnostic results to the higher-level system. The system state switches from normal control or partial soft degradation to partial hard lockout. In this state, the constraint / setpoint scheduling module locks or reduces the temperature setpoint SP_T of the hard-locked zone to a safe insulation temperature, converges its P_soft_limit to zero, and sets the zone's weight w to zero in the TrustedFrame, preventing it from participating in the calculation of the temperature T_hat. The power budget and collaborative allocation module accordingly sets the PowerCommand of the locked zone to zero. Simultaneously, while maintaining or decreasing the total power, it redistributes power to the remaining healthy zones, ensuring that critical zones such as bearings and bushings still receive the minimum heating power required for anti-seize under the new constraints. If the hard-locked partition is a critical partition that has a decisive impact on the risk of bearing seizure, or if the RiskLevel calculated by the Risk Index module exceeds the preset severe risk threshold, the Start Permit and Record Communication Module will further switch the system state to a global protection state, output a StartPermit signal that prohibits startup or requires shutdown, and record the specific interlocked partition and the reason for the interlock in the ReasonCode, retaining only the necessary safety insulation circuit.

[0053] In cases where multiple L1 soft interlock events occur across multiple zones, the actual total power value approaches or reaches the total power limit P_total_max, or the overall RiskLevel exceeds the high-risk threshold, the system state switches from normal control or local degradation state to global power degradation state after the power regulation / interlock / diagnosis module generates the corresponding feedback data packet. Under global power degradation state, the constraint / setpoint scheduling module uniformly tightens the ramp-up of each zone, appropriately increases the ramp-dn of the cooling zone, and reduces the soft power limit P_soft_limit for non-critical zones. The power budget and collaborative allocation module recalculates the power allocation strategy based on parameters such as critical zone safety margins, risk hotspots, and thermal efficiency eta, employing peak shaving and time-sharing polling to reduce grid load and ensure that the anti-shaft seizure requirements of critical zones are prioritized under total power constraints. If a critical partition L2 level hard interlock still occurs or the risk index continues to rise above the severe risk threshold during the global power degradation state, the state machine will further switch to the global protection state. The start-up permission module will issue a prohibition signal for starting, and the corresponding interlocking events, constraint tightening process, and power redistribution process will be completely recorded as trend and event information through the recording communication module.

[0054] After the global protection state or local hard lockout state is released, or after a certain period of time when the risk index falls back to a safe range and no new soft / hard interlock events are detected under global power degradation state, the execution and power regulation / interlock / diagnosis modules notify the supervision and coordination layer via feedback data packets that the current interlock has been cleared and trigger the recovery timer. The system state switches to recovery observation state. Under recovery observation state, the constraint / setpoint scheduling module does not immediately restore the original values ​​of the constraint parameters tightened in the previous state. Instead, it gradually relaxes ramp_up, P_soft_limit, and DeltaT_limit at a predetermined slope within the recovery observation timer t_hold, so that the system smoothly transitions from the degraded state back to the normal control state. The power budget and coordination allocation module correspondingly increases the power quota of each partition step by step, prioritizing the recovery of critical partitions, and then restoring general partitions. If an L1 or L2 interlock event occurs again during the recovery observation period, the state machine immediately switches back to the local soft degradation, local hard locking, global power degradation, or global protection state according to the interlock level and the type of affected partition, triggering a new round of constraint recalculation and power redistribution. Only when the recovery observation period ends and no new interlock or serious diagnostic anomaly is detected during the entire observation period will the system state switch back from the recovery observation state to the normal control state, and the corresponding constraint parameters and power distribution strategy will be restored to the design values ​​corresponding to the operating conditions, thus forming a complete state machine closed loop of "interlock detection - feedback write-back - state switching - constraint recalculation - power redistribution - risk and start-up permission update - recovery observation".

