Monitoring and diagnosis system for DCS (Distributed Control System) of thermal power plant

Abstract

The invention relates to the field of power plant control, and discloses a monitoring and diagnosis system for a DCS (Distributed Control System) of a thermal power plant, which is used for realizing organic collaboration of an equipment layer, a unit layer and a whole plant layer by establishing a whole plant unified dynamic model and a multi-time scale coordination mechanism. Comprising the following steps: establishing a whole-plant unified dynamic model and constructing a performance index description framework; solving a whole plant coordination control optimization problem in real time based on the model and the framework, and generating a multi-level optimization control instruction; state prediction is carried out on the whole-plant dynamic model, a whole-plant health state vector is calculated, an optimization control instruction is output and executed in real time, and a safe operation state evaluation report is generated; and displaying an evaluation report on a DCS monitoring interface, and dynamically adjusting model parameters. According to the invention, collaborative optimization control from an equipment level to a whole plant level is realized, and a complete monitoring and diagnosis system is constructed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power plant control, and in particular to a monitoring and diagnosis system for a DCS control system of a thermal power plant. BACKGROUND

[0002] As an important link of energy supply, the monitoring and diagnosis level of the distributed control system (DCS) of a thermal power plant is directly related to the safety, economy and reliability of the unit operation. With the deepening of the reform of the electricity market and the large-scale access of renewable energy, thermal power units need to participate in grid peak shaving frequently, and the operating conditions are increasingly complex, which puts forward higher requirements for the DCS control system. The traditional DCS control system of a thermal power plant adopts a hierarchical distributed architecture, and each subsystem operates relatively independently, lacking a coordinated optimization mechanism at the whole plant level.

[0003] The existing DCS system usually adopts a hierarchical control mode at the device level and the unit level, lacking a unified dynamic description framework at the whole plant level. The information island phenomenon is serious between subsystems, and the dynamic coupling relationship between main devices such as boilers, steam turbines and generators cannot be effectively modeled, resulting in insufficient coordination of control commands and affecting the overall operation efficiency of the whole plant. The traditional monitoring method is mostly based on threshold alarm and simple logical judgment, lacking quantitative evaluation and trend prediction of the system health status. Stability analysis is often limited to local devices or subsystems, and a unified stability monitoring index system cannot be constructed at the whole plant level, making it difficult to realize early warning and accurate positioning of faults. The existing optimization control method mostly adopts a strategy based on empirical rules or traditional PID control, with a single optimization target, often focusing on the optimization of a certain specific performance indicator (such as thermal efficiency), lacking a multi-objective coordinated optimization mechanism considering thermodynamic performance, mechanical performance, electrical performance and environmental performance. The existing technology lacks an effective whole-plant coordinated performance evaluation system, which cannot accurately identify the coordination bottlenecks between subsystems, restricting the further improvement of the overall operation efficiency of the whole plant, and the coordinated response capability of each system of the unit is insufficient when responding to grid peak shaving requirements.

[0004] Therefore, we propose a monitoring and diagnosis system for a DCS control system of a thermal power plant to solve the above problems. SUMMARY

[0005] The present application provides a monitoring and diagnosis system for a DCS control system of a thermal power plant, which realizes the organic coordination of the device layer, the unit layer and the whole plant level by establishing a unified dynamic model of the whole plant and a multi-time scale coordination mechanism.

[0006] The first aspect of this invention provides a monitoring and diagnostic system for a DCS control system of a thermal power plant. The monitoring and diagnostic system includes: a framework construction module for establishing a unified dynamic model of the entire plant, obtaining the dynamic model product, and constructing a unified description framework product for the plant's performance indicators; a coordinated control module for solving the coordinated control optimization problem of the entire plant in real time based on the dynamic model product and the unified description framework product, generating multi-level optimized control command products; a state prediction module for predicting the state of the dynamic model product, calculating the plant's health state vector product based on the prediction results, constructing a plant-level Lyapunov function for stability monitoring, and generating plant-wide stability monitoring index products; a report generation module for outputting and executing the multi-level optimized control command products in real time, and simultaneously generating a plant-wide safe operation status assessment report product based on the plant-wide health state vector product and the plant-wide stability monitoring index products; and a parameter adjustment module for displaying the plant-wide safe operation status assessment report product in real time on the DCS monitoring interface, and dynamically adjusting the dynamic parameters and coupling coefficients in the dynamic model product based on the assessment results.

[0007] Optionally, in a first implementation of the first aspect of the present invention, a comprehensive optimization objective function is constructed based on the unified description framework product of the plant-wide performance indicators, and an optimization problem description product is generated; based on the plant-wide dynamic model product, equipment operation constraints and system safety boundary conditions are extracted, and a constraint set product is generated; based on the optimization problem description product and the constraint set product, an optimization problem functional is established, and a variational optimization problem product is generated; based on the variational optimization problem product, the optimal control trajectory of the system is obtained, and a preliminary optimization control command product is generated; based on the preliminary optimization control command product and the constraint set product, a multi-level optimization control command product is generated.

[0008] Optionally, in the second implementation of the first aspect of the present invention, a multi-time-scale state predictor is constructed based on the plant-wide dynamic model product to generate multi-time-scale state prediction data products; based on the multi-time-scale state prediction data products, differential feature quantities of the state variables of each device are extracted to generate device differential feature set products; based on the device differential feature set products, health indicators are calculated to generate plant-wide health state vector products; based on the multi-time-scale state prediction data products, a Lyapunov function reflecting the plant-wide operational stability is constructed to generate system stability judgment function products; based on the system stability judgment function products, system stability indicators and stability boundary distances are calculated in real time to generate plant-wide stability monitoring indicator products.

[0009] Optionally, in the third implementation of the first aspect of the present invention, the process of generating the plant-wide stability monitoring index product further includes: establishing an equipment state change trend early warning mechanism based on the equipment differential feature set product, and generating an equipment state trend early warning product; constructing a system dynamic response characteristic analysis model based on the multi-timescale state prediction data product, and generating a dynamic response characteristic evaluation product; identifying potential system instability risk modes based on the equipment state trend early warning product and the dynamic response characteristic evaluation product, and generating a system instability risk early warning product; formulating a graded early warning response strategy based on the system instability risk early warning product, and generating an early warning response strategy product; generating a system stability enhancement control suggestion product based on the early warning response strategy product and the plant-wide health state vector product; and establishing an early warning effect verification mechanism, generating an early warning effect evaluation product by comparing the system stability changes before and after the early warning.

[0010] Optionally, in the fourth implementation of the first aspect of the present invention, the multi-level optimized control instruction products are classified and distributed according to three levels: equipment level, unit level, and plant-wide level, generating hierarchical control instruction execution scheme products; a multi-dimensional judgment matrix is ​​constructed based on the plant-wide health status vector product and the plant-wide stability monitoring index product to generate a safety status judgment benchmark product; based on the hierarchical control instruction execution scheme product and the safety status judgment benchmark product, a matching analysis of control instructions and system status is performed to generate an instruction execution safety assessment product; based on the instruction execution safety assessment product, combined with real-time operating data, a comprehensive plant-wide safety status rating is performed to generate a safety status level assessment product; based on the safety status level assessment product, a plant-wide safety operation status assessment report product is generated.

[0011] Optionally, in the fifth implementation of the first aspect of the present invention, a comprehensive security score is calculated using a multi-dimensional fusion algorithm. The resulting safety status level assessment product is: in, Let i be the weight of the i-th indicator. Let be the score for the i-th indicator, n be the total number of security indicators, and i be the indicator index, representing the i-th specific security assessment indicator.

[0012] Optionally, in the sixth implementation of the first aspect of the present invention, the process of generating the plant-wide safety operation status assessment report product further includes: establishing a dynamic adjustment mechanism for safety boundaries based on real-time operation data and equipment status monitoring information to generate an adaptive safety boundary product; performing boundary compliance verification on the graded control instruction execution scheme product based on the adaptive safety boundary product to generate a boundary verification result product; conducting an operation risk probability assessment based on the boundary verification result product and the safety status level assessment product to generate a risk probability distribution product; formulating graded risk response strategies based on the risk probability distribution product to generate a risk prevention and control strategy product; establishing an assessment effect feedback mechanism to generate an assessment accuracy analysis product by comparing the assessment results with the actual operation status; and optimizing the parameter settings of the multi-dimensional fusion judgment algorithm based on the assessment accuracy analysis product to generate an optimized safety assessment model product.

[0013] Optionally, in the seventh implementation of the first aspect of the present invention, a multi-level visualization display area is configured to display key information of the plant-wide safety operation status assessment report product in real time, generating a visualization monitoring interface product; based on the plant-wide safety operation status assessment report product, system performance deviation data and abnormal operation modes are extracted to generate a model parameter adjustment requirement product; based on the model parameter adjustment requirement product and combined with real-time operating condition data, a dynamic parameter optimization strategy is formulated to generate a parameter adaptive adjustment scheme product; based on the parameter adaptive adjustment scheme product, the plant-wide dynamic model product is corrected online to generate an optimized plant-wide dynamic model product; by comparing the prediction accuracy of the optimized plant-wide dynamic model product with the actual operating data, a model optimization effect evaluation product is generated; based on the model optimization effect evaluation product, the parameter adjustment rule base is dynamically updated to generate an adaptive optimization knowledge base product.

