Electric Aircraft Energy Dynamic Optimization Management System

Abstract

An energy dynamic optimization management system for an electric aircraft, which can optimize in real time to best fit the full-life cycle aging process of the electric aircraft. It includes a monitoring module and multiple management units. Each management unit has a health assessment and fault diagnosis / prediction module (the former) and a dynamic optimization management module (the latter). The former can calculate the health index of the electric aircraft at the current moment and can predict the health states of the electric aircraft at each moment after the current moment. The latter can execute a rule-based control mode and an energy optimization mode at different time sequences, and through integrating the time sequences of different control modes, achieve real-time control of multiple management objects at different levels. When the latter is executing the time sequence of the energy optimization mode, based on the health state of the electric aircraft predicted by the former after the current moment, it performs control with better comprehensive performance for each management object compared to the rule-based control mode.

Description

Technical Field

[0001] The present invention relates to the field of energy management of aircraft, and more particularly to a dynamic optimization management system for the energy of an electric aircraft. Background Art

[0002] In an electric aircraft, a large number of electrified systems and devices are electrically coupled to work together, greatly improving the degree of integration of the energy system. The management and control of this energy system can be described as "pulling one hair and moving the whole body". Although it is possible to choose to manage and control each system and device separately, this is a waste of the limited aircraft energy of the electric aircraft. Therefore, it is necessary to start from an overall perspective and optimize the use of the whole aircraft energy to give full play to the advantages of the electric aircraft and improve the endurance of the electric aircraft.

[0003] In addition, the electrified devices in an electric aircraft exhibit characteristics such as high power, non-linearity, and strong uncertainty. Their dynamic processes change violently and their working states are complex. The steady-state control technology of traditional aircraft energy is difficult to meet the management requirements of the energy system of an electric aircraft, and may even further cause the system to oscillate and collapse. Therefore, for the energy management of an electric aircraft, how to achieve the control of the dynamic processes of the work of the system and devices is also a realistic problem that has to be faced currently.

[0004] In addition, the operating environment of the aircraft is complex, and the aging trends and degrees of the systems and devices distributed at different positions of the aircraft are not the same during the whole life cycle, which brings new challenges to the energy management of the electric aircraft. For example, as the power battery in the electric aircraft is repeatedly used, the performance of the power battery ages, but its aging degree is higher than that of its load devices, and there may be a situation where the power supply capacity of the power battery is insufficient. At this time, if the power battery continues to operate according to the original control strategy, it will be overloaded, which in turn will further accelerate the aging of the power supply. The reality is that each part of the energy system of the electric aircraft may face such problems, and it is urgent to seek a dynamic optimization management and control method for the energy of the electric aircraft that is as suitable as possible for the whole life cycle aging process of the electric aircraft (its various systems and key components). Summary of the Invention

[0005] The present invention is made to solve the above-mentioned technical problems, and its purpose is to provide a dynamic optimization management system for the energy of an electric aircraft that is as suitable as possible for the whole life cycle aging process of the electric aircraft (its various systems and key components).

[0006] To achieve the object of the present invention, there is provided a dynamic optimization management system for the energy system of an electric aircraft, which is used to optimize, manage and control the energy system of the electric aircraft. Among them, the dynamic optimization management system for the energy system of the electric aircraft includes: a monitoring module, which collects energy system parameters in real time and extracts characteristic parameters related to the energy system of the electric aircraft from them; a plurality of management units, which manage a plurality of management objects at different levels of the electric aircraft. Each management unit has a health assessment and fault diagnosis / prediction module and a dynamic optimization management module. The health assessment and fault diagnosis / prediction module evaluates the health status of each management object at the current moment, and calculates the health index of the electric aircraft at the current moment in a bottom-up hierarchical manner. And for the health indexes calculated for each management object at each moment before the current moment, by adding time series information, a change trend curve of the health index of each management object over time is fitted, and the health status of the electric aircraft at each moment after the current moment is predicted in a bottom-up hierarchical manner. The dynamic optimization management module can execute a rule-based control mode and an energy optimization mode for the energy system of the electric aircraft at different time series, and integrate the time series of different control modes by adding the time series of the rule-based control mode between two adjacent time series when executing the energy optimization mode, so as to realize the real-time control of a plurality of management objects at different levels. When the dynamic optimization management module is executing the time series of the energy optimization mode, based on the health status of the electric aircraft predicted by the health assessment and fault diagnosis / prediction module after the current moment, a control with better comprehensive performance for each management object is performed compared with the rule-based control mode.

[0007] According to the above composition, the dynamic optimization management system for the energy system of the electric aircraft of the present invention evaluates the power output, transmission and consumption capabilities of the energy system of the electric aircraft in different health states through power prediction technology, combines health status assessment and prediction with the real-time operation management strategy of the energy system of the electric aircraft, and adjusts the energy management and operation strategies in real time according to the health status of the aircraft throughout its life cycle, thereby improving the service life of the system and equipment, reducing the risk of aircraft damage and maintenance costs, and improving the reliability and operation efficiency of the aircraft.

[0008] In addition, according to the above composition, the dynamic optimization management system for the energy system of the electric aircraft of the present invention executes an energy optimization mode through model predictive control technology to perform real-time optimization management on the intense dynamic process of the energy system of the electric aircraft, thereby making up for the deficiencies of traditional steady-state control and improving the steady-state and transient performance of the aircraft.

[0009] In addition, configured as described above, in the dynamic optimization management system of the electric aircraft energy of the present invention, the dynamic optimization management module can execute a rule-based control mode and an energy optimization mode on the electric aircraft energy system at different time sequences, and integrate different control mode time sequences. Thus, by establishing a two-layer control structure, with the traditional rule-based (such as "worst case") control mode as the lower layer structure, the problem of poor real-time performance caused by the long optimization time of the fault prediction algorithm and the optimization algorithm is solved, and real-time management of the aircraft is achieved.

[0010] Preferably, the monitoring module includes a sensor network module, a data processing module, and a feature extraction module. The sensor network module is used to collect the original data samples of the electric aircraft. The data processing module preprocesses the original data samples to filter out interference signals. The feature extraction module extracts feature parameters that can independently or mutually confirm to characterize a certain / some specific state / characteristics of the electric aircraft from the preprocessed parameters to achieve dimensionality reduction of the data.

