A power plant boiler fault deduction method, device and system

Abstract

The application provides a power plant boiler fault deduction method, device and system, the method comprises the following steps: constructing an operation architecture, and constructing a first simulation operation model and a second simulation operation model of the power plant boiler; synchronously performing online correction on the simulation operation model, and monitoring the output parameters of the first simulation operation model; under abnormal conditions, calling the second simulation operation model to perform accelerated simulation, and obtaining a predicted accident result by combining multi-layer accident chain analysis. Through the parallel double simulation model, the application solves the operation resource conflict problem of real-time working condition simulation and fault deduction, considers the accuracy of real-time monitoring and the timeliness of fault deduction, combines synchronous online correction, and improves the reliability of the fault deduction result; through calling the independent model to perform accelerated simulation, the fault early warning reference can be provided in time, and through combining the multi-layer accident chain analysis, the modeling specification of accident deduction is unified, and the prediction accuracy of the fault spreading path and the final consequence is improved.

Description

Technical Field

[0001] This application relates to the field of thermal power technology, specifically to a method, apparatus and system for simulating boiler failures in thermal power plants. Background Technology

[0002] Boilers in thermal power plants are crucial energy conversion equipment in power production systems, and their operational stability directly affects the safe production and power supply reliability of the entire power plant. Boiler systems have complex structures and variable operating conditions, making them prone to various faults such as water-cooled wall tube rupture, flue gas imbalance, and abnormal combustion. Moreover, these faults spread rapidly and have a wide impact, and failure to handle them promptly can lead to significant economic losses and safety accidents.

[0003] Currently, digital twin technology has been introduced for high-precision simulation in the safety analysis of thermal power plant boilers. This involves constructing a virtual model and a real-time mapping between the virtual model and the physical equipment, which to a certain extent enables automated and high-precision boiler operation monitoring and fault analysis.

[0004] However, existing digital twin-based boiler fault simulation methods still have shortcomings in practical applications, especially in achieving real-time operating condition simulation and rapid fault simulation. They are prone to computational resource conflicts and data interaction interference, making it difficult to balance the accuracy of real-time monitoring with the timeliness of fault simulation. Furthermore, as the equipment's service life increases, the deviation between the simulation model and the actual equipment gradually widens, leading to a decrease in the reliability of the fault simulation results.

[0005] Furthermore, existing accident simulation modeling methods lack unified standards, with significant differences in modeling logic across different fault scenarios. The quantitative analysis capabilities for fault development processes are insufficient, making it difficult to accurately predict the path of fault propagation and the final consequences. The triggering mechanism for fault simulations is not flexible enough, failing to automatically initiate the simulation process based on abnormal changes in equipment operating status, thus hindering the provision of timely fault warnings and handling references for operators.

[0006] In summary, existing digital twin fault simulation technologies for thermal power plant boilers suffer from technical problems such as difficulty in balancing real-time performance and timeliness, low model matching degree, and poor simulation standardization. Summary of the Invention

[0007] This application addresses the problems existing in the prior art by providing a method, apparatus, and system for simulating boiler failures in thermal power plants, thereby resolving the technical issues.

[0008] To achieve the above objectives, the technical solution adopted in this application is as follows: On the one hand, this application provides a method for predicting boiler failures in thermal power plants, which includes: Construct a runtime architecture with logical layering and IP segmentation isolation; Based on the aforementioned operational architecture, a first simulation operation model and a second simulation operation model for a thermal power plant boiler are constructed using digital twin technology; the first simulation operation model and the second simulation operation model operate in parallel. Based on the real-time operating data of the thermal power plant boiler, the first simulation operation model and the second simulation operation model are simultaneously corrected online. Monitor the output parameters of the first simulation running model. If the output parameters of the first simulation running model meet the preset abnormal conditions, call the second simulation running model to perform accelerated simulation and obtain the accelerated simulation results. The accelerated simulation results are analyzed by using a pre-defined multi-layered accident chain to obtain predicted accident results.