[0014] Optionally, in the eighth implementation of the first aspect of the present invention, a coordination optimization module is further included: based on the plant-wide safety operation status assessment report product and actual operation data, the overall coordination performance index of the system is calculated, and a plant-wide coordination performance assessment product is generated; based on the plant-wide coordination performance assessment product, coordination bottlenecks and optimization potential areas among subsystems are identified, and a coordination optimization key area identification product is generated; based on the coordination optimization key area identification product, a hierarchical coordination optimization strategy product is generated; based on the hierarchical coordination optimization strategy product, a coordination parameter optimization scheme product is generated; a coordination optimization effect tracking mechanism is established, and a coordination optimization effect verification product is generated by comparing the system performance changes before and after the adjustment of coordination parameters; based on the coordination optimization effect verification product, the coordination optimization knowledge base and strategy rules are updated, and an adaptive coordination optimization knowledge base product is generated.

[0015] The mechanism of this invention is as follows: by establishing a multi-timescale dynamic model that includes equipment layer, unit layer and plant-wide layer, a global description and collaborative optimization of the entire production process can be achieved; Beneficial effects: A three-level coupled dynamic model including equipment layer, unit layer and plant-wide layer was constructed, which broke through the limitations of traditional hierarchical control, established a multi-time scale coordination mechanism, realized the organic unity of second-level, minute-level and hour-level dynamics, and proposed a unified description framework for plant-wide performance indicators, providing a theoretical basis for comprehensive plant control. Through multi-level coordination and control, the various subsystems work together under a unified framework, eliminating information silos and establishing a complete plant-wide monitoring system, providing technical support for the overall optimization of the production line; By dynamically adjusting parameters and correcting the model online, the system can adapt to different operating conditions, an adaptive optimization knowledge base has been built, continuous improvement of control strategies has been achieved, a dynamic adjustment mechanism for safety boundaries has been established, and the robustness of the system has been improved. By using plant-wide health status vectors and stability monitoring indicators, early warning of faults was achieved, a multi-dimensional safety assessment system was established, operational risks were significantly reduced, a graded risk prevention and control strategy was developed, and the system's emergency response capabilities were enhanced. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of an embodiment of the monitoring and diagnostic system of the DCS control system of a thermal power plant according to the present invention; Figure 2 This is a schematic diagram of another embodiment of the monitoring and diagnostic system of the DCS control system of a thermal power plant in this invention; Figure 3 This is a schematic diagram illustrating the overall safe operation status assessment of the DCS control system of a thermal power plant in an embodiment of the present invention. Figure 4 This is a schematic diagram of an embodiment of the monitoring and diagnostic equipment for the DCS control system of a thermal power plant in this invention. Detailed Implementation

[0017] This invention provides a monitoring and diagnostic system for a DCS control system in a thermal power plant, which achieves organic coordination at the equipment, unit, and plant levels by establishing a unified dynamic model for the entire plant and a multi-timescale coordination mechanism. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0018] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the monitoring and diagnostic system for the DCS control system of a thermal power plant in this invention includes: 101. Framework construction module, used to establish a unified dynamic model of the whole plant based on a multi-time-scale differential algebraic coordination system, to obtain a whole plant dynamic model product with a three-level coupled structure including equipment-level second-level dynamics, unit-level minute-level dynamics and whole plant-level hour-level dynamics, and to construct a unified description framework product of the whole plant performance indicators. It is understood that the executing entity of this invention can be a monitoring and diagnostic device of a DCS control system in a thermal power plant, or it can be a terminal or a server; no specific limitation is made here. This embodiment of the invention will be described using a server as an example.

[0019] It should be noted that in the monitoring and diagnosis of DCS control systems in thermal power plants, this model needs to be compatible with dynamic behaviors at different time scales (equipment level at the second level, unit level at the minute level, and plant-wide level at the hour level) and form a unified descriptive framework for performance indicators. The following example is a typical 600MW coal-fired power plant.

[0020] Thermal power plant configuration: Includes 1 boiler, 1 steam turbine, 1 generator, and auxiliary systems (feedwater and combustion control). The DCS system integrates approximately 5,000 monitoring points (temperature, pressure, flow, and vibration sensors), with data sampling frequency set according to hierarchical differences.

[0021] Modeling Objective: To describe three levels of coupled dynamics using differential-algebraic equations, where: Equipment Level (second-level dynamics): Reflects rapid response processes of sensors and actuators (turbine blade vibration, generator current fluctuations), with dynamic response times within 0.1-10 seconds. Unit Level (minute-level dynamics): Describes the control loop of the boiler-turbine-generator unit (combustion control, steam temperature regulation), with dynamic response times between 10 seconds and 10 minutes. Plant-wide Level (hour-level dynamics): Involves slow processes such as plant-wide economic scheduling and load allocation, with response times exceeding 1 hour.

[0022] Equipment Layer Model (Second-Level): Based on physical mechanisms and sensor data, differential equations are established to describe the transient behavior of key equipment: Turbine Rotor Vibration Model: The state variable is the vibration amplitude (unit: μm), with a sampling frequency of 100Hz (updated every 0.01 seconds). Characteristic frequencies (50Hz fundamental frequency) are extracted through spectrum analysis and coupled with current signals (generator current fluctuation amplitude ±5A). Data Sources: Vibration sensors (accuracy ±0.1μm) and current sensors (accuracy ±0.5A), covering 10 major pieces of equipment.

[0023] Unit-level model (minute-level): Algebraic coordination equations are used to link the equipment-level outputs. Boiler drum water level control loop: Inputs: Feedwater flow rate (second-level data from the equipment level), steam flow rate (minute-level average). State variable: Drum water level (unit: mm). The control setpoint is updated every 30 seconds and coupled with equipment-level vibration data via a PID controller (adjusting the feedwater valve opening when vibration exceeds limits). Coupling method: Second-level vibration data from the equipment level (over-threshold alarm) triggers minute-level adjustments in the unit-level control logic.

[0024] Plant-wide model (hourly level): Slow dynamic equations are constructed based on material / energy balance. Plant-wide coal consumption rate model: The inputs are the average power (MW) output of the unit layer and coal quality data, and the state variable is the hourly coal consumption (tons / hour). Load allocation is optimized once every hour. Coupling method: The minute-level power fluctuation (±2MW) of the unit layer is used as the boundary condition for the plant-wide economic scheduling.

[0025] Coupling is achieved through interface variables of the differential-algebraic system: equipment-level outputs (vibration amplitude, current) serve as inputs to the unit-level equations. Unit-level outputs (boiler efficiency, turbine power) are transmitted to the entire plant level. Key coupling coefficients: when the vibration amplitude exceeds 50 μm, the weight of the unit-level control loop increases by 20%; the correlation coefficient between the overall plant level coal consumption rate and the unit-level power is 0.95.

[0026] Examples of dynamic parameters: Equipment layer time constant: 0.5 seconds (vibration attenuation). Unit layer hysteresis time: 2 minutes (temperature regulation). Plant-wide inertia time: 60 minutes (load response).

[0027] A standardized indicator library is constructed to map the three-level outputs to a unified evaluation dimension: Economic indicators: coal consumption for power supply (g / kWh), plant power consumption rate (%). Safety indicators: equipment health (0-100%), stability margin (based on the initial value of the Lyapunov function). Environmental indicators: NO x Emission concentration (mg / m³) 3 ).

[0028] The framework integrates three levels of data through a weight matrix: vibration data at the equipment level has a weight of 0.3 (second level), efficiency data at the unit level has a weight of 0.5 (minute level), and coal consumption data at the plant level has a weight of 0.2 (hour level).

[0029] The parameter summary table is as follows (data is based on a simplified version of a typical 600MW unit): 102. Coordination and control module, which is used to solve the plant-wide coordination and control optimization problem in real time through variational optimization algorithm based on the plant-wide dynamic model product and the plant-wide performance index unified description framework product, and generate multi-level optimization control instruction products including equipment-level adjustment instructions, unit-level setpoint instructions and plant-wide coordination instructions. It should be noted that, taking a 600MW coal-fired power plant as an example, variational optimization algorithms are used to coordinate the dynamic model of the entire plant (equipment level at the second level, unit level at the minute level, and plant-wide level at the hour level) with the performance index framework, solve the optimization problem in real time, and generate three-level control commands.

[0030] Power plant configuration: One 600MW supercritical unit, including boiler-turbine-generator unit and auxiliary systems. The DCS system integrates 5000+ monitoring points, with sampling frequency set according to hierarchy: equipment level (sensors, actuators) data is updated every second, unit level (main steam temperature control loop) data is aggregated every 30 seconds, and plant-wide level (economic load allocation) data is optimized every hour.