[0011] With the configuration as described above, since the original data samples are preprocessed first to filter out interference signals, and feature parameters meaningful for determining the state of the electric aircraft are extracted from a large number of preprocessed energy system parameters, the computational burden on the dynamic optimization management system of the electric aircraft energy caused by high-dimensional data is reduced, the response degree of the electric aircraft energy optimization is improved, and a basis for subsequent dynamic optimization management is provided.

[0012] Preferably, the dynamic optimization management system of the electric aircraft energy includes: an overall aircraft management unit that manages the entire electric aircraft as the management object; a plurality of system-level management modules. The system-level management modules have a centralized management department that manages the object systems that make up the entire electric aircraft as the management object, and at least one of the plurality of system-level management modules adopts a distributed system energy management architecture with a two-layer structure of centralized + distributed. In the distributed system energy management architecture, there is a key component management department that manages the key components that make up the object system as the management object. The plurality of management units are the overall aircraft management unit, the centralized management departments in each system-level management module, and the key component management departments in the system-level management modules that adopt the distributed system energy management architecture.

[0013] More preferably, a part of the multiple system-level management modules adopt a centralized system energy management architecture, and the multiple system-level management modules adopting a distributed system energy management architecture and a centralized system energy management architecture respectively implement information interaction between each object system and the whole electric aircraft with the whole aircraft management unit through the whole aircraft communication module. In addition, in the system-level management module adopting a distributed system energy management architecture, the centralized management department and the key component management department implement information interaction with each other through the system communication module.

[0014] The electric aircraft energy dynamic optimization management system of the present invention includes a distributed system energy management architecture with a centralized + distributed double-layer structure as the system-level management module. In addition, the whole aircraft management unit can globally optimize the working states of each object system from the whole electric aircraft to improve the comprehensive operation performance of the aircraft. In addition, because this distributed system energy management architecture realizes the interaction and collaborative work of the object system and the key components, it reduces the computational performance requirements of the control algorithm for the processor and improves the control performance.

[0015] As an example, the health assessment and fault diagnosis / prediction modules of multiple management units at different levels have a health assessment department for implementing the health assessment function, a fault diagnosis department for implementing the fault diagnosis function, and a fault prediction department for implementing the fault prediction. The health assessment department assesses the health state of the management object at the current level according to the health state information of the management object at the next lower level, and transmits the assessment result to the upper level. Once the health assessment department finds that the health state of the management object deteriorates, the fault diagnosis department traces and isolates the fault in a top-down manner. Based on the analysis results of the health assessment department and the fault diagnosis department, the fault prediction department predicts the change trend of the health state of each management object level by level from bottom to top, and outputs the aging characteristic parameters.

[0016] Furthermore, or as another example, the dynamic optimization management module of the management unit at each level has a power prediction department. The power prediction departments in the management units at the lowest level under different architectures estimate the change curve of the power demand of their respective management objects according to the influence of factors including preset working requirements, performance degradation, and uncertainty of the management objects at this level on the future operation trend. The management units at a higher level than the lowest level solve the change curve of the power demand of the management objects at this level according to the connection relationship of the management objects at the next lower level and the change curves of the power demand predicted by the power prediction departments of the management units at the next lower level.

[0017] On this basis, the dynamic optimization management module of the management unit at each level has an energy optimization unit. The energy optimization unit refers to the power requirements of each management object, determines the optimal comprehensive performance as the optimization goal, and takes the current and predicted health status information of the management object and the operation requirements of the management object at this level as constraints, and solves the working power with the optimal comprehensive performance for each management object under the power requirements, that is, the optimized operation power. At this time, the management unit at the highest level under different management architectures uses the change curve of the power requirements of the management object at this level as the estimated operation power of the management unit, and combines the current and predicted health status information of the management object obtained by the health assessment and fault diagnosis / prediction module to optimize the estimated operation power to obtain the optimized operation power of each management unit at the highest level. The management unit at a lower level than the highest level solves the estimated operation power of the management unit at the lower level according to the connection relationship of its management object and the optimized operation power optimized by the energy optimization unit of the management unit at the higher level, and then solves the optimized operation power of each management unit at the next lower level.

[0018] Furthermore, the dynamic optimization management module of the management unit at each level also has a dynamic control unit. When the energy optimization mode and the rule-based control mode output control signals at the same time, the dynamic control unit preferentially adopts the output of the control signal of the energy optimization mode, and controls the whole machine and the system in the rule-based control mode during the cycle interval of the output of the control signal of the energy optimization mode.

[0019] In the present invention, the rule-based control mode controls the working power of each management object according to established rules. The energy optimization mode predicts the power requirements, and uses the prediction of the aging status of each management object as the health status obtained by the health assessment and fault diagnosis / prediction module to optimize the control of the working power of each management object. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a schematic diagram showing the architecture of the electric aircraft energy dynamic optimization management system according to the embodiment of the present invention.

[0021] Figure 2 is a schematic block diagram showing the functional structure of the monitoring module in the electric aircraft energy dynamic optimization management system according to the embodiment of the present invention.

[0022] Figure 3It is a schematic diagram showing the general functional structures of the centralized management unit, key component management unit, centralized system energy management architecture, and overall aircraft management unit in the electric aircraft energy dynamic optimization management system according to the embodiments of the present invention.

[0023] Figure 4 It is a schematic diagram showing the operation processes and relationships of the health assessment and fault prediction modules for the overall aircraft, system, and key components at three levels.

[0024] Figure 5 It is a schematic diagram showing the functional structure of the dynamic optimization management module.

[0025] Figure 6 (a), (b), and (c) thereof are schematic diagrams of the timing of control instructions for the rule-based control mode, control instructions for the energy optimization mode, and control instructions output by the dynamic control unit.

[0026] Figure 7 and Figure 8 It is a schematic diagram for explaining the principle of the energy optimization function implemented in the energy optimization unit, where Figure 7 is the output of power prediction and control signals at the k-th moment, Figure 8 is the output of power prediction and control signals at the next (k + 1)-th moment.