[0009] Optionally, constructing the runtime architecture includes: A logical layering is constructed, which includes, from bottom to top, a field layer, a simulation layer, and an interface layer; The simulation layer is divided into two independent IP network segments, which are respectively bound to the first simulation running model and the second simulation running model.

[0010] Optionally, in the accelerated simulation, the first simulation running model runs with a first simulation step size, and the second simulation running model runs with a second simulation step size; The second simulation step size is greater than the first simulation step size.

[0011] Optionally, before calling the second simulation running model to perform accelerated simulation, the operating condition data of the first simulation running model is saved; After the second simulation running model completes the accelerated simulation, the first simulation running model is restored to operation based on the operating condition data.

[0012] Optionally, the preset multi-layered accident chain includes, in sequence, a fault injection layer, a multi-field coupling layer, a standardized hazard layer, a chain action layer, and a final result layer.

[0013] Optionally, the pre-defined multi-layered incident chain construction includes: Deploy parameterized fault source units in the fault injection layer; At least one of logic condition units, physical quantity operation units, and discrete event units is deployed in the multi-field coupling layer; The standardized hazard layer is deployed with at least one of a logic condition unit, a physical quantity calculation unit, and a discrete event unit; the standardized hazard layer is also pre-encapsulated with one of the standardized hazard modes, or a combination of multiple standardized hazard modes, namely, abnormal water level, abnormal temperature, abnormal water pressure, abnormal furnace pressure, abnormal structural damage, or abnormal combustion. Each of the chain action layers is equipped with at least one of the following: a logic condition unit, a physical quantity calculation unit, and a discrete event unit. The discrete event unit is deployed to the final result layer.

[0014] On the other hand, this application also provides a boiler fault simulation device for thermal power plants, which includes a simulation calculation module, a parameter correction module, a monitoring module and an accident analysis module; The simulation computing module is used to run the first simulation running model and the second simulation running model in parallel on the running architecture; The parameter correction module is used to acquire real-time operating data of the boiler in the thermal power plant and simultaneously correct the first simulation operation model and the second simulation operation model online. The monitoring module is used to monitor the output parameters of the first simulation running model. When the output parameters of the first simulation running model meet the preset abnormal conditions, the second simulation running model is called to perform accelerated simulation and obtain the accelerated simulation results. The accident analysis module is used to analyze the accelerated simulation results through a preset multi-layer accident chain and obtain predicted accident results.

[0015] Optionally, it also includes a working condition caching module, which is used to save the working condition data of the first simulation running model before the accelerated simulation starts; The operating condition data is used to restore the operation of the first simulation running model after the accelerated simulation is completed.

[0016] Furthermore, this application also provides a boiler fault prediction system for thermal power plants, which includes a data acquisition unit, a human-machine interface, and the aforementioned boiler fault prediction device for thermal power plants. The data acquisition unit is used to collect real-time operating data of the boiler in the thermal power plant and send the real-time operating data to the parameter correction module of the boiler fault inference device in the thermal power plant. The human-computer interaction interface is used to display the predicted accident results.

[0017] Optionally, the human-computer interaction interface is also used to receive manual deduction instructions input by the user and send the manual deduction instructions to the thermal power plant boiler fault deduction device. The manual simulation command is used to trigger the second simulation running model to perform accelerated simulation under specified working conditions.

[0018] Compared with the prior art, this application has the following advantages: This application resolves the computational resource conflict between real-time operational simulation and fault prediction by using an isolated operating architecture and running dual simulation models in parallel, thus balancing the accuracy of real-time monitoring with the timeliness of fault prediction. Furthermore, by combining real-time operational data with synchronous online correction, it ensures the dynamic matching degree between the simulation model and the physical equipment, improving the credibility of fault prediction results. By automatically calling the second simulation model for accelerated simulation when abnormal output parameters are detected, automatic triggering of fault prediction is achieved, providing timely fault warning references for operators. Finally, through the pre-set multi-layer accident chain analysis, the modeling specifications for accident prediction are unified, improving the prediction accuracy of fault propagation paths and final consequences. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the method described in this application; Figure 2 This is a block diagram of the device in this application; Figure 3 This is a block diagram of the simulation calculation module of this application; Figure 4 This is a schematic diagram of a multi-layered accident chain in a specific embodiment of this application. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] It should be noted that, for ease of understanding, the method steps in the specific embodiments of this application are described in a certain order, but those skilled in the art can change the order of the steps according to actual needs, so this should not be used as a limiting condition.