[0031] Input sources: Plant-wide dynamic model outputs, including equipment-level differential equations (turbine rotor vibration equations, state variable update cycle 0.1 seconds), unit-level algebraic coordination model (coupling relationship between boiler efficiency and turbine power), and plant-wide slow dynamic equations (hourly coal consumption rate model). Unified performance index framework: economic indicators (power supply coal consumption ≤ 295g / kWh), safety indicators (equipment health > 90%), and environmental indicators (NOx). x Emissions ≤50mg / m³ 3 The weights of each indicator are dynamically allocated through the framework (economic weight 0.6, safety weight 0.3, environmental weight 0.1).

[0032] Optimization objective: Solve the three-level coupled coordinated control problem within a 30-second cycle to optimize the overall plant performance while meeting load requirements (currently 600MW) and safety constraints.

[0033] The variational optimization algorithm decomposes the plant-wide optimization problem into three sub-problems for iterative solution by handling multi-timescale constraints. The specific process is as follows: Plant-wide hourly coordination instructions are generated, with an hourly cycle, to optimize the economic dispatch objective. Inputs include real-time power (average 598MW), coal consumption rate (302g / kWh), and grid load instructions uploaded from the unit level. Variational processing: The algorithm adjusts plant-wide variables (load allocation coefficients), calculates the functional extremum of hourly coal consumption rate and environmental constraints, and reduces the hourly coal consumption rate by 2% by coordinating boiler combustion and turbine steam intake, while controlling NO. x Emissions must not exceed limits. Output instructions: Plant-wide coordination instructions, total load setpoint (600MW), power allocation ratio for each unit (Unit A 70%, Unit B 30%). These instructions are updated every hour and sent to the unit level.

[0034] The unit-level minute-level setpoint command is generated on a minute-by-minute cycle, receiving commands from the entire plant level to optimize the unit control loop. Inputs include equipment-level second-level data (feedwater flow fluctuation values) and commands from the entire plant level. Variational processing: The algorithm solves the constrained extrema of the unit-level algebraic equations, minimizing the main steam pressure deviation (setpoint 16.5 MPa, actual value 16.3 MPa). By adjusting the boiler combustion rate setpoint and turbine valve opening, the unit efficiency is improved by 1.5%. Output commands: Unit-level setpoint commands, main steam temperature setpoint (541℃), and drum water level control target (0 mm). Commands are updated every 30 seconds and sent to the equipment level.

[0035] Equipment-level adjustment commands are generated at second-level intervals, executing unit-level commands and rapidly compensating for disturbances. Inputs include real-time sensor data (vibration amplitude, current signal). Variational processing: The algorithm, based on equipment-level differential equations, optimizes local variables (feed pump speed, damper opening) at second-level frequencies. When the vibration sensor detects an amplitude exceeding the limit (>50μm), the lubricating oil valve opening is instantaneously adjusted to restore the vibration value to normal within 2 seconds. Output commands: Equipment-level adjustment commands, actuator opening increment (+5%), and inverter frequency adjustment (-0.5Hz). Commands are updated every second and directly sent to field equipment via the DCS bus.

[0036] The coupling variables in variational optimization are as follows: the plant-wide output (load distribution ratio) serves as the boundary condition for unit-level optimization; the unit-level output (pressure setpoint) is converted into the tracking target for the equipment level. The coupling coefficient is dynamically calibrated using real-time data; when equipment health deteriorates, the safety weight automatically increases by 20%. Real-time solution performance: the server completes a full-cycle optimization calculation every 30 seconds with a latency of less than 1 second. The algorithm prioritizes urgent constraints (vibration exceeding limits), ensuring that safety response takes precedence over economic efficiency.

[0037] Output instructions (multi-level instruction products), the table below shows instruction examples within one optimization cycle (based on 600MW operating conditions): 103. State prediction module, used to predict the state of the dynamic model product of the whole plant using a multi-scale state predictor, calculate the health state vector product of the whole plant through differential feature extraction algorithm based on the prediction results, construct the whole plant-level Lyapunov function for stability monitoring, and generate the whole plant stability monitoring index product. It should be noted that a plant-wide health state vector and stability index are generated through multi-scale state prediction, differential feature extraction, and Lyapunov stability analysis.

[0038] Using the outputs of the plant-wide dynamic model (equipment level at the second level, unit level at the minute level, and plant-wide level at the hour level) as input, a multi-scale state predictor is constructed. This predictor employs a bidirectional gated recurrent neural network (BiGRU) structure to process dynamic data at different time scales: Equipment-level second-level prediction: Based on 100Hz sampling data, predicts the short-term state of key equipment. For turbine bearing vibration signals, it predicts the vibration amplitude change within the next 5 seconds, with the prediction error controlled within ±2μm (normal range 0-50μm). Input data includes real-time vibration sensor sequences (historical 10-second data points), and outputs a vibration prediction curve for the next 5 seconds. Unit-level minute-level prediction: Aggregates equipment-level output and predicts unit-level parameters. Using a 30-second cycle, based on feedwater flow and combustion rate data, it predicts the main steam temperature fluctuation for the next 2 minutes, with a prediction accuracy of ±1.5℃ (set value 541℃). Plant-wide hour-level prediction: Based on unit-level output, it predicts plant-wide economic indicators. Using historical hourly coal consumption data, it predicts the coal consumption rate for power supply for the next 4 hours, with an error not exceeding ±3g / kWh (baseline value 295g / kWh).

[0039] The predictor updates via a rolling time window, with the equipment layer refreshing the prediction results every second, and the unit layer and the plant-wide layer updating every 30 seconds and 1 hour respectively, ensuring real-time performance.

[0040] Based on the prediction results, a differential feature extraction algorithm is used to calculate the health state vector from multi-scale data. This algorithm focuses on the rate of change and trend characteristics of state variables: Equipment-level characteristics: Extracting the envelope derivative of the vibration signal amplitude (achieved through Hilbert transform), calculating the rate of change of vibration energy over time; if the vibration amplitude derivative exceeds 0.5 μm / s for three consecutive times, it is marked as an "accelerated wear" feature. Unit-level characteristics: Performing slope analysis on the main steam temperature prediction curve, calculating the mean and variance of temperature change per minute; when the mean slope exceeds 0.2℃ / min and the variance is higher than 0.05, it reflects a control loop response lag. Plant-wide characteristics: Analyzing the second derivative of the predicted coal consumption rate to identify the efficiency decay trend. A consistently positive coal consumption rate derivative with an absolute value greater than 0.1 g / kWh is considered a positive feature. 2 This indicates a decline in economic efficiency. The health status vector thus forms a five-dimensional array: equipment health (0-100, when the vibration derivative exceeds the limit, the equipment health drops from 95% to 82%; if the coal consumption derivative is abnormal, the energy efficiency attenuation rate is marked as 0.5%).

[0041] Based on the health state vector, a plant-wide Lyapunov function is constructed for stability monitoring. This function takes system state variables (power deviation, pressure fluctuation) as input and evaluates system stability through an energy function. Function design: A weighted function is calculated, using equipment-level vibration amplitude deviation, unit-level temperature setpoint tracking error, and plant-wide coal consumption rate deviation as state variables, calculating their sum of squares. The weights are dynamically adjusted based on real-time health status, with a vibration weight of 0.4 (increased to 0.6 for high-risk situations). Stability judgment: The function outputs stability indicators, including the sign of the "Lyapunov derivative" and the "stability margin." When the derivative is consistently negative and the stability margin is higher than 0.8 (range 0-1), the system is stable; if the derivative turns positive or the margin falls below 0.3, an early warning is triggered.

[0042] The final stability monitoring metrics include: Real-time stability margin: current value 0.75 (safety threshold > 0.5). Critical oscillation frequency: frequency of the dominant oscillation component at the equipment level (45Hz, normal range 40-50Hz). Risk level labels: "Stable", "Critical", "Unstable", and associated with DCS alarm levels.

[0043] 104. The report generation module is used to output the multi-level optimized control command products to the actuators at all levels of the DCS control system of the thermal power plant in real time. At the same time, based on the whole plant health status vector product and the whole plant stability monitoring index product, it generates the whole plant safety operation status assessment report product through a multi-dimensional fusion judgment algorithm. It should be noted that the core tasks of this step include two aspects: first, reliably distributing multi-level optimized control command outputs to the actuators at each level; and second, integrating the plant-wide health status vector and stability monitoring indicators to generate a comprehensive safety assessment report through multi-dimensional judgment algorithms. Implementation is based on the existing architecture of the DCS system (operator stations, controllers, network interfaces, etc.) and is coordinated and executed by the server.