[0055] Working principle: During system operation, the multi-source acquisition module collects temperature, status, and electrical parameters related to the turbine anti-seize shaft heating device, and encapsulates the acquisition time, raw temperature data set, raw electrical parameter data set, and raw status data set into a raw data frame for output. The raw temperature data set can correspond to multiple measurement points such as bearing housing, bearing seat, bearing bush, lubricating oil, and optional shaft seal / sealing; the raw electrical parameter data set can include current, voltage, power, and insulation / leakage related parameters of each heating circuit; the raw status data set can include turning gear engagement, shutdown / restart / trip status, circuit permission / prohibition and switching status, etc. Through the above encapsulation, the timing consistency and traceability of signals from different sources are ensured for subsequent processing modules.

[0056] To avoid misleading control decisions due to anomalies in the original measurement points, this invention includes a reliable measurement and consistency verification module to perform validity assessment and consistency verification on the original data frames. Validity assessment includes at least the identification of open / short circuits, out-of-bounds errors, abrupt changes, and multi-measurement point consistency anomalies. Based on this, the module generates data quality indicator information and assigns usable weights to each temperature measurement point; at the same time, it merges or removes multiple measurement points within the same control zone by reducing their weights to obtain representative temperature data groups; and calculates the temperature change rate data group and temperature difference derived information, finally outputting a reliable data frame.

[0057] Therefore, the system's subsequent operating condition identification, risk assessment, constraint scheduling, power budgeting, and execution protection all use trusted data frames as a unified standard, reducing problems such as "dead pixels causing delays," "misjudged temperature differences," and "incorrect power adjustments" from the source.

[0058] The operating condition state machine and sequential control module identify the current operating condition based on trusted data frames and raw state data sets, and output operating condition identification information. The operating condition identification information includes at least the operating condition mode identifier, mode duration information, and event code information. At the same time, the module includes debouncing and hold logic, that is, when the operating condition switching conditions are met, the switching is not immediate, but is stably maintained according to the hold time parameter to avoid frequent fluctuations in the operating condition under critical conditions. Under different operating conditions, segmented switching and / or gating control are implemented for each heating zone to match the heating strategy with the operating condition and reduce temperature difference shocks during the switching process.

[0059] The bearing seizure risk index calculation module generates risk assessment information based on reliable data frames, temperature difference derivative information, temperature change rate data sets, oil temperature / oil film related conditions, and operating condition information. The risk assessment information includes at least the total risk index or risk level, sub-risk composition information, and hotspot indication information.

[0060] The sub-risk composition information includes at least one or more of the following: temperature difference sub-risk, temperature change rate sub-risk, oil temperature / oil film sub-risk, and operating condition sub-risk; hotspot indication information is used to indicate the heating zone with the greatest risk contribution and its dominant factor. Through the above risk decomposition and hotspot interpretation, the system achieves an interpretable output of "where is the most dangerous and what is the main cause of the danger," providing a basis for subsequent priority strategies for temperature difference management and coordinated power allocation.

[0061] The temperature difference constraint setpoint generation and constraint scheduling module generates and outputs constraint and setpoint packages based on operating condition identification information, risk assessment information, and available weight data groups of measuring points. The constraint and setpoint packages include at least the target temperature setpoint information for each heating zone, heating ramp constraint parameter information and / or cooling ramp constraint parameter information, temperature difference window limit information, and soft power upper limit information.

[0062] This invention emphasizes prioritizing temperature difference management: when the temperature difference exceeds the temperature difference window limit, the module increases the target temperature setpoint information and / or weight of the relatively cold side zone, and limits the temperature ramp constraint parameter information and / or soft power upper limit information of the relatively hot side zone to suppress the expansion of the temperature difference; It should be noted that the core of this step is to output the "set value and constraint boundary", rather than directly output the specific power. This allows the subsequent power budget and partitioned closed loop to converge collaboratively within a unified boundary, avoiding the further deterioration of the temperature difference caused by simply and crudely increasing the power.

