An aero-engine health management algorithm design, test and verification system

Abstract

The application discloses an aero-engine health management algorithm design, test and verification system, and belongs to the technical field of aero-engines. The system comprises a real-time sending simulation unit, a test bench data integration unit, an airborne twin unit, a ground algorithm verification platform unit, a virtual twin, a data mining platform unit and an information pushing platform unit. The test bench data integration unit is used for realizing real-time management and fusion of multi-source test data, the airborne twin unit is used for driving a digital twin model cluster to perform real-time / semi-real-time fault diagnosis, and the ground algorithm verification platform unit and the data mining platform unit are used for realizing the full-process closed loop of the algorithm from offline verification, online test to airborne deployment. The application also has online and offline working modes, and can improve the research and development efficiency, verification accuracy and operation safety of the aero-engine health management algorithm by using historical and real-time data.

Description

Technical Field

[0001] This invention relates to the field of aero-engine technology, and in particular to an integrated architecture for the design, testing and verification of aero-engine health management algorithms. Background Technology

[0002] Aero-engine health management is a key technology for improving engine maintainability, safety, reliability, and reducing operating costs. In the development of aero-engine models and the tackling of key technologies, there is an urgent need for an agile and intelligent algorithm testing and verification platform.

[0003] With the development of big data and artificial intelligence technologies, a vast amount of diverse structured data has been accumulated during aero-engine testing and flight trials. However, existing technologies have not yet systematically conducted data governance and in-depth mining, failing to form high-quality data assets and limiting the development of health management technologies. Currently, in the field of aero-engine health management, there is a lack of an integrated platform that can integrate data governance, algorithm design, offline verification, online testing, and deployment. Therefore, developing an architecture that can fully utilize historical and real-time data to support the entire lifecycle research and verification of intelligent algorithms has significant engineering application value. Summary of the Invention

[0004] The purpose of this invention is to provide an architecture for the design, testing, and verification of aero-engine health management algorithms, in order to solve the problem of the lack of a unified and integrated platform in the existing technology to support the verification of algorithms from design to airborne deployment.

[0005] The technical solution of the present invention: An architecture for designing, testing, and verifying an aero-engine health management algorithm, characterized by comprising the following components: The system includes a real-time transmission simulation unit, a benchtop data integration unit, an airborne twin unit, a ground algorithm verification platform unit, a virtual twin, a data mining platform unit, and an information push platform unit.

[0006] The functions and connections of each unit are as follows: a) Real-time transmission simulation unit: used to simulate and transmit historical datasets in real time according to time series, and connected to the airborne twin unit via communication; this unit has functions of parameter configuration, data loading, batch data transmission, adjustable time multiplier transmission, start, interruption, and transmission status monitoring.

[0007] b) Bench Data Integration Unit: This unit acquires and integrates multi-source heterogeneous bench test data. It receives a first data stream from the digital electronic controller and a second data stream from the engine bench, with the first data stream having a higher sampling frequency than the second. The unit downsamples the first data stream to align it with the second data stream in the time dimension, and then fuses and reassembles the data according to a unified timestamp to generate integrated data. The data acquisition process of this unit is logically and physically isolated from the engine bench test logic.

[0008] c) Airborne twin unit: used to receive synchronous data transmitted by the real-time simulation unit or bench data integration unit, and drive the internally integrated digital twin model cluster to achieve real-time or semi-real-time evaluation of engine status; the digital twin model cluster includes an engine performance baseline model, an engine performance self-adjustment model and diagnostic algorithm components; the unit is also equipped with a supporting host computer unit for synchronously displaying the parameters calculated by the model and the diagnostic results.

[0009] d) Virtual twin: used to explicitly display the real-time parameters generated by the airborne twin unit.

[0010] e) Ground-based algorithm verification platform unit: This unit provides an offline simulation environment and admission assessment before the airborne deployment of algorithms. It loads the offline algorithm module to be verified and inputs the locally stored historical test data into the algorithm module to evaluate the accuracy, false alarm rate and real-time performance indicators of the algorithm in an offline verification manner. This unit has a unified code interaction interface, and the verified algorithm can be deployed to the airborne twin unit after code airborne processing.