[0044] The multi-level instructions (plant-wide coordination instructions, unit-level setpoint instructions, and equipment-level adjustment instructions) generated in step 102 are sent to the corresponding actuators through a dedicated DCS interface. The implementation process emphasizes the real-time performance, security, and hierarchical matching of the instructions: Instruction transmission path and protocol: Plant-wide coordination instructions (hourly level): Sent through the real-time historical data interface (compliant with T / CES218-2023 standard). Hourly coal consumption allocation instructions (unit A load ratio adjusted to 60%, unit B to 40%) are sent as data packets to the DCS real-time historical database, and then forwarded to the unit-level controller via the control command downlink interface. The instruction cycle is 1 hour, and the data transmission rate is ≥100Mbps to ensure seamless instruction switching in case of redundant network failure. Unit-level setpoint instructions (minute level): Sent directly through the controller inter-controller communication interface. The boiler main steam pressure setpoint (16.5MPa) is sent to the boiler coordination controller via the DCS internal communication protocol (ModbusTCP), with a cycle of 30 seconds and a communication delay <250ms, to avoid affecting the stability of the unit-level control loop. Equipment-level adjustment commands (second-level): The water pump speed adjustment command (+3%) is issued via I / O hardwiring or fieldbus and outputs directly to the field frequency converter as a 4–20mA analog signal or ProfibusDP message with a cycle of 1 second and a signal accuracy error of <0.5%. The command is continuously executed through the redundant DPU (distributed processing unit) of the DCS.

[0045] All commands undergo multiple checks before being issued: including command range checks (whether the pressure setpoint is within the safe range of 12–20 MPa), rate of change limits (load command fluctuation ≤ 3 MW per minute), and comparison with the actuator feedback status (an alarm is triggered when the valve opening command deviates from the actual position by more than 5%).

[0046] If the network is interrupted or the actuator fails, the DCS automatically activates the instruction retention policy: it retains the last valid instruction, marks the quality bit as "bad value", and notifies the server to recalculate the backup instruction.

[0047] The plant-wide health status vector product (equipment health, unit control stability score) and the plant-wide stability monitoring index product (Lyapunov stability margin) based on the 103-step process are used to generate an evaluation report through a multi-dimensional fusion judgment algorithm. The process is as follows: Multi-dimensional data fusion: Input data includes real-time health status vector (5 dimensions, equipment health 82%, unit control stability score 0.75), stability indicators (real-time stability margin 0.68, critical oscillation frequency 45Hz), and DCS real-time parameters (main steam temperature deviation ±2℃, vibration amplitude 45μm). Fusion rules: A combination of weighted scoring and logical judgment is used: Economic dimension (weight 0.3): For every 1g / kWh increase in the deviation of power supply coal consumption (295g / kWh) from the set value, 5 points are deducted. Safety dimension (weight 0.4): When equipment health <90% or stability margin <0.5, a "high risk" label is triggered; when vibration amplitude >50μm, an "equipment warning" is superimposed. Environmental dimension (weight 0.2): NO x Emission concentration (50mg / m³) 3 When the limit is exceeded, mark it as "Environmental protection deviation". Dynamic correlation analysis: If the stability margin decreases (0.68→0.55) and the boiler over-temperature event occurs simultaneously, the correlation is determined as "control loop response lag".

[0048] Assessment Report Output: Content Structure: The report includes a comprehensive score (0-100 points), risk level (normal / concern / warning / hazard), detailed diagnosis, and improvement suggestions. Comprehensive Score: 85 points (based on weighted calculation: economic efficiency 28 points + safety 38 points + environmental protection 19 points). Risk Level: "Concern" (due to stability margin of 0.55 approaching the threshold of 0.5). Diagnostic Details: Lists key issues, "Turbine vibration amplitude is approaching the limit (45μm / 50μm), it is recommended to check the bearing lubrication system"; "The main steam temperature control response delay of the unit layer is 2 minutes, and the PID parameters need to be optimized." Source Information: Linked to specific equipment (feed pump B) and control loop (boiler main control loop ID: MCS-01).

[0049] Generation cycle and real-time performance: The report is updated every 5 minutes, with a server processing delay of less than 1 second, ensuring real-time synchronization with the DCS monitoring interface (operator station).

[0050] Command issuance performance: Equipment-level command issuance success rate reaches 99.9%, unit-level command cycle error < ±0.5 seconds, and plant-wide command transmission packet loss rate < 0.1%. Assessment report accuracy: Based on historical data verification, diagnostic accuracy rate > 92%, false alarm rate < 3%, helping power plants reduce unplanned downtime by approximately 15%.

[0051] 105. Parameter adjustment module, used to display the plant-wide safety operation status assessment report product in real time on the DCS monitoring interface, and dynamically adjust the dynamic parameters and coupling coefficients in the plant-wide dynamic model product according to the assessment results, so as to realize the adaptive optimization of the plant-wide control system. It should be noted that in the DCS monitoring and diagnosis of a 600MW coal-fired power plant, the core objectives of step 105 include two aspects: first, to display the plant-wide safety operation status assessment report in real time on the DCS monitoring interface; and second, to dynamically adjust the parameters in the plant-wide dynamic model based on the assessment results to achieve adaptive optimization.

[0052] The plant-wide safety operation status assessment report (including comprehensive score, risk level, diagnostic details, etc.) generated through 104 steps is displayed to operators in real time through the DCS monitoring interface. The DCS system typically provides interfaces such as overview screen, trend curves, and alarm list for centralized management of critical information.

[0053] Display Content and Layout: Overall View: The main interface displays key plant parameters in a flowchart format. Real-time bar graphs or color blocks (green for normal, yellow for warning, red for danger) are embedded in the main interface, showing equipment health (turbine health 82%), stability margin (0.75), and overall score (85 points). A scrolling list in the corner displays diagnostic details such as "Feedwater pump vibration is too high; it is recommended to check bearing lubrication." Alarm Overview: When the assessment report identifies a risk (stability margin drops to 0.4, triggering a "critical" state), the DCS automatically pops up an alarm window, recording the event time, level, and description: "2023-10-01 14:05:00, unit-level main steam temperature control response delay, risk level: attention," accompanied by a voice prompt. Trend Curve Interface: Supports operators in retrieving historical data, comparing parameter changes, and displaying the equipment health curve over the past 8 hours (slowly decreasing from 95% to 82%), overlaid with the vibration amplitude curve to help analyze the attenuation trend. Real-time performance guarantee: Data update cycle is consistent with DCS sampling frequency (seconds at the device level, hours at the plant-wide level). Interface response time is less than 0.5 seconds, ensuring that operators can intervene immediately.

[0054] Based on the diagnostic results in the evaluation report (equipment performance degradation, control loop lag), the dynamic parameters and coupling coefficients in the plant-wide dynamic model output are automatically adjusted to achieve model self-correction and control system optimization. Dynamic parameter adjustment: Equipment layer example: When the evaluation report indicates abnormal turbine vibration characteristics (vibration derivative continuously exceeding limits), the model automatically increases the equipment layer time constant (from 0.5 seconds to 0.8 seconds) to smooth transient responses. Simultaneously, the weight of vibration data in the performance index framework is reduced (from 0.3 to 0.2), minimizing its impact on the overall score. Unit layer example: If the main steam temperature control response delay exceeds 2 minutes, the model dynamically corrects the lag time of the unit layer algebraic equations (reducing it from 2 minutes to 1.5 minutes) and optimizes the PID parameters (increasing the proportional coefficient by 10%) to improve the adjustment speed.

[0055] Coupling Coefficient Calibration: When the evaluation report shows a weakened correlation between the unit layer and the overall plant layer (the correlation coefficient between coal consumption rate and power is below 0.9), the model recalculates the coupling coefficient through regression analysis, increasing the power weight from the unit layer output to the overall plant layer from 0.95 to 0.98 to strengthen the economic correlation. For subsystems with frequent alarms (water supply system), the coupling strength between them and the global stability function is reduced (coefficient reduced by 15%) to prevent the spread of local disturbances. Optimization Effect Verification: After adjustment, the server compares the performance indicators (coal consumption rate deviation, stability margin) before and after the adjustment. If the indicators improve after adjustment (coal consumption rate reduced by 1.5%), the parameters are fixed; otherwise, it is rolled back to the historical stable version.

[0056] The table below uses one adaptive optimization cycle as an example (based on DCS interface operation): In this embodiment of the invention, a unified description framework for the performance indicators of the entire plant is constructed, which is compatible with dynamic behaviors at different time scales. This framework integrates data with various characteristics into a single system, facilitating comprehensive evaluation and analysis of the plant's performance. It also enables data exchange and comparison between different levels, providing a unified standard for coordinated control of the entire plant. A variational optimization algorithm is used to solve the coordinated control optimization problem of the entire plant in real time. The optimization problem is decomposed into three sub-problems for iterative solving, addressing the optimization needs at different time scales at the plant level, unit level, and equipment level. This approach efficiently handles constraints at multiple time scales, optimizing the overall performance indicators of the entire plant while meeting load requirements and safety constraints. This achieves rational allocation and efficient utilization of plant resources, improving the economic efficiency and safety of thermal power plant operation. A multi-scale state predictor is constructed using a bidirectional gated recurrent neural network structure to process dynamic data at different time scales. At the equipment level, it predicts the short-term state of key equipment at the second level; at the unit level, it predicts unit-level parameters at the minute level; and at the plant-wide level, it predicts the plant-wide economic indicators at the hour level. The prediction results are updated via a rolling time window to ensure real-time performance, providing forward-looking information for thermal power plant operation and enabling early detection of potential problems. The multi-level optimized control command output is output to the actuators at each level in real time through a dedicated DCS interface, emphasizing command real-time performance, security, and hierarchical matching. The command transmission path and protocol are designed according to the characteristics of different command levels, and undergo multiple checks before being issued. In the event of network interruption or actuator failure, the command retention strategy is automatically activated to ensure reliable command execution and guarantee the stable operation of the thermal power plant. Based on the evaluation results, the dynamic parameters and coupling coefficients in the dynamic model output of the entire plant are dynamically adjusted to achieve adaptive optimization of the entire plant's control system. For problems occurring at different levels, the model parameters and coupling relationships are automatically adjusted to better adapt the model to the actual operating conditions of the thermal power plant, improve the performance and stability of the control system, reduce manual intervention, and lower operating costs.