[0063] The total power budget and collaborative allocation module budgets and allocates power under the constraint of the total power upper limit information, outputs power control command information and priority information for each partition, and outputs power-limited feedback information. When the total power is limited, the system prioritizes ensuring the minimum requirements of critical partitions, and then performs weighted allocation by combining the available weight data groups of measurement points, hot spot indication information and / or equivalent thermal efficiency parameters. If necessary, peak shaving control and / or time-sharing allocation can be performed to improve the safety and controllability in scenarios with limited total power.

[0064] The zone closed-loop temperature control module generates power demand information for each zone based on the target temperature setpoint information and ramp constraint parameter information, and receives power-limited feedback information. When the power-limited feedback information indicates the presence of a limited flag or allocation reduction ratio, the zone closed-loop temperature control module performs anti-integral saturation and / or degradation control to avoid temperature overshoot caused by integral accumulation during the limited period after power release, thereby improving the stability of zone control under total power constraints.

[0065] The execution and power regulation / interlock / diagnosis module converts power control command information into actual heating output and outputs execution feedback information. The execution feedback information includes at least actual output power information, interlock status information and / or interlock cause coding information, and health diagnosis result information. This module includes hard interlocks and soft interlocks. The trigger threshold of hard interlocks is higher than that of soft constraint thresholds and prioritizes the execution of cut-off or limiting to ensure safety boundary priority. Furthermore, when the interlock is triggered or the health diagnosis is abnormal, the execution module feeds back the interlock status information and / or health diagnosis result information to the temperature difference constraint setpoint generation and constraint scheduling module and the total power budget and collaborative allocation module, triggering upstream recalculation of constraint convergence and power reallocation, so that the system can still form closed-loop convergence under abnormal conditions instead of just remaining at "passive disconnection", thereby ensuring that the upstream will not continuously generate control outputs that break through hard boundaries.

[0066] The start-up permission and recording communication module integrates risk assessment information, temperature difference window satisfaction status, temperature change rate constraint satisfaction status, and interlock status information to output start-up permission information. The start-up permission information includes at least start-up permission / prohibition indication, risk level information, and cause code information. At the same time, the module records the trend, event, and audit records of operating condition identification information, constraint and set value packages, power control command information, execution feedback information, and alarm events, and sends them to the DCS / upper-level system through the communication interface to meet the needs of operation management, fault tracing, and start-up decision-making.

[0067] Example 2: To verify the effectiveness of the system of the present invention in actual use scenarios, the following example is taken: "preheating before restarting (including turning gear operation) after a long period of static storage following a cold shutdown and in a low ambient temperature". The functional verification and effect evaluation of the intelligent control system of the turbine anti-seize shaft heating device of the present invention are conducted. Specific examples are as follows: Example: Functional verification and effect evaluation of cold preheating + turning gear operation This embodiment addresses a scenario where the unit is shut down in a cold state, the ambient temperature is low, and the shutdown time is relatively long. In this situation, the temperature distribution in components such as the bearing housing / bearing seat / bearing bush is uneven, the lubricating oil temperature is low, and preheating and turning gear operation are required before restarting. The main risks in this scenario are: 1. Slow temperature rise on the cold side leads to a widening temperature difference; 2. Rapid heating on the hot side causes localized overheating and thermal shock; 3. When total heating power is limited, critical zones cannot be guaranteed; 4. The vibration during the switching between turning gear engagement and preheating processes causes control instability; 5. Sensor malfunctions or loop malfunctions cause the control strategy to deviate from the safety boundary.

[0068] Therefore, the verification objectives of this embodiment include: temperature difference window management capability, slope constraint effectiveness, total power budget allocation rationality, closed-loop integrity of interlocking / diagnostic write-back and recalculation, and interpretability of startup license output.

[0069] In this embodiment, the system is configured with at least the following heating zones and measuring points: Zone Z1 (critical bearing area): Multiple temperature measuring points are set around the heating circuit of the bearing housing / bearing seat / bearing bush. Zone Z2 (Lubricating Oil Zone): Corresponding to the oil tank / oil pipe / oil heater circuit, oil temperature measuring points are set; Zone Z3 (Auxiliary structure area, optional): Set temperature measuring points in areas such as the heating circuit of shaft seals / seals or adjacent structural components.