[0011] f) Data mining platform unit: This is an offline analysis platform deployed on the server side, forming a parallel computing architecture with the airborne twin unit. It is used to run high-computation algorithm models that do not require real-time execution. This unit performs deep calculations on the stored historical data to obtain the long-term evolution pattern of the engine and sends the mining results through the information push platform.

[0012] g) Information push platform unit: used to push information to relevant professional systems or personnel according to specified requirements based on the results of the data mining platform unit.

[0013] Furthermore, the architecture has both offline and online operating modes; In offline mode, historical data is used to drive the airborne twin unit to perform algorithm verification through data playback via the real-time transmission simulation unit. In online mode, the test data is accessed in real time through the bench data integration unit, driving the airborne twin unit to perform real-time algorithm verification and status monitoring.

[0014] Furthermore, in the online working mode, the bench data integration unit and the engine bench test system use one-way communication to ensure that the online monitoring process does not affect test safety.

[0015] Furthermore, the diagnostic algorithm components integrated within the airborne twin unit include algorithms for diagnosing air path faults, vibration faults, and sensor faults, as well as algorithms for control quality identification, steady-state performance monitoring, and virtual perception of performance parameters.

[0016] The beneficial effects of this invention are: This paper presents an integrated architecture for the design, testing, and verification of aero-engine health management algorithms, which connects the entire process from data governance, offline algorithm verification, online real-time testing to airborne deployment, and solves the problems of platform fragmentation and inconsistent processes in existing technologies.

[0017] By setting up both online and offline working modes, historical data can be used to accelerate algorithm maturation, while real-time data can be accessed for online verification and data accumulation, thereby improving the efficiency and flexibility of algorithm development and verification.

[0018] By co-designing the airborne twin unit and the ground-based algorithm verification platform, the consistency and reliability of the algorithm from ground-based R&D to airborne application are ensured.

[0019] By using a data mining platform and an onboard twin unit-based offloading computing architecture, the collaboration between onboard real-time diagnostics and server-side in-depth analysis was achieved, optimizing the utilization of computing resources. Attached Figure Description

[0020] Figure 1 This is a diagram of the overall system.

[0021] Figure 2 This is an architecture block diagram and data interaction diagram of the online working mode of the present invention.

[0022] Figure 3 This is an architecture block diagram and data interaction diagram of the offline working mode of the present invention. Detailed Implementation

[0023] Example 1: The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] Example 1: The system consists of a real-time transmission simulation unit, a benchtop data integration unit, an airborne twin unit, a virtual twin, a ground-based algorithm verification platform unit, a data mining platform unit, and an information push platform unit. The functions of each component are described below: a) Real-time transmission simulation unit: This unit simulates the real-time transmission of historical datasets in a time-series manner, and connects to the airborne twin unit via communication. The real-time transmission simulation unit includes functions such as parameter configuration, data loading, batch data transmission, adjustable time-rate transmission, start, interrupt, and transmission status monitoring.

[0025] b) Bench Data Integration Unit: This unit acquires and integrates multi-source heterogeneous bench test data. It receives a first data stream from the digital electronic controller and a second data stream from the engine bench, with the first data stream having a higher sampling frequency than the second. The unit downsamples the first data stream to align it with the second data stream in the time dimension, and then fuses and reassembles the data according to a unified timestamp to generate integrated data. This unit has data transfer and storage functions, and its data acquisition process is logically and physically isolated from the engine bench test, ensuring independent test operation.