[0024] like Figure 1 As shown, this application provides a method for simulating boiler failures in thermal power plants, the main steps of which include: First, construct a runtime architecture with logical layering and IP segmentation isolation.

[0025] The logical layering, from bottom to top, includes the field layer, simulation layer, and interface layer. The simulation layer is divided into two independent IP network segments, designated as the first and second segments, corresponding to the first and second simulation operation models. The field layer shares the same IP segment as data acquisition and is responsible for communicating with data acquisition-related functional modules and reading acquired data. The simulation layer is responsible for separating the computational activities of the first and second simulation operation models. The interface layer outputs computational results and interfaces with memory and parameter correction modules.

[0026] In terms of operational architecture, a first and second simulation operation model of the thermal power plant boiler are constructed based on digital twin technology, and these two models operate in parallel. Both the first and second simulation operation models include a controlled object model and a controller model. Furthermore, the controlled object model consists of mathematical model parameters describing the physical characteristics of the thermal power plant boiler, including thermal inertia coefficient, drag coefficient, heat transfer coefficient, volume coefficient, etc., with units determined according to the type of physical quantity and calibrated through on-site measurements or equipment factory parameters. The controller model consists of parameters describing the logic of the boiler control system, including the proportional gain, integral time, and derivative time of the PID controller, as well as the action threshold and delay time of interlocking protection.

[0027] The first and second simulation models can perform two types of simulations. One involves constructing initial operating conditions based on typical faults in the fault database, using the controlled object and controller variables listed in the I / O point table. The variables in the I / O point table include those related to the interaction between the controlled object and the controller, as well as setpoints and controller parameters. The other approach involves both models receiving real-time operating data to achieve high-precision tracking of the physical system's operating conditions.

[0028] The above simulations are based on initial operating conditions and physical information. These two modes can be switched through the user interface.

[0029] Based on real-time operating data of thermal power plant boilers, the first and second simulation operating models are simultaneously corrected online; real-time operating data input is received, and the parameters of the controlled object model and controller model are corrected. At the same time, the operator's operating instructions to the equipment are synchronized, so that the same operation is performed in the simulation model.

[0030] During the normal operation of the first and second simulation models, the output parameters of the first simulation model are monitored. Specifically, the monitoring employs one or more of the following methods: parameter exceeding threshold alarm, continuous time exceeding threshold alarm, or preset time point exceeding threshold alarm. The preset alarm information format includes the model, variable name, threshold, alarm judgment method, and non-real-time simulation time. Continuous time exceeding threshold refers to the number of simulation cycles in which the parameter exceeds the threshold before triggering an alarm; a typical value is 10 cycles, used to avoid false alarms caused by momentary interference.

[0031] If the output parameters of the first simulation model meet the preset abnormal conditions, the second simulation model is invoked for accelerated simulation, and the accelerated simulation results are obtained. The preset abnormal conditions are parameter threshold conditions that trigger the accelerated simulation, determined by the thermal power plant operating procedures and equipment safety specifications, including upper and lower limits for process parameters such as temperature, pressure, flow rate, and liquid level.

[0032] Before invoking the second simulation model for accelerated simulation, the operating condition data of the first simulation model is saved. After the accelerated simulation of the second simulation model is completed, the first simulation model is restored based on the operating condition data. The scheduler is responsible for periodically scheduling the controlled object model and the controller model. Accelerated simulation is achieved by scheduling the model at different steps within the same physical time. The time calculation module provides the scheduler with the calculation relationship between physical time and simulation time. Therefore, during the accelerated simulation process, the first simulation model runs at the first simulation step size, and the second simulation model runs at the second simulation step size, with the second simulation step size being larger than the first simulation step size.