[0070] The system collects at least the following signals: Raw temperature data set (temperature at each measuring point); Status raw data group (shutdown / preheating / turning gear activation, loop enabled / disabled, etc.); Raw data set of electrical parameters (current / voltage / power, insulation / leakage status, etc. for each circuit).

[0071] The total power limit information is set to P_total_max (configured according to the on-site power supply capacity), and a configurable threshold system is enabled (temperature upper / lower limit threshold, temperature difference window limit threshold, temperature change rate threshold, model confidence threshold, hold time parameter, etc.).

[0072] After the system starts its preheating function, the multi-source acquisition module forms the original data frame. The reliable measurement and consistency verification module performs open circuit / short circuit, out-of-bounds, abrupt change, and consistency anomaly judgment on the temperature measurement points, generates data quality flag information, and obtains representative temperature data groups and usable weighted data groups for measurement points. At the same time, it outputs temperature change rate data groups and temperature difference derived quantity information to form a reliable data frame.

[0073] Verification point A (robustness of bad points): In this stage, a short-term jump (such as a sudden rise or fall) is deliberately set or simulated at a certain non-critical measuring point. The system should reduce the weight of the measuring point or remove it to keep the representative temperature and temperature difference derivative stable and not trigger the false heating surge or the false risk escalation.

[0074] The operating condition state machine identifies the current mode as "cold preheating" and starts the de-jittering and holding logic; then, when the conditions for turning gears to be engaged are met, the mode switches to the "turning gears + preheating" combined mode.

[0075] The system first opens the gating of the lubricating oil zone and the critical bearing zone according to the sequential control strategy, so that the oil temperature and the critical bearing zone enter the controlled temperature rise range first, and then the auxiliary zones are gradually opened.

[0076] Verification Point B (Switching Stability): Changes in state such as turning gear engagement / disengagement and preheating phase transitions should not cause mode jitter; mode duration information and hold time parameters should ensure smooth switching of operating conditions and avoid frequent fluctuations in power commands.

[0077] The bearing seizure risk index calculation module calculates risk assessment information based on reliable data frames and outputs hotspot indication information. If the temperature difference between the critical bearing area and the lubricating oil is large or the temperature change rate is abnormal, the hotspot should point to the corresponding partition and give the dominant factor (temperature difference or change rate or oil temperature condition).

[0078] Verification point C (hotspot explanation): When the temperature difference widens, the hotspot should fall on the critical partition that caused the temperature difference and drive subsequent constraint scheduling to tighten the temperature difference window governance.

[0079] The temperature difference constraint setpoint generation and constraint scheduling module generates constraint and setpoint packages (target temperature setpoint, heating / cooling ramp, temperature difference window limit, soft power limit). When the temperature difference derived quantity indicates that the temperature difference window limit is exceeded, the system executes the "temperature difference management priority" strategy: increase the target temperature setpoint information and / or weight of the relatively cold side zone, while limiting the ramp and / or soft power limit of the relatively hot side zone.

[0080] Verification point D (effectiveness of temperature difference management): When the temperature difference exceeds the limit, the system should show a trend of "raising the cold side / pressing the hot side", rather than blindly increasing the power as a whole; after the temperature difference gradually returns to the window, the ramp and the set value should recover smoothly.

[0081] Under the constraint of the total power limit, the total power budget and collaborative allocation module performs critical partition protection and performs weighted allocation based on weight, hot spot and efficiency, and performs peak shaving or time-sharing polling when necessary; at the same time, it outputs power-limited feedback information to the partition closed loop.

[0082] The zoned closed-loop temperature control module generates zoned power requirements based on the target temperature and ramp, and performs anti-integral saturation / degradation under constrained feedback to prevent temperature overshoot after the constraint is removed.

[0083] Verification point E (stability under constraints): When the total power is insufficient, the critical bearing zone should obtain a minimum power; the closed-loop control should not exhibit significant temperature overshoot or power oscillation after the constraints are lifted.