[0026] c) Onboard Twin Unit: This unit receives synchronous data transmitted from the real-time simulation unit or bench data integration unit and drives the internally integrated digital twin model cluster to achieve real-time or semi-real-time evaluation of engine status. The digital twin model cluster includes an engine performance baseline model, an engine performance self-adjustment model, and diagnostic algorithm components. The performance baseline model calculates benchmark values ​​for temperature, pressure, and unmeasurable parameters such as thrust, fuel consumption rate, and surge margin at various airflow cross-sections of the engine under standard conditions. The performance self-adjustment model synchronously tracks the individual state of the tested engine based on real-time input data through parameter correction and outputs measured estimates of cross-section parameters and unmeasurable parameters of the same category as the baseline model. The diagnostic algorithm components include algorithms for diagnosing airflow faults, vibration faults, and sensor faults, as well as algorithms for control quality identification, steady-state performance monitoring, and virtual perception of performance parameters. Diagnostic conclusions are derived through comprehensive analysis of the benchmark values ​​and measured estimates. This unit also has a supporting host computer unit for synchronously displaying the parameters calculated by the models and the diagnostic results.

[0027] d) Virtual twin: This part is used to make the real-time parameters of the airborne twin unit explicit.

[0028] e) Ground Algorithm Verification Platform Unit: This unit provides an offline simulation environment and admission assessment before the airborne deployment of algorithms. It loads the offline algorithm module to be verified and inputs locally stored historical test data (including real data without anomalies and real data with anomalies) into the algorithm module to evaluate the algorithm's accuracy, false alarm rate, and real-time performance indicators through offline verification. The verification covers algorithms including: engine airflow and vibration fault detection, life prediction and parameter estimation, fault location and degradation detection, and trend prediction and threshold analysis algorithms. This unit has a unified code interaction interface. Verified algorithms are subsequently deployed to the airborne twin unit through code airborne processing. Simultaneously, the verified algorithm model can be directly applied to the system's offline analysis phase, thus ensuring the consistency and reliability of the algorithm from ground development to airborne application.

[0029] f) Data Mining Platform Unit: This is an offline analysis platform deployed on the server side. It forms a parallel computing architecture with the airborne twin unit and is used to run computationally intensive algorithm models that do not require real-time execution. This unit integrates non-real-time algorithms verified by the ground-based algorithm verification platform unit, specifically including: engine life prediction and parameter estimation algorithms, and engine trend prediction and threshold analysis algorithms. This unit performs deep computation on stored historical data to obtain the long-term evolution patterns of the engine and links the mining results to professional users through an information push platform. By distributing computational tasks according to real-time requirements, this unit achieves collaboration between real-time airborne diagnostics and deep server-side analysis.

[0030] g) Information push platform unit: Based on the data mining platform unit, relevant results are pushed to relevant professionals according to specified requirements for decision-making by each professional, while strengthening collaboration among multiple professionals.

[0031] like Figure 2 As shown, in online working mode, the operation flow of this architecture is as follows: During the aero-engine bench testing, the bench data integration unit receives data in real time from both the digital electronic controller (which acquires the high-frequency first data stream) and the engine bench measurement and control system (which acquires the low-frequency second data stream). This unit downsamples, aligns, and reassembles the data to generate standardized integrated data, which is then stored locally.

[0032] The fused real-time data is transmitted to the airborne twin unit via a communication link. To ensure experimental safety, this communication link is unidirectional.

[0033] Meanwhile, the algorithm code that has completed offline verification in the ground algorithm verification platform unit is deployed to the airborne twin unit after code airborne processing.

[0034] The airborne twin unit receives real-time data and drives the internal performance baseline model, performance self-adjustment model and various diagnostic algorithms to run in parallel. It calculates the deviation between the benchmark values ​​and measured estimates of key engine parameters online and performs real-time diagnosis of faults in the air path, vibration, and sensors.

[0035] The diagnostic results and key model parameters are sent to the host computer unit, which then sends this data into a virtual twin for 3D visualization or charting, allowing test engineers and health management experts to monitor the engine status in real time.

[0036] Offline working mode such as Figure 3 As shown, in offline working mode, the operation flow of this architecture is as follows: Historical test or flight test data of aero-engines are compiled into high-quality datasets and loaded into the real-time transmission simulation unit, the ground algorithm verification platform unit, and the data mining platform unit, respectively.

[0037] The real-time transmission simulation unit simulates the generation process of historical data according to a set time multiplier (such as 1x speed, slow motion, or fast motion) and sends the data stream to the airborne twin unit.