[0027] (Symbol Explanation)

[0028] 100 Electric aircraft energy dynamic optimization management system

[0029] 110 Monitoring module

[0030] 111 Sensor network (functional module)

[0031] 112 Data processing (functional module)

[0032] 113 Feature extraction (functional module)

[0033] 120 System-level management module

[0034] 121 Distributed system energy management architecture

[0035] 121a Centralized management department

[0036] 121b Key component management department

[0037] 121c System communication module

[0038] 122 System energy management architecture

[0039] 122a Centralized management department

[0040] 130 Whole Machine Communication Module

[0041] 140 Whole Machine Management Unit

[0042] 200 Electric Aircraft Energy System

[0043] 210 Energy Generation / Storage System

[0044] 220 Energy Conversion / Transmission System

[0045] 230 Energy Conversion / Utilization System

[0046] 300 Ground PHM Module

[0047] A Health Assessment and Fault Diagnosis / Prediction Module

[0048] A11 Whole Machine Health Assessment Department

[0049] A12 Whole Machine Fault Diagnosis Department

[0050] A13 Whole Machine Fault Prediction Department

[0051] A21 System Health Assessment Department

[0052] A22 System Fault Diagnosis Department

[0053] A23 System Fault Prediction Department

[0054] A31 Key Component Health Assessment Department

[0055] A32 Key Component Fault Diagnosis Department

[0056] A33 Key Component Fault Prediction Department

[0057] B Dynamic Optimization Management Module

[0058] B1 Power Prediction Department

[0059] B2 Energy Optimization Department

[0060] B3 Dynamic Control Department

[0061] C1 Rule-Based Control Mode

[0062] C2 Energy Optimization Mode Specific Embodiments

[0063] Hereinafter, in conjunction with the accompanying drawings, the specific embodiments of the present invention will be described in detail. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to unduly limit the protection scope of the present invention.

[0064] Figure 1It is a schematic diagram showing the architecture of the electric aircraft energy dynamic optimization management system 100 according to an embodiment of the present invention.

[0065] As Figure 1 shown, the electric aircraft energy dynamic optimization management system 100 according to an embodiment of the present invention is used to optimize, manage, and control the electric aircraft energy system 200 of an electric aircraft (not shown). Among them, the electric aircraft energy system 200 has an energy generation / storage system 210 composed of multiple lithium batteries, an energy conversion / transmission system 220 as a power distribution system, and an energy conversion / utilization system 230 for realizing the operation and work tasks of the electric aircraft.

[0066] The energy conversion / transmission system 220 as a power distribution system is mainly based on a redundant and fault-tolerant power network, and realizes the conversion, transmission, and distribution of energy through various types of power conversion devices such as DC / DC, AC / DC, DC / AC, and devices such as cables, connectors, fuses, power distribution, and protection devices.

[0067] In addition, the energy conversion / utilization system 230 may include, but is not limited to, at least any one of the electric aircraft's electric environmental control system, electric de-icing system, fly-by-wire control system, electric braking system, electric fuel pump system, avionics system, and electric propulsion system, etc.

[0068] The electric aircraft energy dynamic optimization management system 100 according to an embodiment of the present invention includes a monitoring module 110. The monitoring module 110 can, for example, collect energy system parameters in real time through intelligent sensing technology, and can extract characteristic parameters related to the electric aircraft energy system 200 from a large number of collected (and pre-processed) energy system parameters as the original input values for optimizing and managing the electric aircraft (each system and device). The extracted characteristic parameters can independently or mutually confirm the characteristic parameters representing a certain / some specific state / characteristics of the electric aircraft.

[0069] Figure 2 It is a schematic block diagram schematically showing the functional structure of the monitoring module 110 in the electric aircraft energy dynamic optimization management system 100 according to an embodiment of the present invention. As an example, as Figure 2As shown, the monitoring module 110 includes three functional modules: a sensor network 111, data processing 112, and feature extraction 113. Further, the functional module of the sensor network 111 is used to collect the original data samples of the electric aircraft. In addition, the functional module of data processing 112 filters out interference signals through, for example, preprocessing techniques, thereby improving the sampling accuracy and credibility of the data samples. Additionally, the functional module of feature extraction 113 extracts feature parameters that can independently or mutually confirm to characterize a certain / some specific states / characteristics of the electric aircraft from the preprocessed energy system parameters through, for example, data fusion methods, in order to achieve dimensionality reduction of the data. At the same time, by extracting feature parameters that are meaningful for determining the state of the electric aircraft from a large number of energy system parameters, the computational burden on the energy dynamic optimization management system 100 of the electric aircraft caused by high-dimensional data is reduced, the responsiveness of the energy optimization of the electric aircraft is improved, and a basis for subsequent dynamic optimization management is provided.

[0070] The monitoring module 100 monitors the health state feature parameters of each key component, system, and the aircraft as a whole. The core of its design lies in the layout of the sensor network 111, that is, how to comprehensively grasp the health state of the electric aircraft with as few sensors as possible. Specifically, after collecting the original data samples of the electric aircraft, the monitoring module 110 combines the structure of the electric aircraft energy system 200 to analyze the fault modes and their impacts, and then establishes a testability model based on a multi-signal graph to describe the relationship between the health state of the energy system and the system characteristic parameters of the test. Then, through a preliminary analysis of the testability model, the monitoring module 110 can, in the stage of data processing 112: (1) identify and eliminate the feature parameters irrelevant to the system health state, that is, irrelevant parameters; (2) find the redundant parameters that can simultaneously characterize the health state of a certain system; (3) find the system health states that cannot be described by the feature parameters or the health states that cannot be distinguished, and further analyze these states to clarify the relevant feature parameters; (4) further divide the data into local parameters that are only related to the health states of the system and key components and global parameters that are also related to the health state of the aircraft as a whole, in order to reduce the communication and monitoring burden of the operation data of the electric aircraft. The monitoring module 110 extracts the system parameters (local parameters and global parameters) that characterize the system health state in the subsequent feature extraction 113 stage.