[0084] The execution and power regulation / interlock / diagnostic module translates power control commands into actual output power and monitors hardware / soft interlocks and health diagnostic results in real time.

[0085] During the verification process, an insulation abnormality or circuit fault can be simulated in a certain auxiliary zone, triggering hard interlocking or soft interlocking limit. The execution module outputs interlocking status information / cause coding information and writes it back to the temperature difference constraint setpoint generation and constraint scheduling module and the total power budget and collaborative allocation module, triggering constraint convergence and power redistribution recalculation.

[0086] Verification point F (closed-loop integrity): After interlocking is triggered, the upstream should automatically recalculate and transfer or limit the power from the fault zone, and the critical zone control should remain controllable; the system should not continue to generate instructions that break through hard boundaries.

[0087] The start-up permission and recording communication module integrates risk assessment information, temperature difference window satisfaction status, rate of change constraint satisfaction status, and interlock status information to output start-up permission information and generate cause code information; at the same time, it records the trend / event / audit of operating condition identification information, constraint and set value packages, power commands, execution feedback, interlock and diagnostic events and sends them to DCS / supervisor system.

[0088] Verification point G (explainable and traceable): When startup is not allowed, the cause code can clearly point to the main factors such as excessive temperature difference, excessive rate of change, insufficient oil temperature, or failure to release interlock; and the complete process link data can be replayed at the DCS terminal.

[0089] In this embodiment, at least the following indicators are used to evaluate the effect: 1. Temperature difference window compliance rate: The percentage of the temperature difference that remains within the window limit throughout the entire preheating process; 2. Overshoot and Impact Suppression: This indicates whether the ramp during the temperature rise process satisfies the constraints and whether there is significant overshoot. 3. Critical Zone Protection Level: When total power is limited, do critical zones continuously receive minimum power and maintain safe temperature boundaries? 4. Control stability: Whether the power command and temperature curve are smooth and free from frequent fluctuations during the switching of operating conditions (cold preheating → turning gear operation); 5. Abnormal fallback capability: After interlocking is triggered, does it achieve closed-loop convergence of "execution of limit / cut-off → write-back → upstream recalculation → redistribution"? 6. Reliability and interpretability of start-up permission: Whether the start-up permission / prohibition, risk level, and reason code are consistent with the on-site situation and are traceable.

[0090] Verified in the above scenarios, this invention forms a complete closed loop through "unified caliber of trusted data frames + debouncing and sequential control under operating conditions + risk hotspot driving + priority constraint scheduling for temperature difference management + guaranteed total power budget and weighted allocation + restricted feedback to resist saturation / degradation + interlocking / diagnostic write-back and recalculation + start-up permission and audit records". In actual use scenarios of cold preheating and turning gear operation, it can effectively suppress the risk of temperature difference expansion and thermal shock. Even under conditions of limited total power and abnormal circuits, it still maintains priority satisfaction of key partition boundaries and provides interpretable start-up permission and traceable records, thereby achieving safe, stable and coordinated control of anti-seize shaft heating.

[0091] Example 3: Hot restart (short shutdown time, high but uneven component temperature) To verify the effectiveness of the system of the present invention under hot restart conditions, a restart scenario with a short downtime and the overall temperature of the unit still at a high level but uneven temperature distribution in each heating zone is used as an example for functional verification and effect evaluation. In this scenario, the temperature of the bearing housing / bearing seat / bearing bush and the lubricating oil area has not completely dropped. However, due to the differences in heat dissipation paths and thermal inertia of different parts, there is a risk of rapid amplification of temperature difference and continued heating of local hot side leading to thermal shock.

[0092] In this scenario, the system uses a multi-source acquisition module to collect temperature, status, and electrical parameters and form raw data frames. The reliable measurement and consistency verification module identifies open circuits, short circuits, out-of-bounds, sudden changes, and consistency anomalies at temperature measurement points, generates data quality flags and available weights for measurement points, and forms reliable data frames representing temperature, temperature change rate, and temperature difference derivatives. The operating condition state machine and sequential control module identify the entry into the hot restart mode and enable debouncing and holding logic to avoid power command jitter during mode switching. The bearing seizure risk index calculation module outputs the risk level, hot spot partitions, and dominant factors as directional inputs for constraint scheduling and power allocation.