[0057] Please see Figure 2 , Figure 3 Another embodiment of the monitoring and diagnostic system for the DCS control system of a thermal power plant in this invention includes: 201. Framework construction module, used to establish a unified dynamic model of the whole plant based on a multi-timescale differential algebraic coordination system, to obtain a whole plant dynamic model product with a three-level coupled structure including equipment-level second-level dynamics, unit-level minute-level dynamics and whole plant-level hour-level dynamics, and to construct a unified description framework product of the whole plant performance indicators. Specifically, a second-level dynamic model is established at the equipment level to describe the rapid dynamic characteristics of major equipment such as boilers, steam turbines, and generators, generating the equipment-level dynamic model product; a minute-level dynamic model is established at the unit level to describe the coordinated operation characteristics between various process units, generating the unit-level dynamic model product; a plant-wide hour-level dynamic model is established to describe the changing characteristics of the overall plant performance and economic indicators, generating the plant-wide dynamic model product; based on the equipment-level, unit-level, and plant-wide dynamic model products, a coordination relationship between the three levels of models is established through a time-scale coupling mechanism, generating a unified plant-wide dynamic model product; based on the unified plant-wide dynamic model product, a plant-wide performance indicator system including thermodynamic performance indicators, mechanical performance indicators, electrical performance indicators, and environmental performance indicators is constructed, generating a unified description framework product for plant-wide performance indicators.

[0058] It should be noted that models describing the rapid dynamic characteristics of major equipment such as boilers, steam turbines, and generators are established. The boiler model focuses on describing the dynamic changes in pressure, temperature, and flow rate of the working fluid (water and steam) in the furnace combustion and steam-water system (steam drum, water-cooled walls, superheater, and reheater) on a time scale of seconds. The dynamic response of the steam-water separator and water-cooled walls (using a spiral tube and vertical tube arrangement) in ultra-supercritical once-through boilers is particularly critical. The steam turbine model focuses on the second-level response of steam flow rate, speed, and power when the control valve opening changes.

[0059] A unit-level minute-level dynamic model is established, taking process units (pulverizing system, steam-water system, and combustion system) as objects, to describe the coordinated operation characteristics between the equipment within the unit. The pulverizing unit model describes the minute-level dynamic process from the raw coal bunker, coal feeder, coal mill to the burner, including the impact of changes in pulverized coal fineness, conveying concentration, and air-coal ratio on boiler combustion stability. The unit-level model reflects the coupling relationships between equipment and the coupling relationship between feedwater flow rate and main steam temperature; its dynamic processes are typically on the minute level.

[0060] A dynamic hourly model of the entire plant was established to describe the changing characteristics of the plant's overall performance and economic indicators, including hourly statistical indicators or slowly varying parameters such as net efficiency, coal consumption for power generation, plant power consumption rate, and pollutant (sulfide) emission concentration. This model focuses on the comprehensive impact of equipment performance and unit coordination on the plant's overall economic and environmental performance over a longer time scale.

[0061] The time-scale coupling and integration with the plant-wide unified dynamic model integrates the three-layer model into a unified plant-wide dynamic model. The second-level dynamic output of the equipment layer (boiler main steam pressure and temperature) serves as the input boundary for the minute-level model of the unit layer. The operating status of the unit layer (pulverized coal quality) then affects the hour-level economic indicators (coal consumption) of the entire plant. This coupling ensures a coherent description from rapid equipment dynamics to slow plant-wide performance.

[0062] Based on the aforementioned unified dynamic model, a performance index system is constructed: Thermodynamic performance indicators: boiler efficiency, turbine heat rate, and overall plant thermal efficiency. Mechanical performance indicators: unit vibration characteristics and bearing temperature trends. Electrical performance indicators: generator output, power factor, and plant power consumption rate. Environmental performance indicators: flue gas desulfurization (FGD) system desulfurization efficiency and pollutant emission concentrations.

[0063] 202. Coordination and control module, which is used to solve the plant-wide coordination and control optimization problem in real time through variational optimization algorithm based on the plant-wide dynamic model product and the plant-wide performance index unified description framework product, and generate multi-level optimization control instruction products including equipment-level adjustment instructions, unit-level setpoint instructions and plant-wide coordination instructions. Specifically, based on the unified description framework of the plant's performance indicators, a comprehensive optimization objective function considering thermodynamic, mechanical, electrical, and environmental performance is constructed, generating an optimization problem description product. Based on the plant's dynamic model product, equipment operation constraints and system safety boundary conditions are extracted, generating a constraint set product. Based on the optimization problem description product and the constraint set product, the principle of variational method is applied to establish an optimization problem functional including state variables and control variables, generating a variational optimization problem product. Based on the variational optimization problem product, the optimal control trajectory of the system is obtained through the Euler-Lagrange equation solution method, generating a preliminary optimized control command product. Based on the preliminary optimized control command product and the constraint set product, the feasibility of the control commands is corrected and the coordination is verified, generating a multi-level optimized control command product.

[0064] Furthermore, the process of performing feasibility correction and coordination verification of control commands to generate multi-level optimized control command products also includes: performing dynamic constraint evaluation on the preliminary optimized control command products based on real-time operating condition data and safety boundary conditions to generate real-time constraint evaluation products; identifying the safety margin between control commands and operating boundaries based on the real-time constraint evaluation products to generate dynamic constraint adjustment strategy products; adaptively correcting control commands that exceed the safe operating boundary based on the dynamic constraint adjustment strategy products to generate safety boundary protection command products; updating the dynamic constraint parameters in the constraint set products based on the safety boundary protection command products to generate adaptive constraint set products; reconstructing the constraints of the optimization problem based on the adaptive constraint set products to generate reconstructed optimization problem products; and verifying the coordination of control commands based on the reconstructed optimization problem products to ensure the consistency between commands at each level and generate verified multi-level optimized control command products.

[0065] It should be noted that a comprehensive objective function, taking into account safety, economy, and environmental protection, is constructed based on a unified description framework for the plant's performance indicators. Specific indicators include: thermodynamic performance (coal consumption target for power generation ≤ 265 g / kWh), mechanical performance (turbine vibration amplitude < 75 μm), electrical performance (AGC load tracking error < ± 1.2 MW), and environmental performance (NOx). x Emission concentration <45mg / m³ 3 By weighting and integrating the above indicators, a mathematical expression is formed with the goal of "lowest total operating cost".

[0066] The operating boundaries are extracted from the unified dynamic model of the entire plant: Equipment level constraints: main steam temperature fluctuation range ≤565±5℃, steam drum water level deviation ±30mm; Unit level constraints: coal mill outlet air-coal ratio 1.8-2.2, fuel-water ratio accuracy ±0.05; Plant-wide constraints: peak load range 300-1000MW, AGC response rate ≥9MW / min. The variational method is applied to integrate the objective function and constraints into an optimization functional that includes state variables (main steam pressure, load) and control variables (coal feed rate, valve opening).

[0067] The optimal control trajectory is solved using the Euler-Lagrange equations, generating preliminary commands: Plant-wide: Distribute load commands to the unit level (current cycle target load 650MW); Unit level: Calculate the total boiler fuel consumption of 210t / h and the turbine control valve opening of 68%; Equipment level: Decompose the coal feed rate to 6 coal mills (35t / h / mill), and set the forced draft fan airflow to 850km³. 3 / h.

[0068] Dynamic constraint adjustment: Upon detecting an excessive fluctuation in the actual main steam pressure (16.8 MPa, exceeding the safety boundary of 17.0 MPa), a real-time constraint assessment is triggered. Using a dynamic constraint adjustment strategy, the coal feed rate command is reduced by 3%, and the turbine control valve opening is simultaneously adjusted to 65%, generating a safety boundary protection command. Coordination verification: The matching between the unit-level fuel-water ratio and the equipment-level feedwater flow command is checked. If the deviation is >5%, the optimization problem is reconstructed to correct the feedwater pump speed command, ensuring coordination among all levels of commands.

[0069] The resulting multi-level optimized control commands are distributed and executed by the DCS system, reducing the unit's coal consumption for power supply by 1.5 g / kWh and NOx under deep peak shaving conditions. x Emissions reduced by 3 mg / m³ 3 Furthermore, the main steam pressure fluctuation is controlled within ±0.3MPa.