[0071] Combined with the implementation problems in actual projects, the data processing 112 needs to consider the impact of the number of tests on the test effect. If the number of tests is too small, the healthy state of the energy system may not be recognized. If the number of tests is too large, it will cause waste of resources. Therefore, based on the mapping relationship between the healthy state described by the testability model and the characteristic parameters, for example, based on a multi-objective genetic algorithm, etc., with the test cost (the number of parameters, the information content of the parameters, the extraction cost of the parameters) as the optimization objective, the characteristic parameters of the test are preferably selected, so as to obtain the layout of the sensor network 111 for the healthy state of the aircraft energy system, and avoid repeated, multiple and unnecessary acquisition of the same or equivalent characteristic parameters representing the same state / characteristic.

[0072] In the electric aircraft energy dynamic optimization management system 100 according to the embodiment of the present invention, the characteristic parameters extracted by the feature extraction 113 of the monitoring module 110 are output to the system-level management module 120. The system-level management module 120 can select the following two different architectures for system management according to the type of the system, the complexity of the component devices, or the number of control devices, etc.:

[0073] (1) For a system with a relatively wide function and spatial distribution (also called a "distributed object system"), because the component devices are complex, and there may be more than 3 control devices for independent and partially redundant management of multiple key components of the system, therefore, a distributed system energy management architecture 121 is selected;

[0074] (2) For a system with a relatively concentrated function and spatial distribution (also called a "centralized object system"), because the component devices are simple, and there may be only 3 or less than 3 control devices for backup management of the system, therefore, a centralized system energy management architecture 122 is selected.

[0075] More preferably, as Figure 1 shown, the aforementioned distributed system energy management architecture 121 adopts a two-layer structure of centralized + distributed, that is, it has a centralized management department 121a and a key component management department 121b. The centralized management department 121a takes the corresponding object system (distributed object system) as the management object, and performs system-level energy optimization management according to the healthy state of each key component. In addition, the key component management department 121b takes the corresponding key component as the management object, and realizes the optimization control of the working state of the key component as the management object.

[0076] The centralized management department 121a and the key component management department 121b realize information interaction with each other through the system communication module 121c, including but not limited to the reporting of component and system state parameters and the issuance of control and request signals, etc.

[0077] The central management unit 121a of the distributed system energy management architecture 121 and the central management unit 122a of the centralized system energy management architecture 122 can send optimization control signals to the electric aircraft energy system 200.

[0078] In addition, the distributed system energy management architecture 121 and the centralized system energy management architecture 122 are respectively connected to the whole aircraft communication module 130 to realize information interaction between each system (i.e., each distributed object system and each centralized object system) and the whole aircraft, including but not limited to the reporting of system and whole aircraft status parameters, and the sending down of control commands and request signals, etc.

[0079] The system communication module 121c and the whole aircraft communication module 130 can select a suitable communication method according to the requirements of communication reliability, the number of nodes, data volume, transmission rate, distance, etc. in the actual application scenario. For example, any one or more of power line carrier communication, ARINC 429, ARINC 664, ARINC 825, TTP, 1553B, and 5G communication can be selected, thereby further establishing the communication architecture in the electric aircraft energy dynamic optimization management system 100.

[0080] In addition, the whole aircraft communication module 130 is communicatively connected to the whole aircraft management unit 140. In addition, the whole aircraft management unit 140 is communicatively connected to the ground PHM module 300. The ground PHM module 300 performs health management on the aircraft platform throughout its life cycle, supports maintenance and repair management decisions, realizes fault prediction of components and systems, and at the same time enables the ground maintenance system to timely dispatch spare parts, maintenance equipment, and maintenance personnel, providing a basis for the logistics support and repair decision-making of the transport platform, and reducing the maintenance time and cost.

[0081] Figure 3 It is a schematic diagram showing the general functional structures of the central management unit 121a and the key component management unit 121b of the distributed system energy management architecture 121, the central management unit 122a of the centralized system energy management architecture 122, and the whole aircraft management unit 140 in the electric aircraft energy dynamic optimization management system 100 according to the embodiments of the present invention. Figure 4 It is a schematic diagram showing the operation process and relationship of the health assessment and fault diagnosis / prediction module A for the whole aircraft, system, and key components as three levels. Figure 5 It is a schematic diagram showing the functional structure of the dynamic optimization management module B. Figure 6 The (a), (b), and (c) are schematic diagrams of the timing of the control instructions of the rule-based control mode C1, the control instructions of the energy optimization mode C2, and the control instructions output by the dynamic control unit B3.

[0082] The centralized management unit 121a and the critical component management unit 121b of the distributed system energy management architecture 121, the centralized management unit 122a of the centralized system energy management architecture 122, and the whole machine management unit 140 have the same functional modules, that is, as Figure 3 shown, they have a health assessment and fault diagnosis / prediction module A and a dynamic optimization management module B.

[0083] The health assessment and fault diagnosis / prediction module A of the electric aircraft energy dynamic optimization management system 100 according to the embodiment of the present invention can be further divided into three levels: the whole machine, the system, and the critical components. By adopting a two-layer structure of centralized + distributed in the system (distributed system energy management architecture 121), it is possible to perform dynamic optimization management and control on the whole electric aircraft, the system (distributed system energy management architecture 121), and the critical components.

[0084] The health assessment and fault diagnosis / prediction module A has three functions: health assessment, fault diagnosis, and fault prediction. The health assessment function evaluates the health status of the current control level based on the health status information of the system or critical components at the next control level, and transmits the evaluation result to the upper level. Once a decline in the health status of the management object is detected through the health assessment function, the fault diagnosis function will be activated to trace back from top to bottom and perform fault isolation. The fault prediction function is based on the analysis results of the health assessment function and the fault diagnosis function, predicts the change trend of the health status of the management object level by level from bottom to top, and outputs the aging characteristic parameters.

[0085] In Figure 4 the whole machine management unit 140, the health assessment and fault diagnosis / prediction module A includes a whole machine health assessment unit A11, a whole machine fault diagnosis unit A12, and a whole machine fault prediction unit A13. In addition, the health assessment and fault diagnosis / prediction module A in the centralized management unit 121a of the distributed system energy management architecture 121 and the centralized management unit 122a of the centralized system energy management architecture 122 includes a system health assessment unit A21, a system fault diagnosis unit A22, and a system fault prediction unit A23. Furthermore, the health assessment and fault diagnosis / prediction module A in the critical component management unit 121b of the distributed system energy management architecture 121 includes a critical component health assessment unit A31, a critical component fault diagnosis unit A32, and a critical component fault prediction unit A33.