[0093] Subsequently, the temperature difference constraint setpoint generation and constraint scheduling module outputs a constraint and setpoint package, including the target temperature setpoint, heating / cooling ramp constraints, temperature difference window limit, and soft power limit. When the temperature difference exceeds the temperature difference window limit, the system executes a temperature difference management priority strategy. This strategy increases the target setpoint and / or weight of the relatively cold side partition, while limiting the heating ramp and / or soft power limit of the relatively hot side partition, thus suppressing further expansion of the temperature difference. The total power budget and collaborative allocation module provides a safety net for key partitions under the total power limit constraint and performs weighted allocation based on weight, hotspot, and / or efficiency information. Simultaneously, it sends power-limited feedback information to the partition closed-loop temperature control module. The partition closed-loop temperature control module then performs anti-integral saturation and / or degradation control to avoid temperature overshoot after the limitation is lifted. The execution and power adjustment module converts the power control command into actual output and sends back the actual power and electrical parameters to ensure stable closed-loop operation.

[0094] Through the above process, controlled convergence of temperature difference and smooth constraint of the heating ramp can be achieved during the hot restart phase, preventing the high-temperature zone from continuing to "heat up rapidly" and causing thermal shock, while the cold side zone receives more targeted compensation heating; the switching of operating conditions remains stable under the de-jittering effect, and the power command is smooth and continuous; this embodiment shows that the present invention does not require additional special hardware modifications, and can achieve feasible control effects of prioritizing temperature difference management, controlling the ramp and coordinating power distribution in the hot restart scenario by relying only on temperature / state / electrical parameter acquisition and conventional power actuators, and can output interpretable start-up permission and upload recorded data.

[0095] Example 4: Rapid recovery after tripping (frequent interlocking, more severe power limitation) To verify the effectiveness of the system of the present invention in the case of rapid recovery after tripping, the functional verification and effect evaluation are carried out by taking the scenario of rapid recovery after the unit trips as an example. This scenario is usually accompanied by significant temperature distribution disturbances, abnormal temperature change rate, and interlocking events such as insulation / leakage or circuit failure in some heating circuits. In addition, there are often stricter total power upper limit constraints to ensure power supply safety, which makes temperature difference control, power distribution and safety backup more difficult.

[0096] In this scenario, the multi-source acquisition module forms the original data frame, the reliable measurement and consistency verification module outputs the reliable data frame and performs gating processing on bad points; the operating condition state machine and sequential control module identify the entry into the fault / trip recovery mode and put it into segmented gating according to the sequential control strategy, prioritizing the opening of key partitions to meet the thermal boundary conditions. After the bearing risk index calculation module outputs the risk level and hot spot partition factor, the temperature difference constraint setpoint generation and constraint scheduling module correspondingly tightens the constraint and setpoint package, that is, sets the temperature difference window limit to a more conservative range, reduces the allowable temperature rise slope and compresses the soft power upper limit to reduce the risk of secondary impact and temperature difference expansion.

[0097] Due to stricter total power constraints, the total power budget and collaborative allocation module implements a critical partition safety net under the total power upper limit, and performs weighted allocation based on weight, hotspot, and / or efficiency information. When necessary, peak shaving and time-sharing polling methods are used to compensate for heating non-critical partitions. At the same time, power-limited feedback information is sent to the partition closed-loop temperature control module to trigger anti-integral saturation / degradation control. The execution and power regulation / interlocking / diagnosis modules implement power commands and monitor hard and soft interlocks. When an interlock is triggered or a diagnostic anomaly occurs, the execution module prioritizes cutting off or limiting the power and outputs the interlock status / cause code and health diagnosis results. At the same time, the above information is written back to the constraint scheduling module and the power budget module to trigger upstream recalculation, so that the system can automatically complete constraint convergence and power redistribution, avoid the upstream from continuously issuing invalid or out-of-bounds control outputs to faulty partitions, and ensure that critical partitions are continuously under control.