[0033] The accelerated simulation results are analyzed and predicted accident outcomes are obtained by using a pre-defined multi-layered accident chain. The pre-defined multi-layered accident chain consists of a fault injection layer, a multi-field coupling layer, a normalized hazard layer, a chain action layer, and a final result layer. The pre-defined multi-layered accident chain is constructed by custom combination of four basic units: parameterized fault source unit, logic condition unit, physical quantity calculation unit, and discrete event unit. The parameters and calculation results are uniformly expressed as several basic physical quantities such as length, mass, time, current, temperature, pressure, flow rate, and percentage data.

[0034] The fault injection layer deploys only parameterized fault source units, and the final result layer deploys only discrete event units. The multi-field coupling layer, normalized hazard layer, and interlocking action layer each deploy at least one of the following: logic condition units, physical quantity calculation units, and discrete event units. Specifically, the normalized hazard layer pre-encapsulates one or more normalized hazard modes, such as abnormal water level, abnormal temperature, abnormal water pressure, abnormal furnace pressure, structural damage, or abnormal combustion. The multi-field coupling layer describes the relationship between the source fault and other operating conditions or physical parameters and is user-defined. The normalized hazard layer has a preset fixed classification and cannot be user-defined. The interlocking action layer is related to the control and protection equipment equipped in the power plant and can be user-defined.

[0035] On the other hand, this application as Figure 2 and Figure 3 As shown, a fault simulation device for thermal power plant boilers is also provided, including a simulation calculation module, a parameter correction module, a monitoring module, an accident analysis module, an operating condition generation module, a simulation scheduling module, a monitoring module, and an operating condition cache module.

[0036] The simulation computation module is used to run the first simulation model and the second simulation model in parallel on the runtime architecture.

[0037] The parameter correction module is used to acquire real-time operating data of the boiler in the thermal power plant, and to simultaneously correct the first and second simulation operating models online. It is also used to receive external equipment operation commands and to simultaneously send the equipment operation commands to the first and second simulation operating models.

[0038] The monitoring module is used to monitor the output parameters of the first simulation running model. When the output parameters of the first simulation running model meet the preset abnormal conditions, the second simulation running model is called to perform accelerated simulation and obtain the accelerated simulation results.

[0039] The accident analysis module is used to analyze the accelerated simulation results through a preset multi-layered accident chain and obtain the predicted accident results.

[0040] The operating condition cache module is used to save the operating condition data of the first simulation running model before the accelerated simulation starts. The operating condition data is used to restore the operation of the first simulation running model after the accelerated simulation is completed.

[0041] The operating condition generation module includes a fault database, an operating condition generator, and an I / O point table. It is used to construct the initial operating conditions for simulation and send them to the simulation calculation module. The fault database stores parameter data for various typical faults in thermal power plant boilers. The operating condition generator, combined with the controlled objects and controller variables listed in the I / O point table, constructs the initial operating conditions for simulation based on typical faults in the fault database. It can also receive real-time operating data forwarded by the data acquisition module to achieve high-precision operating condition tracking of the physical system. The two modes can be switched through the user interface. The variables in the I / O point table include variables related to the interaction between the controlled objects and the controller, as well as setpoints, controller parameters, and other variables.

[0042] The simulation scheduling module includes a scheduler and a time calculation module, used to implement the periodic scheduling of the simulation model. The scheduler is responsible for periodically scheduling the controlled object model and the controller model, accelerating the simulation by scheduling the model a fixed number of times within the same physical time frame. The time calculation module provides the scheduler with the calculation relationship between physical time and simulation time.