[0070] 203. State prediction module, which is used to predict the state of the dynamic model product of the whole plant using a multi-scale state predictor, calculate the health state vector product of the whole plant through differential feature extraction algorithm based on the prediction results, construct the whole plant-level Lyapunov function for stability monitoring, and generate the whole plant stability monitoring index product. Specifically, based on the dynamic model product of the entire plant, a multi-timescale state predictor including second-level, minute-level, and hour-level predictors is constructed to generate multi-timescale state prediction data products; based on the multi-timescale state prediction data products, differential feature quantities of the state variables of each device are extracted to generate device differential feature set products; based on the device differential feature set products, health indicators of the thermal system, mechanical system, electrical system, and environmental protection system are calculated to generate plant-wide health state vector products; based on the multi-timescale state prediction data products, a Lyapunov function reflecting the overall plant operation stability is constructed to generate system stability judgment function products; based on the system stability judgment function products, system stability indicators and stability boundary distances are calculated in real time to generate plant-wide stability monitoring indicator products.

[0071] Furthermore, the process of generating plant-wide stability monitoring index products also includes: establishing an equipment state change trend early warning mechanism based on the equipment differential feature set product, and generating equipment state trend early warning products; constructing a system dynamic response characteristic analysis model based on multi-timescale state prediction data products, and generating dynamic response characteristic evaluation products; identifying potential system instability risk modes based on the equipment state trend early warning products and dynamic response characteristic evaluation products, and generating system instability risk early warning products; formulating graded early warning response strategies based on the system instability risk early warning products, and generating early warning response strategy products; generating system stability enhancement control suggestion products based on the early warning response strategy products and combined with the plant-wide health state vector products; and establishing an early warning effect verification mechanism, generating early warning effect evaluation products by comparing system stability changes before and after the early warning.

[0072] It should be noted that, based on a unified dynamic model for the entire plant, multi-scale predictors at the second, minute, and hourly levels are established. Second-level predictions (1-5 seconds) focus on rapid parameters such as boiler main steam pressure (target value 25.5±0.3MPa) and turbine speed (3000±2rpm); minute-level predictions (5-15 minutes) focus on unit coordination characteristics, air-coal ratio (set value 1.8-2.2), and enthalpy distribution of the steam-water system; hourly predictions (1-4 hours) assess the plant's total coal consumption for power generation (baseline value ≤265g / kWh) and NO₂. x Emission concentration (constraint value <45mg / m³) 3 Economic and environmental indicators, etc. The predictor is trained using historical data and builds a parameter change trajectory model based on the operating data of the past 30 days to predict the state of future time windows.

[0073] Differential characteristic quantities of key parameters extracted from the predicted data: Thermal system: main steam temperature change rate (normal range ±0.5℃ / s), boiler efficiency deviation (health threshold ≥94%); Mechanical system: turbine shaft vibration acceleration (warning value >2m / s²). 2The following parameters are considered: bearing temperature gradient (limit <3℃ / min); electrical system: generator active power fluctuation rate (normal <±0.8MW / s), excitation voltage stability. By weighted fusion of the characteristic quantities of each system, a plant-wide health status vector is generated: [thermal health 0.92, mechanical health 0.88, electrical health 0.95, environmental health 0.90], where the closer the value is to 1, the higher the health status.

[0074] A Lyapunov function based on the system energy function is constructed, with inputs including the deviations of state variables predicted at multiple scales (errors between power setpoints and actual values, and pressure fluctuation amplitude). System stability is determined by real-time calculation of the function derivative: if the derivative remains negative (stability index > 0.85), the system is stable; if it approaches zero or becomes positive (index ≤ 0.7), an instability warning is triggered. Simultaneously, the distance to the stability boundary and the normalized distance between the current operating point and the critical conditions (surge boundary, oscillation boundary) are calculated (threshold set at 0.2).

[0075] Trend Warning: When a sustained positive increase in the differential characteristic of the coal mill outlet temperature is detected (rate >0.1℃ / s for 5 consecutive minutes), a status trend warning is generated, indicating potential coal blockage. Instability Risk Identification: Combined with dynamic response analysis, if a lag in drum water level regulation is found during load abrupt changes (overshoot >15%), it is judged as a coordinated instability risk, triggering a graded warning (yellow warning). Closed-Loop Optimization: Based on the warning results, stability enhancement suggestions are generated, such as reducing the load change rate (from 9MW / min to 6MW / min) or adjusting the airflow lead control parameters. The effectiveness of the warning is verified by comparing the changes in stability indicators before and after the warning (overshoot reduced to 8%).

[0076] 204. The report generation module is used to output the multi-level optimized control command products to the actuators at all levels of the DCS control system of the thermal power plant in real time. At the same time, based on the whole plant health status vector product and the whole plant stability monitoring index product, it generates the whole plant safety operation status assessment report product through a multi-dimensional fusion judgment algorithm. Specifically, the multi-level optimized control command products are classified and distributed according to three levels: equipment level, unit level, and plant-wide level, generating hierarchical control command execution scheme products. Based on the plant-wide health status vector product and the plant-wide stability monitoring index product, a multi-dimensional judgment matrix including thermal health, mechanical health, electrical health, and system stability is constructed to generate a safety status judgment benchmark product. Based on the hierarchical control command execution scheme product and the safety status judgment benchmark product, a matching analysis between control commands and system status is performed to generate a command execution safety assessment product. Based on the command execution safety assessment product, combined with real-time operating data, a multi-dimensional fusion judgment algorithm is used to conduct a comprehensive plant-wide safety status rating to generate a safety status level assessment product. Based on the safety status level assessment product, a plant-wide safety operation status assessment report product including safety early warning suggestions, control optimization suggestions, and maintenance and repair suggestions is generated. An command execution feedback mechanism is established to feed back the actual execution effect of the hierarchical control command execution scheme product to the safety status assessment process, generating a dynamic early warning adjustment product.

[0077] Furthermore, the process of generating the plant-wide safety operation status assessment report product also includes: establishing a dynamic adjustment mechanism for safety boundaries based on real-time operation data and equipment status monitoring information to generate an adaptive safety boundary product; verifying the boundary compliance of the graded control instruction execution scheme product based on the adaptive safety boundary product to generate a boundary verification result product; conducting an operational risk probability assessment based on the boundary verification result product and the safety status level assessment product to generate a risk probability distribution product; formulating graded risk response strategies based on the risk probability distribution product to generate a risk prevention and control strategy product; establishing an assessment effect feedback mechanism to generate an assessment accuracy analysis product by comparing the assessment results with the actual operating status; and optimizing the parameter settings of the multi-dimensional fusion judgment algorithm based on the assessment accuracy analysis product to generate an optimized safety assessment model product.

[0078] It should be noted that the optimized control commands are distributed hierarchically: plant-wide coordination commands (load setpoint 650MW, power supply coal consumption target ≤265g / kWh); unit-level setpoint commands (total boiler fuel consumption 210t / h, air-coal ratio 1.8-2.2); and equipment-level adjustment commands (coal feeder speed deviation ±2%, turbine control valve opening 65%). Commands are transmitted to each actuator via the DCS network, and rate limits are set (load change rate ≤9MW / min) to prevent impact.

[0079] Based on the plant-wide health state vector [thermal health 0.92, mechanical health 0.88, electrical health 0.95, system stability 0.90] and stability indices (Lyapunov stability > 0.85), a judgment matrix is ​​constructed. Weights are determined using the analytic hierarchy process (AHP) to highlight key indicators.

[0080] An example of a safety status judgment matrix is ​​shown in the table below: A comprehensive security score is calculated using a multi-dimensional fusion algorithm. in Let i be the weight of the i-th indicator. This is the score for the i-th indicator. If the calculated score S≥90, it is rated as "Level A (Safe)", 80≤S<90 is rated as "Level B (Caution)", and S<80 is rated as "Level C (Risk)". Real-time data verifies the feasibility of the instruction. If the deviation between the coal feeding instruction and the coal mill current is found to be >5%, an instruction execution safety alarm is triggered.

[0081] Based on the safety rating (Level A in this example), the report generates the following safety warning: Safety Warning Recommendation: Main steam pressure fluctuations are approaching the upper limit (16.8 MPa / 17.0 MPa), indicating a need to monitor combustion stability. Control Optimization Recommendation: Fine-tune the air-fuel ratio by +0.1 to reduce NO. x Emissions (currently 45 mg / m³) 3 (Approaching the limit). Maintenance and inspection recommendations: The coal mill bearing temperature trend is abnormal (continuously rising by 0.3℃ / h), and it is recommended to check it within 72 hours.

[0082] Establish a dynamic adjustment mechanism for safety boundaries: When the unit load rises or falls rapidly, the main steam temperature fluctuation range is adaptively widened (temporarily adjusted from ±5℃ to ±8℃) to avoid frequent tripping. Based on the boundary verification results, the risk probability distribution is calculated (main steam overpressure risk probability <2%), and a graded response strategy is formulated (the load change rate is reduced when the probability >5%).

[0083] By comparing the deviation between the actual coal consumption for power generation (266g / kWh) and the target value (265g / kWh) after the command is executed, the fuel-water ratio coupling coefficient in the dynamic model of the whole plant is optimized, thereby improving the prediction accuracy to 95%.