[0098] Through the above process, even under conditions of frequent interlocking and limited total power, the system can still maintain an operating state of "prioritizing critical zones, controllable temperature difference windows, and convergent control outputs." When the interlock is not released, the start-up permission and recording communication module outputs a prohibition and provides a reason code. After the interlock is released and the risk decreases, the output is allowed. The operating conditions, constraint packages, power commands, execution feedback, interlocking and diagnostic events are recorded as trends / events / audits and sent to the DCS. This embodiment shows that the present invention, through the closed-loop mechanism of "execution interlock fallback + write-back trigger recalculation + power redistribution," can still achieve a feasible and stable anti-seize shaft collaborative heating control effect in harsh scenarios of rapid recovery after tripping, and has clear and interpretable start-up criteria and traceable recording capabilities.

[0099] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. An intelligent control system for a turbine anti-shaft seizure heating device, characterized in that, include: A multi-source acquisition module is used to acquire temperature, status and electrical parameters, and output raw data frames. The raw data frames include at least: acquisition time information, temperature raw data set, electrical parameter raw data set and status raw data set. The reliable measurement and consistency verification module is used to perform validity judgment and consistency verification on the original data frame and output a reliable data frame. The reliable data frame includes at least: a representative temperature data group, a data group of available weights for measurement points, data quality flag information, a temperature change rate data group, and temperature difference derived quantity information. The operating condition state machine and sequential control module are used to identify the operating condition based on the trusted data frame and perform segmented switching and / or gating control, and output operating condition identification information. The operating condition identification information includes at least: operating condition mode identifier, mode duration information and event encoding information, and the operating condition state machine and sequential control module includes debouncing and hold logic. The bearing seizure risk index calculation module is used to calculate and output risk assessment information based on temperature difference, temperature change rate, oil temperature / oil film conditions and operating condition information. The risk assessment information includes at least: total risk index or risk level, sub-risk composition information and hot spot indication information. The temperature difference constraint setpoint generation and constraint scheduling module is used to generate and output constraint and setpoint packages based on the operating condition identification information, risk assessment information and available weight data group of measuring points. The constraint and setpoint packages include at least: target temperature setpoint information for each heating zone, heating ramp constraint parameter information and / or cooling ramp constraint parameter information, temperature difference window limit information and soft power upper limit information for each heating zone. The total power budget and collaborative allocation module is used to budget and allocate power under the constraint of the total power upper limit information, and output the power control command information and priority information of each heating zone, while also outputting power-limited feedback information. The zoned closed-loop temperature control module is used to generate power demand information for each heating zone based on the target temperature setpoint information and ramp constraint parameter information, and to perform anti-integral saturation and / or degradation control according to the power-limited feedback information. The execution and power regulation / interlock / diagnosis module is used to convert the power control command information into actual heating output and output execution feedback information. The execution feedback information includes at least: actual output power information of each heating zone, interlock status information and / or interlock cause coding information, and health diagnosis result information. The start-up permission and recording communication module is used to output start-up permission information by integrating the risk assessment information, temperature difference window satisfaction status, temperature change rate constraint satisfaction status, and interlock status information. The start-up permission information includes at least: start-up permission / prohibition indication, risk level information, and cause code information. It is also used to perform trend recording and event auditing of operating condition identification information, constraint and set value package, power control command information, execution feedback information, and alarm events, and send them to the DCS / supervisor system. When the interlock is triggered or the health diagnosis is abnormal, the execution and power regulation / interlock / diagnosis module feeds back the interlock status information and / or health diagnosis result information to the temperature difference constraint setpoint generation and constraint scheduling module and the total power budget and collaborative allocation module, so as to trigger the upstream recalculation of constraint convergence and power reallocation. Furthermore, the system also includes an online thermal inertia identification module, which is used to identify the thermal inertia time constant parameters, equivalent thermal efficiency parameters, and thermal response hysteresis parameters of each heating zone online using the actual output power information as a reference, and output model confidence information. The output of the online thermal inertia identification module is used by the temperature difference constraint setting value generation and constraint scheduling module and / or the total power budget and collaborative allocation module to adaptively adjust the constraints and power allocation. The system operates in a layered, cyclical manner, comprising: a fast loop centered on the execution and power regulation / interlocking / diagnosis module; a control loop consisting of the multi-source acquisition module, the reliable measurement and consistency verification module, the total power budget and collaborative allocation module, and the execution and power regulation / interlocking / diagnosis module; a strategy loop consisting of the operating condition state machine and sequential control module, the bearing seizure risk index calculation module, and the temperature difference constraint setpoint generation and constraint scheduling module; a slow identification loop consisting of the thermal inertia online identification module; and a background recording loop consisting of the start-up permission and recording communication module.

2. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The reliable measurement and consistency verification module performs validity checks on temperature measurement points for open circuits, short circuits, out-of-bounds, abrupt changes, and consistency anomalies to generate the data quality flag information, and forms the representative temperature data group and the available weighted data group for measurement points based on the data quality flag information.

3. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The debouncing and holding logic of the working condition state machine and the sequential control module includes a holding time parameter, and the working condition identification information includes mode duration information; The operating condition state machine and sequential control module are used to perform segmented switching and / or gating control on the heating zone to achieve stable maintenance of the operating condition switching process.

4. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The model confidence information output by the online thermal inertia identification module corresponds to the preset model confidence threshold. When the model confidence information is lower than the preset model confidence threshold, the online thermal inertia identification module keeps the thermal inertia time constant parameter, equivalent thermal efficiency parameter and thermal response hysteresis parameter of the previous moment unchanged or enters a frozen state.

5. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The sub-risk composition information includes at least one or more of the following: temperature difference sub-risk, temperature change rate sub-risk, oil temperature / oil film sub-risk, and operating condition sub-risk. The hot spot indication information is used to indicate the heating zone with the greatest risk contribution and its dominant factor.

6. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, When the temperature difference constraint setpoint generation and constraint scheduling module detects that the temperature difference exceeds the temperature difference window limit information, it executes a temperature difference management priority strategy. This strategy involves increasing the target temperature setpoint information and / or the weight corresponding to the available weight data group of the measuring point in the relatively cold side zone, and limiting the temperature rise slope constraint parameter information and / or soft power upper limit information in the relatively hot side zone, in order to suppress the expansion of the temperature difference.

7. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The total power budget and collaborative allocation module performs a minimum allocation to key partitions under the constraint of the total power upper limit information, and performs weighted allocation based on the available weighted data group of measurement points, hot spot indication information and / or equivalent thermal efficiency parameters; and the total power budget and collaborative allocation module is further used to perform peak shaving control and / or time-sharing polling allocation.

8. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The power-limited feedback information includes at least a limited flag information and / or a reduction ratio information. The zoned closed-loop temperature control module performs anti-integral saturation and / or degradation control based on the power-limited feedback information to avoid temperature overshoot after the power limitation is lifted.

9. The intelligent control system for a turbine anti-shaft seizure heating device according to claim 1, characterized in that, The execution and power regulation / interlocking / diagnosis module includes hard interlocking and soft interlocking. The trigger threshold of hard interlocking is higher than that of soft constraint and it takes priority to execute cut-off or limiting. The startup permission and recording communication module outputs the startup permission information based on a configurable threshold system. The configurable threshold system includes at least one or more of the following: upper temperature limit threshold and / or lower temperature limit threshold, temperature difference window limit threshold, temperature change rate threshold, available weight threshold for measurement points, model confidence threshold, upper total power limit threshold, and de-jitter hold time threshold. It also performs trend recording, event recording, and audit recording of operating condition identification information, risk assessment information, constraint and setpoint package, power control command information, interlock status information, and health diagnosis result information.