[0043] The monitoring module includes a data alarm module, a manual / automatic switching module, and an accident simulation module. The data alarm module generates alarm signals, employing one or more of the following: parameter exceeding threshold alarm, continuous time exceeding threshold alarm, or preset time point exceeding threshold alarm. The preset alarm information format includes the model, variable name, threshold, alarm judgment method, and non-real-time simulation time. The manual / automatic switching module sets the scheduler's scheduling signals. In automatic mode, it automatically schedules the scheduler for rapid simulation when parameters are abnormal. In manual mode, it receives information from the operation interface and schedules the scheduler for operational condition simulation. The accident simulation module constructs custom fault chains and generates accident simulation results. It uses four basic units—parameterized fault source units, logic condition units, physical quantity calculation units, and discrete event units—to construct accident chains through custom combinations. Parameters and calculation results are uniformly expressed as several basic physical quantities: length, mass, time, current, temperature, pressure, flow rate, and percentage data.

[0044] Furthermore, this application also provides a boiler fault simulation system for thermal power plants, including a data acquisition unit, a human-machine interface, and a boiler fault simulation device for thermal power plants.

[0045] The data acquisition unit includes a data acquisition module, sensors, and a signal conversion module. It is used to collect real-time operating data from the boiler in the thermal power plant and synchronously send this data to the parameter correction module and the operating condition generation module. The sensors receive signals from the physical system, digitize them, and send them to the signal conversion module. The signal conversion module converts the digitized signals into a standard format and sends them to the parameter correction module for model parameter correction and to the operating condition generation module for real-time operating condition tracking.

[0046] The human-machine interface includes an alarm display interface and an operation interface. The alarm display interface shows the predicted accident results and receives manual simulation commands input by the user. These commands are then sent to the boiler fault simulation device in the thermal power plant. The manual simulation commands trigger the second simulation model to perform accelerated simulation under specified operating conditions. The alarm display interface receives and displays alarm information from the data alarm module. Users can switch between automatic and manual fault simulation modes through the operation interface.

[0047] This application utilizes a parameter correction module to receive input from the signal transfer module of the data acquisition module, correcting the parameters of the preset controlled object model and controller model to ensure the accuracy of the model for the current physical system. In the early stages of an accident, the accelerated model automatically and rapidly performs simulation calculations to obtain the consequences after the fault has spread for a period of time. Furthermore, to ensure that the conventional and accelerated models can operate simultaneously without interference, the system adopts a logical layering and IP segmentation approach to build its overall architecture. Specifically, the module is divided into a field layer, a simulation layer, and an interface layer from bottom to top. The simulation layer is responsible for separating the computational activities of the conventional and accelerated models, and it has two different IP segments, one for the conventional model and one for the accelerated model. The accident simulation module presets five layers of fault events: a fault injection layer, a multi-field coupling layer, a standardized hazard layer, a chain reaction layer, and a final result layer. Each layer implements a different function, and these rules are used to simulate the accident.

[0048] Example 1 This embodiment takes the water-cooled wall temperature control and detection system of a thermal power plant boiler as an example. The temperature sensors are arranged at key nodes of the water-cooled wall, and the signal conversion module is implemented using an AM85C30-16JI signal driver and an MSC21X microcontroller. The signal is transmitted to the operating condition generator for input to the simulation calculation module for simulation calculation, and the signal is simultaneously sent to the parameter correction module.

[0049] Assuming that in the initial model, the dynamic model of the controlled object, the water-cooled wall, is set as follows: Because the initial model has low accuracy or the model's dynamic characteristics change during equipment use, the parameter correction module remodels after receiving sensor data, and the model parameters should be changed accordingly. Then the parameter correction module will send the modified parameters to the controlled object model.

[0050] Meanwhile, the fault database stores the operating conditions of small water-cooled wall rupture accidents. When the user switches to the initial operating conditions constructed by the fault database, the fault database directly sends the initial variable parameters of the small water-cooled wall rupture accident operating conditions to the controller model and the controlled object model.

[0051] Regarding fault alarms, the upper limit of the water-cooled wall temperature is set to 850℃, and the following three alarm modes are adopted: parameter exceeding threshold alarm, parameter exceeding threshold alarm within a continuous time period, and parameter exceeding threshold alarm at a preset time point.