[0086] In each distributed object system, the key component health assessment unit A31 selects different health assessment methods according to the characteristics of different key components, uses the health state characteristic parameters of the corresponding key components, combines the experimental data and empirical data of the key components, and uses the health assessment method to evaluate the health state level of the key components, and sends the health state information of the key components to the centralized management unit 121a of the distributed system energy management architecture 121.

[0087] Next, the system health assessment unit A21 of each distributed object system establishes a corresponding fault propagation model of the distributed object system according to the connection relationship of each key component, uses the health state information of each key component as input, and calculates the health state (or health index HeI) of each distributed object system from point to area, and transmits the health state information (or health index HI) to the upper-level overall machine health assessment unit A11.

[0088] At the same time, the system health assessment unit A21 of each centralized object system selects different health assessment methods according to the characteristics of the systems and key devices in different centralized object systems, uses the health state characteristic parameters of the corresponding centralized object system, combines the experimental data and empirical data, and uses the health assessment method to evaluate the health state level, and transmits the health state information (or health index HI) to the upper-level overall machine health assessment unit A11.

[0089] In addition, the overall machine health assessment unit A11 of the overall machine management unit 140 establishes a fault propagation model of the overall machine according to the connection relationship of each system (that is, each distributed object system and each centralized object system), uses the health state information (or health index HeI) of each system as input, and calculates the health state (or health index HI) of the overall machine from point to area.

[0090] That is, when evaluating the health state of the electric aircraft, the health assessment and fault diagnosis / prediction module A performs health assessment on each centralized object system in the order of the system health assessment unit A21 the overall machine health assessment unit A11, and for each distributed object system, it performs health assessment in the order of the key component health assessment unit A31 the system health assessment unit A21 the overall machine health assessment unit A11.

[0091] The purpose of fault diagnosis is to detect and isolate the occurring faults and ensure the normal operation of the system. At the same time, there is a certain mapping relationship between the fault state and the health state, that is, when the health state parameters drop significantly, there must be one or some systems that have failed. Therefore, in the electric aircraft energy dynamic optimization management system 100 of the present invention, the fault diagnosis unit does not run continuously, but runs in a condition-triggered manner, and according to the diagnostic strategy formulated based on testability, faults are located from the whole machine, systems to components, from top to bottom.

[0092] That is, the whole machine fault diagnosis unit A12 of the whole machine management unit 140 cyclically detects each signal in turn based on the fault diagnosis strategy formulated based on testability, locates the system where the fault occurs, and sends the diagnosis result to the centralized management units 121a and 122a of the energy management architecture corresponding to the faulty system.

[0093] In addition, the health assessment and fault diagnosis / prediction module A of the whole machine receives the health state characteristic parameters and aging curves of each system (i.e., each distributed object system and each centralized object system) from the whole machine communication module 130, combines the massive operation data stored in the ground PHM module 300 to conduct a health state assessment and fault prediction on the electric aircraft as a whole, and returns the prediction result to the ground PHM module 300 to help carry out the maintenance and repair of the electric aircraft.

[0094] The system fault diagnosis unit A22 of each distributed object system cyclically detects each signal in turn based on the fault diagnosis strategy formulated based on testability, locates the key components where the fault occurs, sends the diagnosis result to the key component management unit 121b of the faulty component, and the key component fault diagnosis unit A32 locates the specific type and cause of the fault.

[0095] At the same time, the system fault diagnosis unit A22 of the centralized object system locates the specific type and cause of the fault according to the characteristic signal of the specific fault.

[0096] That is, when diagnosing the faults of the electric aircraft, the health assessment and fault diagnosis / prediction module A conducts fault diagnosis and location on each centralized object system in the order of the whole machine fault diagnosis unit A12 The system fault diagnosis unit A22, and conducts fault diagnosis and location on each distributed object system in the order of the whole machine fault diagnosis unit A12 The system fault diagnosis unit A22 The key component fault diagnosis unit A32.

[0097] The system fault prediction unit A23 of each distributed object system incorporates time information into the network nodes of the corresponding distributed object system, takes the aging characteristic curves of each key component predicted by the key component fault prediction unit A33 as input, and analyzes and predicts the aging characteristic curve of the corresponding distributed object system through the fault propagation mechanism analysis.

[0098] The system fault prediction unit A23 of each centralized object system selects different fault prediction methods according to the characteristics of the systems or key components in different centralized object systems, and analyzes and predicts the aging characteristic curve of the corresponding centralized system.

[0099] The whole machine fault prediction unit A13 of the whole machine management unit incorporates time information into the network nodes connected to each system, takes the aging characteristic curves of each system (i.e., each distributed object system and each centralized object system) as input, and analyzes and predicts the aging characteristic curve of the whole machine through the fault propagation mechanism analysis.

[0100] That is, when predicting the potential faults (such as aging, etc.) of each key component, system and the whole machine of the electric aircraft, the health assessment and fault diagnosis / prediction module A predicts the potential faults of each centralized object system in the order of the system fault prediction unit A23 The whole machine fault prediction unit A13, and predicts the potential faults of each distributed object system in the order of the key component fault prediction unit A33 The system fault prediction unit A23 The whole machine fault prediction unit A13.

[0101] Based on the results of the health assessment and fault diagnosis / prediction module A, the dynamic optimization management module B performs dynamic optimization management on the management object and adopts a two-layer control structure.

[0102] Specifically, as Figure 5 shown, the dynamic optimization management module B has a power prediction unit B1, an energy optimization unit B2 and a dynamic control unit B3.