[0084] 205. Parameter adjustment module, used to display the plant-wide safety operation status assessment report product in real time on the DCS monitoring interface, and dynamically adjust the dynamic parameters and coupling coefficients in the plant-wide dynamic model product according to the assessment results, so as to realize the adaptive optimization of the plant-wide control system. Specifically, a multi-level visualization display area is configured in the DCS monitoring interface to display key information from the plant-wide safety operation status assessment report in real time, generating a visualization monitoring interface product. Based on the plant-wide safety operation status assessment report product, system performance deviation data and abnormal operation modes are extracted to generate a model parameter adjustment requirement product. Based on the model parameter adjustment requirement product, combined with real-time operating condition data, a dynamic parameter optimization strategy is formulated to generate a parameter adaptive adjustment scheme product. Based on the parameter adaptive adjustment scheme product, the equipment dynamic parameters, unit coupling coefficients, and plant-wide coordination parameters in the plant-wide dynamic model product are corrected online to generate an optimized plant-wide dynamic model product. A closed-loop verification mechanism is established to generate a model optimization effect evaluation product by comparing the prediction accuracy of the optimized plant-wide dynamic model product with actual operating data. Based on the model optimization effect evaluation product, the parameter adjustment rule base is dynamically updated to generate an adaptive optimization knowledge base product.

[0085] It should be noted that the DCS monitoring interface for the 1000MW ultra-supercritical unit has undergone adaptive optimization. The DCS monitoring interface is divided into three dedicated visualization areas: Plant-wide overview: centrally displayed key indicators such as coal consumption for power supply (265.3 g / kWh), load rate (92%), and plant-wide safety level (Level A). Unit-level diagnostic area: the left-hand column displays unit health status information such as boiler efficiency (94.2%), turbine vibration (72 μm), and desulfurization efficiency (96.5%). Equipment-level early warning area: the right-hand scrolling alerts provide real-time warnings such as abnormal coal mill bearing temperature trends (0.3℃ / h increase) and excessive steam drum water level fluctuations (±28 mm). All data is updated every second via a time-series database and uses red / yellow / green color coding to distinguish the urgency of the status.

[0086] Model parameter tuning requirement extraction and strategy generation: Based on the security assessment report, performance deviations are identified, and parameter optimization requirements are automatically generated. Examples of model parameter tuning strategies are shown in the table below. Parameter correction: Based on the adjustment strategy, the fuel-water ratio coupling coefficient in the equipment layer dynamic model is updated online (0.85→0.83), and the pressure-coal feedforward weight in the unit layer coordination model is optimized simultaneously (+15%).

[0087] Closed-loop verification: Comparing the model predictions with actual operating data after parameter adjustment, the main steam pressure prediction error was reduced from ±1.2MPa to ±0.5MPa after correction, and the boiler efficiency prediction accuracy was improved to 96.5%.

[0088] Establish a rule base for tracking the effect of parameter adjustment: When the load change rate is >8MW / min, the fast response mode is automatically activated, and the fluctuation range of the steam drum water level is temporarily relaxed to ±35mm; if the model accuracy is improved by >3% after 5 cumulative corrections, the parameter adjustment strategy is marked as a high-priority recommended scheme.

[0089] 206. The Coordination and Optimization Module is used to establish a plant-wide coordination and optimization performance evaluation system. Based on the plant-wide safety operation status assessment report and actual operation data, it calculates the overall system coordination performance indicators and generates a plant-wide coordination performance evaluation product. Based on the plant-wide coordination performance evaluation product, it identifies coordination bottlenecks and optimization potential areas among subsystems and generates a coordination optimization key area identification product. Based on the coordination optimization key area identification product, it formulates targeted coordination strategy adjustment schemes and generates a hierarchical coordination optimization strategy product. Based on the hierarchical coordination optimization strategy product, it refines the coordination parameters at the equipment level, unit level, and plant-wide level and generates a coordination parameter optimization scheme product. It establishes a coordination optimization effect tracking mechanism, and generates a coordination optimization effect verification product by comparing the system performance changes before and after the coordination parameter adjustment. Based on the coordination optimization effect verification product, it updates the coordination optimization knowledge base and strategy rules, forms a closed-loop optimization learning mechanism, and generates an adaptive coordination optimization knowledge base product.

[0090] It should be noted that a comprehensive coordination performance index is constructed based on the safety assessment report and actual operating data. This index (C) can be quantified as the unit's agility in responding to grid dispatch commands (load response rate ≥ 9 MW / min), main steam pressure stability (fluctuation range ± 0.3 MPa), coal consumption for power generation (265 g / kWh), and NOx emission concentration (≤ 45 mg / m³). 3 The weighted average performance of () is expressed by the following quantitative formula: in, to The stress stability score is calculated based on the weights assigned according to the operational objectives. .

[0091] By analyzing the coordination performance assessment results, key bottlenecks affecting overall performance were identified. It was found that when the unit experienced load increases and decreases within the 500-700MW load range, the boiler main control and turbine main control commands were out of sync, resulting in significant fluctuations in main steam pressure (up to ±0.8MPa) and an increase in power generation coal consumption of approximately 1.5g / kWh. The table below shows a simplified assessment analysis: An example of coordination performance bottleneck identification is shown in the table below: Based on this, a hierarchical coordination and optimization strategy was formulated: the distribution curve of AGC load commands was adjusted at the plant level; the feedforward coefficient of the boiler main control was optimized at the unit level (increased by 10%); and the positioning control parameters of the coal feeder and damper were finely adjusted at the equipment level.

[0092] Parameter adjustments were implemented for the identified key areas: In the 500-700MW load range, the feedforward coefficient of the boiler main control system to the load change rate was increased from 0.85 to 0.92; the air-coal ratio curve of the combustion system was fine-tuned, optimizing the air-coal ratio from 1.78 to 1.82 at the 580MW load point. An effect tracking mechanism was established, comparing data from one week before and after the adjustment: the standard deviation of main steam pressure fluctuation decreased from 0.52MPa to 0.35MPa, the load response rate increased to 8.8MW / min, and the coal consumption for power generation decreased by 0.7g / kWh.

[0093] The parameter adjustment scheme (580MW load point, wind-to-coal ratio 1.82) and its resulting performance improvement (coal consumption reduction of 0.7g / kWh) are stored as a successful case in the adaptive coordinated optimization knowledge base. The system learns such optimization rules, and when similar operating conditions (similar load range, similar coal characteristics) are detected again, it can automatically recommend or apply the verified parameter adjustment strategy. By continuously comparing the changes in performance indicators before and after optimization, the knowledge base is constantly updated, enabling the unit's coordinated control performance to continuously improve itself in a cyclical manner.

[0094] In this embodiment of the invention, by establishing a unified dynamic model for the entire plant and multi-level optimized control commands, coordinated control at the equipment level, unit level, and plant level is achieved, effectively solving the coordination bottleneck problem between various subsystems. After adjusting the boiler main control feedforward coefficient and air-coal ratio curve in a specific load range, the standard deviation of main steam pressure fluctuation is significantly reduced, the load response rate is improved, the coal consumption for power supply is reduced, and the unit's responsiveness to grid dispatch commands and overall operational coordination are improved. The multi-scale state prediction and stability monitoring functions can monitor the operating status of equipment and the entire plant in real time and provide early warning of potential risks. By constructing a Lyapunov function to determine system stability and establishing an early warning mechanism based on the differential characteristic set of equipment, the system can promptly identify system instability risk modes and formulate graded early warning response strategies, effectively preventing accidents and ensuring the safe and stable operation of the system. A comprehensive optimization objective function considering thermodynamic, mechanical, electrical, and environmental performance enables the unit to achieve both economic and environmental optimization while ensuring safe operation. A closed-loop adaptive optimization mechanism allows the system to dynamically adjust model parameters and coordination strategies based on real-time operating data and safety assessment results, and continuously update the optimization knowledge base. The DCS monitoring interface displays key information from the plant-wide safety operation status assessment report in real time. Through multi-level visualization display areas, operators can intuitively understand the operating status of the entire plant, units, and equipment, promptly identify anomalies, and take corresponding measures, improving monitoring efficiency and management level, and providing strong support for the safe, economical, and environmentally friendly operation of thermal power plants.

[0095] Figure 4 This is a schematic diagram of the structure of a monitoring and diagnostic device for a DCS control system in a thermal power plant, provided by an embodiment of the present invention. The monitoring and diagnostic device 300 for the DCS control system in a thermal power plant can vary considerably due to differences in configuration or performance. The device 300 includes a transmitter 301, a receiver 302, and a processor 303. The processor 303 can also be a controller. Figure 4 The device is designated as "controller / processor 303". Optionally, the device 300 may also include a modem processor 305, which may include an encoder 306, a modulator 307, a decoder 308, and a demodulator 309.