[0052] For the first type of alarm setting, the preset alarm message format is as follows: The data alarm module monitors the temperature value output by the controlled object model in real time. When it detects that the temperature exceeds the preset value of 520℃, the scheduler changes the real-time simulation scheduling mode to short-time rapid scheduling, achieving the effect of the fault condition spreading for 300 seconds. The controller model and the controlled object model store the operating conditions before the non-real-time simulation in the memory, and retrieve the initial operating conditions from the memory after the non-real-time high-speed simulation ends.

[0053] For the second type of alarm setting, the preset alarm message format is as follows: The data alarm module monitors the temperature value output by the controlled object model in real time. If it finds that the temperature exceeds the preset value of 520℃, the scheduler will simulate nine more cycles in each simulation cycle. If the temperature exceeds 520℃ for ten consecutive simulation cycles, an alarm signal will be generated and non-real-time simulation will start. If the temperature does not exceed 520℃ for ten consecutive simulation cycles, the memory will return to the initial operating condition.

[0054] For the third type of alarm setting, the preset alarm message format is as follows: The data alarm module monitors the temperature value output by the controlled object model in real time, and the scheduler simulates backward for 300 seconds in each simulation cycle. If the output data at the time point after 300 seconds exceeds 520℃, an alarm signal is generated.

[0055] After completing the rapid fault simulation, the operating condition generator returns to the operating condition state before the rapid fault simulation.

[0056] Example 2 This embodiment uses a proportional-integral-derivative (PID) control strategy as an example, and the expression is as follows: ; In the formula, For the Laplace operator; The Laplace transform of the controller output; This is the Laplace transform of the system deviation.

[0057] The conversion yields the discrete-domain PID expression: ; In the formula, This is a transformation operator used to convert a continuous domain transfer function into a discrete domain recursive formula; The unit delay operator represents the signal at the previous sampling time. The z-transform of the output of the discrete controller; This is the z-transform of the discrete deviation.

[0058] That is: ; Further converted to a time-domain expression: ; In the formula, The controller output value at the sampling time; This is the sampling time sequence number.

[0059] Therefore, when the parameter correction module corrects, it takes the corresponding order parameter and modifies the current parameter accordingly; when there is no parameter correction, it takes the parameter from the previous time step and fills it into the next time step.

[0060] Example 3 like Figure 4 As shown, this embodiment takes boiler flue gas imbalance fault as an example. The accident simulation module presets a quantifiable custom fault chain, receives data from the data alarm module, and sends the accident simulation results to the human-machine interface.

[0061] The parameters corresponding to the air-smoke imbalance fault in the parameterized fault source unit include: induced draft fan frequency conversion command 50%, indicating that the induced draft fan is operating at 50% of its rated speed; induced draft fan feedback opening 25%, indicating that the actual opening of the induced draft fan damper is 25%; induced draft fan current 38A, the normal operating current; blower fan damper jamming; blower fan current 65A, the abnormal current when the damper is jammed; and multiple parameters are used to obtain intermediate events and final result events based on user-defined conditions.

[0062] The multi-field coupling layer describes the relationship between air-smoke imbalance faults and furnace temperature, furnace pressure, and combustion efficiency, and obtains the changes in relevant physical quantities through air-smoke-fuel coupling calculations; the standardized hazard layer matches two types of standardized hazard modes: abnormal furnace pressure and abnormal combustion; the interlocking action layer corresponds to protective actions such as flame detection actions and fuel reduction; and the final result layer generates the final event of main fuel tripping.

[0063] Finally, it should be noted that the above content is only used to illustrate the technical solution of this application, and is not intended to limit the scope of protection of this application. Simple modifications or equivalent substitutions made by those skilled in the art to the technical solution of this application shall not depart from the substance and scope of the technical solution of this application.

Claims

1. A method for predicting boiler failures in thermal power plants, characterized in that, include: Construct a runtime architecture with logical layering and IP segmentation isolation; Based on the aforementioned operational architecture, a first simulation operation model and a second simulation operation model for a thermal power plant boiler are constructed using digital twin technology; the first simulation operation model and the second simulation operation model operate in parallel. Based on the real-time operating data of the thermal power plant boiler, the first simulation operation model and the second simulation operation model are simultaneously corrected online. Monitor the output parameters of the first simulation running model. If the output parameters of the first simulation running model meet the preset abnormal conditions, call the second simulation running model to perform accelerated simulation and obtain the accelerated simulation results. The accelerated simulation results are analyzed by using a pre-defined multi-layered accident chain to obtain predicted accident results.