[0103] The power prediction unit B1 in the lowest-level management unit in different management architectures (i.e., the key component management department 121b of the distributed system energy management architecture 121, and the centralized management department 122a of the centralized system energy management architecture 122) estimates the change curve of the estimated power demand of its respective management object based on the influence of factors including but not limited to preset working requirements, performance degradation, and uncertainties on the future operation trend of its respective management object. The higher-level management unit (the centralized management department 121a of the distributed system energy management architecture 121, which is higher-level compared to the key component management department 121b; and the overall machine management unit 140, which is higher-level compared to the centralized management department 122a of the centralized system energy management architecture 122 and the key component management department 121b (and the centralized management department 121a) of the distributed system energy management architecture 121) solves the change curve of the power demand of the management object at this level based on the connection relationship of the lower-level management object and the change curves of the power demand predicted by the power prediction units B1 of each management unit at the lower level, and finally uses the change curve of the power demand at the overall machine level as the estimated operating power at this level.

[0104] The energy optimization unit B2 refers to the estimated operating power of each management object, takes the energy utilization rate, efficiency, cost, reliability, safety, etc. of the management object as optimization goals, and uses the current and predicted health status information of the management object obtained by the health assessment and fault diagnosis / prediction module A and the flight operation requirements as constraint conditions to solve the operating power (optimized operating power) with the optimal comprehensive performance for each management object under the predicted power demand. At this time, the highest-level management unit in different management architectures combines the current and predicted health status information of the management object obtained by the health assessment and fault diagnosis / prediction module A to optimize the estimated operating power to obtain the respective optimized operating powers of the highest-level management units. The lower-level management units decompose the optimized operating power of the higher level into the estimated operating powers of the energy optimization units B2 of each lower-level management unit according to the connection relationship of the management objects at this level, and use this to optimize and solve the optimized operating powers of the lower-level management units.

[0105] The dynamic control unit B3 can output control commands in two different modes: the rule-based control mode ( Figure 5 denoted by C1 in Figure 5 ) and the energy optimization mode ( denoted by C2 in

[0106] The rule-based control mode C1 controls the working power of each management object according to established rules (e.g., the "worst-case" rule). Since the aging status of each electrified component is not predicted, the working power determined by the rule-based control mode C1 is not the working power with the optimal comprehensive performance for each management object due to the different aging degrees of each management object.

[0107] On the contrary, the energy optimization mode C2 predicts the power demand and makes optimal control of the working power of each management object by using the prediction of the aging status of each management object obtained by the key component fault prediction unit A33, the system fault prediction unit A23, and the whole machine fault prediction unit A13 of the health assessment and fault diagnosis / prediction module A. As described above, the energy optimization mode C2 is the working power with the optimal comprehensive performance for each management object.

[0108] The rule-based control mode C1 and the energy optimization mode C2 output control signals at different frequencies (time sequences), and realize real-time control of the whole machine, system, and key components through the integration of different control mode time sequences. Preferably, for every N times ( Figure 6 for example, 3 times in (a) of) the control signal of the rule-based control mode C1 is output, the control signal of the energy optimization mode C2 is output once at the same time sequence. Then, as Figure 6 shown in (c), the dynamic control unit B3 preferentially adopts the output of the control signal of the energy optimization mode C2 when the control signals of the energy optimization mode C2 and the rule-based control mode C1 are output simultaneously, and controls the whole machine and system with the rule-based control mode C1 during the interval between the output of the control signals of the energy optimization mode C2. In this way, the control cycle of the electric aircraft can be shortened, the response speed of the aircraft to different working conditions can be improved, and the negative impacts brought by the defects such as high computational complexity and long time consumption existing in the energy optimization mode C2 can be avoided, meeting the dynamic performance requirements of the electric aircraft and its system.

[0109] Next, the specific algorithm when the energy optimization mode C2 is executed will be described.

[0110] Considering the unity of health assessment and fault prediction, the system and whole machine health assessment units A21, A11 and the system and whole machine fault prediction units A23, A13 use unified characteristic parameters to reflect the aging degree of the electric aircraft. In this specification, the health index HI is used as a unified index to measure the health status of different systems and the whole machine.

[0111] In the electric aircraft energy dynamic optimization management system 100 of the present invention, the health assessment and fault diagnosis / prediction functions in each key component management unit 121b can be expressed by formulas (1) and (2).

[0112] HeI i = f i (Para i ) (1)

[0113] Para i = g i (t) (2)

[0114] Wherein:

[0115] f i represents the health index HeI of the key component i i and the variation relationship between it and the characteristic parameter Para i can reflect the current health state of the key component i;

[0116] g i predicts the variation relationship between the characteristic parameter Para of the key component based on the principle and a large amount of experimental data i and the running time t, which can reflect the aging characteristics of the key component.

[0117] That is to say, in the electric aircraft energy dynamic optimization management system 100, the health state of the system is closely related to each component including the key components in the system. The change of the component performance state will affect other components that have physical or information connection with it, and then affect the performance of the entire system. Therefore, considering the structural characteristics of the system, in the present invention, with connectivity analysis as the main body, a fault propagation model F j of the system j is established, and then according to the fault propagation model F j , from the component health state (HeI1,..., HeI i ,..., HeI M , 1 < i < M, M is the number of key components in the system j), the health state HI j of the system j is calculated, so as to comprehensively reflect the health state HI j of the system and the components, as shown in the following formula (3).

[0118] HI j = F j (HeI1,…, HeI i ,…, HeI M ) (3)

[0119] Furthermore, for each system, in the same way, from the health state of the system (HI1,..., HI i ,..., HI N , 1 < j < N, N is the number of systems in the electric aircraft), the health state HI of the whole electric aircraft is calculated, as shown in the following formula (4).

[0120] HI = F(HI1, …, HI j , …, HI N ) (4)

[0121] Fault prediction adds the time - series information of the health index on the basis of health assessment, that is, formula (2), and calculates the change trend curve of the health index over time successively from key components, systems to the whole machine.

[0122] For the lowest - level management unit (the key component management department 121b) in different management architectures (taking the architecture with the key component management department 121b as an example), the power prediction department B1 in the key component management department 121b comprehensively considers the influence of factors including but not limited to preset work requirements, performance degradation, and uncertainties on the future operation trend of the management objects at this level, and estimates the change curve of the estimated power requirements of their respective management objects.