[0096] In one example, transmitter 301 modulates (e.g., analog-to-analog conversion, filtering, amplification, and up-conversion, etc.) the output sample and generates an uplink signal, which is transmitted via an antenna to an access network device. On the downlink, the antenna receives the downlink signal transmitted by the access network device. Receiver 302 modulates (e.g., filtering, amplification, down-conversion, and digitization, etc.) the signal received from the antenna and provides an input sample. In modem processor 305, encoder 306 receives service data and signaling messages to be transmitted on the uplink and processes (e.g., formatting, encoding, and interleaving) the service data and signaling messages. Modulator 307 further processes (e.g., symbol mapping and modulation) the encoded service data and signaling messages and provides an output sample. Demodulator 309 processes (e.g., demodulates) the input sample and provides a symbol estimate. Decoder 308 processes (e.g., deinterleaving and decoding) the symbol estimate and provides decoded data and signaling messages to device 300. Encoder 306, modulator 307, demodulator 309, and decoder 308 can be implemented by a combined modem processor 305. These units perform processing according to the radio access technology adopted by the radio access network (e.g., LTE and other evolved systems access technologies). It should be noted that when device 300 does not include modem processor 305, the above-mentioned functions of modem processor 305 can also be performed by processor 303.

[0097] The processor 303 controls and manages the operation of the device 300, and is used to execute the processing procedures performed by the device 300 in the above embodiments of this disclosure. For example, the processor 303 is also used to execute various steps of the transmitting or receiving device in the above method embodiments, and / or other steps of the technical solutions described in the embodiments of this disclosure.

[0098] Furthermore, the device 300 may also include a memory 304 for storing program code and data for the device 300.

[0099] Understandable Figure 4 Only a simplified design of device 300 is shown. In practical applications, device 300 can include any number of transmitters, receivers, processors, modem processors, memory, etc., and all devices that can implement the embodiments of this disclosure are within the protection scope of the embodiments of this disclosure.

[0100] The present invention also provides a monitoring and diagnostic device for a DCS control system of a thermal power plant. The monitoring and diagnostic device for a DCS control system of a thermal power plant includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the monitoring and diagnostic system for the DCS control system of the thermal power plant in the above embodiments.

[0101] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the monitoring and diagnostic system of the DCS control system of the thermal power plant.

[0102] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0103] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0104] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A monitoring and diagnostic system for a DCS control system of a thermal power plant, characterized in that, The method comprises the following steps: a framework modeling module is used to establish a unified dynamic model of the whole plant, obtain a dynamic model product of the whole plant, and construct a unified description framework product of performance indexes of the whole plant; a coordinated control module is used to solve a coordinated control optimization problem of the whole plant in real time according to the dynamic model product of the whole plant and the unified description framework product of performance indexes of the whole plant, and generate a multi-level optimization control instruction product; a state prediction module is used to perform state prediction on the dynamic model product of the whole plant, calculate a health state vector product of the whole plant based on a prediction result, construct a Lyapunov function of the whole plant for stability monitoring, and generate a stability monitoring index product of the whole plant; a report generation module is used to output the multi-level optimization control instruction product for execution in real time, and generate a safe operation state evaluation report product of the whole plant based on the health state vector product of the whole plant and the stability monitoring index product of the whole plant; a parameter adjustment module is used to display the safe operation state evaluation report product of the whole plant on a DCS monitoring interface in real time, and dynamically adjust dynamic parameters and coupling coefficients in the dynamic model product of the whole plant according to an evaluation result.

2. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 1, characterized in that, According to the unified description framework product of performance indexes of the whole plant, a comprehensive optimization objective function is constructed, and an optimization problem description product is generated; Based on the dynamic model product of the whole plant, equipment operation constraint conditions and system safety boundary conditions are extracted, and a constraint condition set product is generated; According to the optimization problem description product and the constraint condition set product, an optimization problem functional is established, and a variational optimization problem product is generated; Based on the variational optimization problem product, an optimal control trajectory of the system is obtained, and a preliminary optimization control instruction product is generated; Based on the preliminary optimization control instruction product and the constraint condition set product, a multi-level optimization control instruction product is generated.

3. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 1, characterized in that, According to the dynamic model product of the whole plant, a multi-time scale state predictor is constructed, and a multi-time scale state prediction data product is generated; According to the multi-time scale state prediction data product, differential characteristic quantities of state variables of each equipment are extracted, and an equipment differential characteristic set product is generated; Based on the equipment differential characteristic set product, a health degree index is calculated, and a health state vector product of the whole plant is generated; Based on the multi-time scale state prediction data product, a Lyapunov function reflecting the running stability of the whole plant is constructed, and a system stability judgment function product is generated; Based on the system stability judgment function product, a system stability degree index and a stability boundary distance are calculated in real time, and a stability monitoring index product of the whole plant is generated.

4. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 3, characterized in that, In the process of generating the stability monitoring index product of the whole plant, the following steps are further included: Based on the equipment differential characteristic set product, an equipment state change trend early warning mechanism is established, and an equipment state trend early warning product is generated; According to the multi-time scale state prediction data product, a system dynamic response characteristic analysis model is constructed, and a dynamic response characteristic evaluation product is generated; Based on the equipment state trend early warning product and the dynamic response characteristic evaluation product, a potential system instability risk mode is identified, and a system instability risk early warning product is generated; According to the system instability risk early warning product, a hierarchical early warning response strategy is formulated, and an early warning response strategy product is generated; Based on the early warning response strategy product, combined with the plant health state vector product, a system stability enhancement control suggestion product is generated; An early warning effect verification mechanism is established, and by comparing the system stability changes before and after early warning, an early warning effect evaluation product is generated.

5. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 1, characterized in that, The multi-level optimization control instruction product is classified and distributed according to the device level, unit level and plant level three levels, and a hierarchical control instruction execution scheme product is generated; Based on the plant health state vector product and the plant stability monitoring index product, a multi-dimensional judgment matrix is constructed, and a safety state judgment benchmark product is generated; According to the hierarchical control instruction execution scheme product and the safety state judgment benchmark product, the matching analysis of control instruction and system state is carried out, and an instruction execution safety evaluation product is generated; According to the instruction execution safety evaluation product, combined with real-time operation data, a comprehensive safety state rating of the plant is carried out, and a safety state level evaluation product is generated; Based on the safety state level evaluation product, a plant safety operation state evaluation report product is generated.

6. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 5, characterized in that, A comprehensive safety score is calculated by a multi-dimensional fusion algorithm , obtaining a safety state rating product: wherein, is the weight of the i-th indicator, is the score of the i-th indicator, n is the total number of safety indicators, i is the index of the indicator, representing the i-th specific safety assessment indicator.

7. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 5, characterized in that, In the process of generating the plant safety operation state evaluation report product, it also includes: Based on real-time operation data and equipment state monitoring information, a safety boundary dynamic adjustment mechanism is established, and a self-adaptive safety boundary product is generated; According to the self-adaptive safety boundary product, the boundary compliance verification of the hierarchical control instruction execution scheme product is carried out, and a boundary verification result product is generated; Based on the boundary verification result product, combined with the safety state level evaluation product, the operation risk probability is evaluated, and a risk probability distribution product is generated; According to the risk probability distribution product, a hierarchical risk response strategy is developed, and a risk prevention and control strategy product is generated; An evaluation effect feedback mechanism is established, and by comparing the evaluation results with the actual operation state, an evaluation accuracy analysis product is generated; Based on the evaluation accuracy analysis product, the parameter settings of the multi-dimensional fusion judgment algorithm are optimized, and an optimized safety evaluation model product is generated.

8. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 1, characterized in that, A multi-level visual display area is configured to display the key information of the plant safety operation state evaluation report product in real time, and a visual monitoring interface product is generated; Based on the plant safety operation state evaluation report product, system performance deviation data and operation abnormal mode are extracted, and a model parameter adjustment demand product is generated; According to the model parameter adjustment demand product, combined with real-time operation condition data, a dynamic parameter optimization strategy is developed, and a parameter self-adaptive adjustment scheme product is generated; According to the parameter self-adaptive adjustment scheme product, the plant dynamic model product is corrected online, and an optimized plant dynamic model product is generated; By comparing the prediction accuracy of the optimized plant dynamic model product with the actual operation data, a model optimization effect evaluation product is generated; Based on the model optimization effect evaluation product, the parameter adjustment rule library is dynamically updated, and a self-adaptive optimization knowledge base product is generated.

9. The monitoring and diagnostic system of a thermal power plant DCS control system according to claim 1, characterized in that, It also includes a coordination optimization module: Based on the plant safety operation state evaluation report product and the actual operation data, the system overall coordination performance index is calculated, and a plant coordination performance evaluation product is generated; According to the whole plant coordination performance evaluation product, a coordination bottleneck and an optimization potential area between subsystems are identified, and a coordination optimization key area identification product is generated; A hierarchical coordination optimization strategy product is generated based on the coordination optimization key area identification product; A coordination parameter optimization scheme product is generated according to the hierarchical coordination optimization strategy product; A coordination optimization effect tracking mechanism is established, and a coordination optimization effect verification product is generated by comparing the system performance changes before and after the coordination parameter adjustment; According to the coordination optimization effect verification product, the coordination optimization knowledge base and strategy rules are updated, and a self-adaptive coordination optimization knowledge base product is generated.

Images