2. The method for simulating boiler failures in thermal power plants according to claim 1, characterized in that, The construction of the aforementioned runtime architecture includes: A logical layering is constructed, which includes, from bottom to top, a field layer, a simulation layer, and an interface layer; The simulation layer is divided into two independent IP network segments, which are respectively bound to the first simulation running model and the second simulation running model.

3. The method for simulating boiler failures in thermal power plants according to claim 1, characterized in that, In the accelerated simulation, the first simulation running model runs with a first simulation step size, and the second simulation running model runs with a second simulation step size. The second simulation step size is greater than the first simulation step size.

4. The method for simulating boiler failures in thermal power plants according to claim 1, characterized in that, Before calling the second simulation running model for accelerated simulation, save the operating condition data of the first simulation running model; After the second simulation running model completes the accelerated simulation, the first simulation running model is restored to operation based on the operating condition data.

5. The method for simulating boiler failures in thermal power plants according to claim 1, characterized in that, The pre-defined multi-layered accident chain includes, in sequence, a fault injection layer, a multi-field coupling layer, a standardized hazard layer, a chain action layer, and a final result layer.

6. The method for simulating boiler failures in thermal power plants according to claim 5, characterized in that, The pre-defined multi-layered incident chain construction includes: Deploy parameterized fault source units in the fault injection layer; At least one of logic condition units, physical quantity operation units, and discrete event units is deployed in the multi-field coupling layer; The standardized hazard layer is deployed with at least one of a logic condition unit, a physical quantity calculation unit, and a discrete event unit; the standardized hazard layer is also pre-encapsulated with one of the standardized hazard modes, or a combination of multiple standardized hazard modes, namely, abnormal water level, abnormal temperature, abnormal water pressure, abnormal furnace pressure, abnormal structural damage, or abnormal combustion. Each of the chain action layers is equipped with at least one of the following: a logic condition unit, a physical quantity calculation unit, and a discrete event unit. The discrete event unit is deployed to the final result layer.

7. A fault prediction device for thermal power plant boilers, characterized in that, It includes a simulation calculation module, a parameter correction module, a monitoring module, and an accident analysis module; The simulation computing module is used to run the first simulation running model and the second simulation running model in parallel on the running architecture; The parameter correction module is used to acquire real-time operating data of the boiler in the thermal power plant and simultaneously correct the first simulation operation model and the second simulation operation model online. The monitoring module is used to monitor the output parameters of the first simulation running model. When the output parameters of the first simulation running model meet the preset abnormal conditions, the second simulation running model is called to perform accelerated simulation and obtain the accelerated simulation results. The accident analysis module is used to analyze the accelerated simulation results through a preset multi-layer accident chain and obtain predicted accident results.

8. The boiler fault prediction device for thermal power plants according to claim 7, characterized in that, It also includes a working condition caching module, which is used to save the working condition data of the first simulation running model before the accelerated simulation starts; The operating condition data is used to restore the operation of the first simulation running model after the accelerated simulation is completed.

9. A boiler fault prediction system for thermal power plants, characterized in that, Includes a data acquisition unit, a human-machine interface, and the boiler fault simulation device for thermal power plants as described in any one of claims 7-8; The data acquisition unit is used to collect real-time operating data of the boiler in the thermal power plant and send the real-time operating data to the parameter correction module of the boiler fault inference device in the thermal power plant. The human-computer interaction interface is used to display the predicted accident results.

10. The boiler fault prediction system for thermal power plants according to claim 9, characterized in that, The human-computer interaction interface is also used to receive manual deduction instructions input by the user and send the manual deduction instructions to the thermal power plant boiler fault deduction device. The manual simulation command is used to trigger the second simulation running model to perform accelerated simulation under specified working conditions.

Images