[0123] Correspondingly, the centralized management department 121a in the distributed system energy management architecture 121, as a management unit at a higher level compared to the key component management department 121b, solves the change curve of the power requirements of the management objects at this level according to the connection relationship of the management objects at the next lower level and the change curve of the power requirements estimated by the power prediction department B1 of the key component management department 121b. Then, combined with the change curve of the power requirements estimated by the power prediction department B1 of the centralized management department 121a, it solves the change curve of the power requirements of the management objects at the whole - machine level. This is the estimated operating power of the management objects at the whole - machine level.

[0124] Next, the energy optimization department B in the whole - machine management unit 140 uses the change curve of the power requirements of the management objects at the whole - machine level as the estimated operating power, and optimizes the estimated operating power in combination with the current and predicted health status information of the management objects obtained by the health assessment and fault diagnosis / prediction module to obtain the optimized operating power of the whole - machine management unit 140. Then, the centralized management departments 121a and 122a decompose the optimized operating power of the whole - machine management unit 140 into the estimated operating power of the energy optimization departments B2 of the centralized management departments 121a and 122a at the lower level (system level) according to the connection relationship between systems, and optimize and solve the optimized operating power of the centralized management departments 121a and 122a at the lower level (system level). Then, it is further decomposed into the estimated operating power of the energy optimization department B2 of the key component management department 121b at the lower - level (component level).

[0125] The energy optimization unit B2 in the key component management unit 121b analyzes the relationship between the input / output power of each key component and the change of its characteristic parameters based on mechanism analysis, and then combines the relationship between the health status index and the characteristic parameters to establish a power optimization model for the key components, so as to optimize the power of the key components.

[0126] More specifically, for the power prediction and optimization of key components, a physical model of each key component needs to be established, the relationship between its power and the degree of aging is analyzed, characteristic parameters that can reflect power requirements are extracted, and the power requirement PPowr of key component i is established i and the characteristic parameter Para i function, as shown in formula (5), and combined with formula (2), predict the future power requirements of key components.

[0127] PPow i = h i (Para i ) (5)

[0128] The power prediction unit B in each object system can use the Monte Carlo method to simulate the power combinations under different operating conditions of the electric aircraft, and then use the support vector machine-based method to learn the simulated system power data, so as to provide a basis for realizing the energy optimization function.

[0129] The specific steps are as follows:

[0130] Analyze the possible change range and its probability distribution of each variable factor under the system operating conditions, construct a probability model or a stochastic model of the system power combination, and generate random numbers on the computer according to the distribution of each random variable in the model to achieve a sufficient number of random numbers required for one simulation process;

[0131] According to the distribution characteristics of the random variables, use the stratified sampling and importance sampling methods to sample each random variable, conduct multiple simulation and simulation experiments, statistically analyze the results of the simulation experiments, and obtain the probability solution or stochastic solution of the system power and the accuracy estimation of the solution;

[0132] Construct a data sample according to the data obtained from the simulation, select the kernel function and the penalty constant, obtain and solve the dual optimization problem, obtain the corresponding optimal solution, and select a component that satisfies the penalty constant range from the optimal solution, calculate the decision function, and then the support vector machine model for system power prediction can be obtained;

[0133] Considering the uncertainty of the load operation, taking the peak power of the high-power pulse load as the power prediction error, add it to the system power prediction model, and finally obtain the predicted power PPwr of system j j .

[0134] PPwrj = H j (PPowr1, …, PPowr i , …, PPowr M ) (6)

[0135] Next, using the predicted power of each system (PPwr1, ..., PPwr i , ..., PPwr N ), the predicted power PPwr of the entire electric aircraft is calculated and obtained.

[0136] PPwr = H(PPwr1, …, PPwr j , …, PPwr N ) (7)

[0137] The energy optimization function implemented in the energy optimization unit B2 adopts the rolling horizon method to achieve Figure 7 , Figure 8 the optimization process shown in, where Figure 7 , Figure 8 the upper half shows the relationship between power and time. Among them, the moment k is the current moment. To the left (e.g., moments k - 1, k - 2) are the past moments, and to the right (e.g., moments k + 1, k + 2, …… k + p) are the future moments (i.e., the prediction horizon). In addition, the thick solid line depicts the actual operating power of the electric aircraft at the past moments, the dotted line depicts the estimated operating power obtained according to the actual operating conditions (e.g., pilot commands, etc.), and the dashed line depicts the optimized operating power based on the current health state of the management object.

[0138] As shown in Figure 5 and Figure 7 , at a certain moment k during the aircraft operation, the power prediction unit B1 of the dynamic optimization management module B obtains the power demand at this time (depicted by the dotted line). Considering the system power balance, the current health state, and the output of the future power prediction, the energy optimization unit B2 uses intelligent optimization algorithms (such as NSGA - III, NSGA - II, stochastic programming, etc.) to search for the optimal energy optimization control strategy for the entire energy system, and obtains the control instruction sequence for the future moments k + 1, k + 2, ..., k + c (also known as the "control horizon"). The first value of the control instruction sequence is used as the control signal for the controlled object and is maintained until the moment k + 1. At the moment k + 1, as shown in Figure 8 , the current and future states of the system are updated, and the above steps are repeated. It can be seen that the prediction horizon p and the control horizon c are continuously rolling forward on the time axis, and the dynamic optimization management module B executes the above process at each moment.

[0139] The optimization model of the energy optimization unit B2 can be represented, for example, by the mathematical model shown in the following formula (8), where it includes the objective function of energy optimization control (including but not limited to fuel consumption FC, failure rate FR, efficiency η, response speed RT, etc.) and constraint conditions (including but not limited to power prediction curve PPwr, health state HI, node power balance matrix Pwr, etc.). The specific modeling process will not be elaborated here. Based on formula (8), the energy optimization unit B2 can solve the current optimal control strategy by means of optimization algorithms such as NSGA-III.

[0140]

[0141] Those skilled in the art can easily think of other advantages and modifications. Therefore, in a broader sense, the present invention is not limited to the specific details and representative embodiments shown and described here. Therefore, modifications can be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.