[0038] The ground-based algorithm verification platform unit loads the offline algorithm module to be verified and uses historical data to conduct comprehensive offline testing and evaluation of the algorithm, verifying its accuracy, false alarm rate, and real-time performance. After successful verification, the algorithm code is split into two paths: one path is processed by airborne processing and sent to the airborne twin unit; the other path is sent to the data mining platform unit.

[0039] The airborne twin unit receives simulated data streams from the real-time transmission simulation unit, runs the airborne algorithm, and performs offline fault diagnosis. The diagnostic results are displayed on the virtual twin by the host computer for retrospective analysis.

[0040] As a server-side platform, the data mining platform unit utilizes its powerful computing capabilities to run computationally intensive algorithms such as lifespan prediction and trend analysis to deeply mine long-term historical data and analyze the patterns of engine performance degradation.

[0041] The generated analysis reports, trend charts, and other results are transmitted to the information push platform unit. This unit automatically pushes the information to different professional departments such as design, maintenance, and management according to preset rules, providing support for decision-making.

[0042] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A system for designing, testing, and verifying an aero-engine health management algorithm, characterized in that, include: The system includes a real-time transmission simulation unit, a benchtop data integration unit, an airborne twin unit, a ground-based algorithm verification platform unit, a virtual twin, and a data mining platform unit. The real-time transmission simulation unit is used to simulate and transmit historical datasets in real time according to time series, and is connected to the airborne twin unit; The bench data integration unit is used to acquire and integrate multi-source heterogeneous bench test data, and send the integrated data to the airborne twin unit; The airborne twin unit is used to receive data sent by the real-time simulation unit or the bench data integration unit, drive the internally integrated digital twin model cluster to perform engine status assessment, and output diagnostic results. The ground-based algorithm verification platform unit is used to load offline algorithms and perform offline verification using historical data before the algorithm is deployed on an airborne basis. The virtual twin is used to receive the output parameters of the airborne twin unit and to visualize them; The data mining platform unit is a server-side platform, located alongside the airborne twin unit, and is used to run non-real-time, computationally intensive algorithm models to perform in-depth mining of historical data.

2. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 1, characterized in that: It also includes an information push platform unit, which is used to push the results of the data mining platform unit to a designated target.

3. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 1, characterized in that: The test bench data integration unit receives a first data stream from the digital electronic controller and a second data stream from the engine test bench, with the sampling frequency of the first data stream being higher than that of the second data stream. This unit downsamples the first data stream to align it with the second data stream in the time dimension, and then fuses and reassembles the data according to a unified timestamp.

4. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 1, characterized in that... The digital twin model cluster integrated within the airborne twin unit includes: an engine performance baseline model for calculating engine standard state reference parameters, a performance self-adjustment model for tracking individual engine states, and a diagnostic algorithm component that includes algorithms for diagnosing air path faults, vibration faults, and sensor faults.

5. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 1, characterized in that: The ground algorithm verification platform unit loads the offline algorithm to be verified and inputs the historical test data stored locally. The accuracy, false alarm rate and real-time performance indicators of the algorithm are evaluated in an offline verification manner. The verified algorithm is deployed to the airborne twin unit after code airborne processing to ensure the consistency of the algorithm from ground verification to airborne application.

6. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 1, characterized in that: The data mining platform unit integrates algorithms including engine life prediction and parameter estimation algorithms, and engine trend prediction and threshold analysis algorithms.

7. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 1, characterized in that: It has both offline and online working modes. In the offline working mode, the real-time transmission simulation unit uses historical data to simulate real-time transmission and drive the airborne twin unit to work. In online working mode, the bench data integration unit accesses real bench test data and drives the airborne twin unit to work in real time.

8. The system for designing, testing, and verifying the aircraft engine health management algorithm according to claim 7, characterized in that: In online operating mode, the bench data integration unit and the engine bench test system use a one-way communication link to ensure that online monitoring does not affect test safety.

9. The system for designing, testing, and verifying an aero-engine health management algorithm according to claim 7, characterized in that: In offline working mode, algorithms verified by the ground algorithm verification platform unit are simultaneously sent to the data mining platform unit and, after airborne processing, sent to the airborne twin unit.

Images