Claims

1. An electric aircraft energy dynamic optimization management system for optimizing, managing, and controlling the energy system of an electric aircraft, characterized in that the electric aircraft energy dynamic optimization management system includes: a monitoring module that collects energy system parameters in real time and extracts characteristic parameters related to the electric aircraft energy system therefrom; multiple management units at different levels that manage multiple management objects at different levels of the electric aircraft, each level of management unit has a health assessment and fault diagnosis / prediction module and a dynamic optimization management module, the health assessment and fault diagnosis / prediction module evaluates the health status of each management object at the current moment, calculates the health index of the electric aircraft at the current moment in a bottom-up manner, and for the health indices calculated for each management object at each moment before the current moment, fits the change trend curve of the health index of each management object over time by adding time series information, and predicts the health status of the electric aircraft at each moment after the current moment in a bottom-up manner, the dynamic optimization management module can execute a rule-based control mode and an energy optimization mode on the electric aircraft energy system at different time series, and integrates the time series of different control modes by adding the time series of the rule-based control mode between two adjacent time series when executing the energy optimization mode to achieve real-time control of multiple management objects at different levels, when the dynamic optimization management module is in the time series of executing the energy optimization mode, based on the health status of the electric aircraft predicted by the health assessment and fault diagnosis / prediction module after the current moment, it performs control with better comprehensive performance for each management object compared to the rule-based control mode, where the rule-based control mode controls the working power of each management object according to established rules, the energy optimization mode predicts the power demand and optimally controls the working power of each management object by using the prediction of the aging condition of each management object as the health status obtained by the health assessment and fault diagnosis / prediction module.

2. The electric aircraft energy dynamic optimization management system according to claim 1, characterized in that the monitoring module includes a sensor network module, a data processing module, and a feature extraction module, the sensor network module is used to collect the original data samples of the electric aircraft, the data processing module preprocesses the original data samples to filter out interference signals, the feature extraction module extracts characteristic parameters from the preprocessed parameters that can independently or mutually confirm to characterize a certain / some states / characteristics of the electric aircraft to achieve data dimensionality reduction.

3. The electric aircraft energy dynamic optimization management system according to claim 1, characterized in that the electric aircraft energy dynamic optimization management system includes: an overall aircraft management unit that manages the electric aircraft as a whole as a management object; Multiple system-level management modules, where the system-level management module has a centralized management unit that manages the object systems constituting the entire electric aircraft as management objects, and at least one of the multiple system-level management modules adopts a distributed system energy management architecture with a two-layer structure of centralized + distributed. In the distributed system energy management architecture, there is a critical component management unit that manages the critical components constituting the object system as management objects. The multiple management units are the overall aircraft management unit, the centralized management units in each system-level management module, and the critical component management units in the system-level management modules adopting the distributed system energy management architecture.

4. The electric aircraft energy dynamic optimization management system according to claim 3, characterized in that A part of the multiple system-level management modules adopts a centralized system energy management architecture. The multiple system-level management modules adopting the distributed system energy management architecture and the centralized system energy management architecture respectively implement information interaction between each object system and the entire electric aircraft with the overall aircraft management unit through the overall aircraft communication module.

5. The electric aircraft energy dynamic optimization management system according to claim 3, characterized in that In the system-level management module adopting the distributed system energy management architecture, the centralized management unit and the critical component management unit implement information interaction with each other through the system communication module.

6. The electric aircraft energy dynamic optimization management system according to claim 1, characterized in that The health assessment and fault diagnosis / prediction module has a health assessment unit for implementing the health assessment function, a fault diagnosis unit for implementing the fault diagnosis function, and a fault prediction unit for implementing fault prediction. The health assessment unit assesses the health status of the management object at the current level based on the health status information of the management object at the lower level as the management object, and transmits the assessment result to the upper level. Once the health assessment unit finds that the health status of the management object has declined, the fault diagnosis unit traces and isolates the fault in a top-down manner by level. Based on the analysis results of the health assessment unit and the fault diagnosis unit, the fault prediction unit predicts the change trend of the health status of each management object level by level from bottom to top, and outputs the aging characteristic parameters.

7. The electric aircraft energy dynamic optimization management system according to claim 1, characterized in that The dynamic optimization management module of the management unit at each level has a power prediction unit. The power prediction units in the management units at the lowest level under different architectures comprehensively consider the influence of factors including preset working requirements, performance degradation, and uncertainty of the management objects at this level on the future operation trend, and estimate the change curve of the power demand of their respective management objects. The management units at a higher level than the lowest level solve the change curve of the power demand of the management objects at this level according to the connection relationship of the management objects at the lower level and the change curves of the power demands predicted by the power prediction units of the management units at the lower level.

8. The dynamic optimization management system for the energy of an electric aircraft according to claim 7, wherein the dynamic optimization management module of the management unit at each level has an energy optimization section, the energy optimization section determines the optimal comprehensive performance as the optimization goal with reference to the power requirements of each management object, and uses the current and predicted health status information of the management object and the operation requirements of the management objects at this level as constraint conditions to solve for the working power with the optimal comprehensive performance for each management object under the power requirements, that is, the optimized operation power.

9. The dynamic optimization management system for the energy of an electric aircraft according to claim 8, wherein the management unit at the highest level under different management architectures uses the change curve of the power requirements of the management objects at this level as the estimated operation power of the management unit, and optimizes the estimated operation power in combination with the current and predicted health status information of the management object obtained by the health assessment and fault diagnosis / prediction module to obtain the optimized operation power of each management unit at the highest level, the management units at lower levels compared to the highest level solve for the estimated operation power of the management units at the next lower level based on the connection relationships of the management objects at the lower level through the optimized operation power optimized by the energy optimization sections of the management units at the higher level, and then solve for the optimized operation power of each management unit at the next lower level.

10. The dynamic optimization management system for the energy of an electric aircraft according to claim 9, wherein the dynamic optimization management module of the management unit at each level further has a dynamic control section, when the control signals are output simultaneously in the energy optimization mode and the rule-based control mode, the dynamic control section preferentially adopts the output of the control signal in the energy optimization mode, and controls the whole aircraft and the system in the rule-based control mode during the cycle interval of the output of the control signal in the energy optimization mode.

Images