WO2024016966A1 - 一种电力电子产品的多层控制系统及方法 - Google Patents

一种电力电子产品的多层控制系统及方法 Download PDF

Info

Publication number
WO2024016966A1
WO2024016966A1 PCT/CN2023/103039 CN2023103039W WO2024016966A1 WO 2024016966 A1 WO2024016966 A1 WO 2024016966A1 CN 2023103039 W CN2023103039 W CN 2023103039W WO 2024016966 A1 WO2024016966 A1 WO 2024016966A1
Authority
WO
WIPO (PCT)
Prior art keywords
layer
power electronic
model
signal
electronic product
Prior art date
Application number
PCT/CN2023/103039
Other languages
English (en)
French (fr)
Inventor
刘宁
林霖
单忠伟
孔令伟
王振世
马艳丽
李多强
Original Assignee
联合汽车电子有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 联合汽车电子有限公司 filed Critical 联合汽车电子有限公司
Publication of WO2024016966A1 publication Critical patent/WO2024016966A1/zh

Links

Classifications

    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00Programme-control systems
    • G05B19/02Programme-control systems electric
    • G05B19/04Programme control other than numerical control, i.e. in sequence controllers or logic controllers
    • G05B19/042Programme control other than numerical control, i.e. in sequence controllers or logic controllers using digital processors
    • G05B19/0423Input/output
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00Program-control systems
    • G05B2219/20Pc systems
    • G05B2219/24Pc safety
    • G05B2219/24215Scada supervisory control and data acquisition

Definitions

  • the present invention relates to the field of power electronic products, and in particular to a multi-layer control system and method for power electronic products.
  • the present invention provides a multi-layer controller for power electronic products. Control systems and methods.
  • the multi-layer control system for power electronic products at least includes: a first-layer controller and a second-layer controller; wherein the first-layer controller is configured to: obtain the power electronic product Send the first status signal to the second-layer controller according to the first status signal detected by the sensor; and receive the control reference signal fed back by the second-layer controller, and generate the control reference signal based on the control reference signal.
  • the control signal of the power electronic product is output to the power electronic product;
  • the second layer controller is configured to: obtain the power electronic product according to the first status signal and a model that simulates the power electronic product. a second status signal other than the first status signal; and generating the control reference signal according to the first status signal, the second status signal and the control target of the power electronic product, and outputting it to the Describe the first layer controller.
  • the multi-layer control system for power electronic products can assist power electronics based on an external controller (second-layer controller) without adding additional hardware or additional costs.
  • the product's controller (first-layer controller) can generate more optimized control parameters, thereby improving the control of power electronic products, enriching the functions of power electronic products, and improving the performance of power electronic products.
  • a third-layer controller is also included; wherein the third-layer controller is configured to: based on the historical first state fed back by the first-layer controller The signal and the historical control signal, the historical second state signal fed back by the second-layer controller, and the historical control reference signal identify and optimize the model parameters of the model, and output the model parameters to the second-layer controller; The second layer controller is further configured to update the model based on the received model parameters.
  • the third-layer controller is further configured to: in response to the change of the model parameter exceeding the preset threshold, change the model that changes beyond the preset threshold.
  • the parameters are sent to the first layer controller, so that the first layer controller diagnoses the hardware status of the power electronic product based on the model parameters.
  • the model for simulating the power electronic product in the second-layer controller includes: a mechanism model, and/or a big data model; wherein the mechanism model In order to simulate the mathematical physical model of the power electronic product, the simulated state variables of the power electronic product other than the first state signal are obtained based on the first state signal as the second state signal; the big data
  • the model represents a mapping relationship between the first state signal and/or the simulated state variable of the power electronic product and the expected state variable of the power electronic product, so as to determine the relationship between the first state signal and/or the simulated state variable and the expected state variable of the power electronic product.
  • the expected state variable outputted by the state variable is the second state signal.
  • the third-layer controller is further configured to: identify and optimize model parameters of the mathematical physical model and/or the big data model, and output The model parameters are passed to the second layer controller.
  • the second-layer controller in response to the model including the mechanism model and the big data model, is further configured to: according to the first The state signal and the mechanism model obtain the simulated state variable; input the first state signal and the simulated state variable into the big data model to obtain the expected state variable as the second state signal; and
  • the control reference signal is generated according to the second status signal output by the big data model.
  • the multi-layer control system is applied to vehicles,
  • the first-layer controller is a power electronic product controller of the vehicle,
  • the second-layer controller is an on-board computing unit of the vehicle, and
  • the third-layer controller is one-to-one corresponding to the vehicle. Cloud controller.
  • the multi-layer control system is applied in the vehicle field. Due to the rapid development of vehicle electronic and electrical architecture and intelligent network technology, vehicles have appeared on-board domain controllers, vehicle-mounted area controllers and vehicle-mounted computers.
  • the computing power calculation unit can use these controllers on the vehicle as external controllers for on-board power electronics products, so that it can be based on the vehicle electrical architecture already installed on the vehicle without increasing the cost of the vehicle and parts. Enrich the functions of power electronic products, improve the performance of power electronic products, and provide more benefits for the entire vehicle.
  • resources other than the controller product can be used to optimize the performance (control accuracy) of the controller product (the chip used by the controller remains unchanged). , system power, system efficiency and robustness, etc.), a cost-reducing solution to achieve high control performance of controller products.
  • the multi-layer control method includes: obtaining a first status signal of the power electronic product based on sensor detection by a first-layer controller;
  • the layer controller obtains a second state signal of the power electronic product other than the first state signal based on the first state signal and a model that simulates the power electronic product;
  • the second layer controller obtains a second state signal of the power electronic product according to the first state signal.
  • the first state signal, the second state signal and the control target of the power electronic product generate a control reference signal; and the first layer controller generates a control signal of the power electronic product according to the control reference signal. signals and output to the power electronic products.
  • the multi-layer control method further includes: a third-layer controller based on the historical first status signal and the historical control signal fed back by the first-layer controller. , identifying and optimizing the model parameters of the model based on the historical second state signal and the historical control reference signal fed back by the second layer controller; and causing the second layer controller to update the model based on the model parameters.
  • the multi-layer control method further includes: in response to the change of the model parameter exceeding a preset threshold, the third layer controller controls the change beyond the preset threshold. Model parameters with threshold changes are sent to the first-layer controller, so that the first-layer controller diagnoses the hardware status of the power electronic product based on the model parameters.
  • the model for simulating the power electronic product in the second-layer controller includes: a mechanism model, and/or a big data model; wherein the mechanism model In order to simulate the mathematical physical model of the power electronic product, the simulated state variables of the power electronic product other than the first state signal are obtained based on the first state signal as the second state signal; the big data
  • the model represents a mapping relationship between the first state signal and/or the simulated state variable of the power electronic product and the expected state variable of the power electronic product, so as to determine the relationship between the first state signal and/or the simulated state variable and the expected state variable of the power electronic product.
  • the expected state variable outputted by the state variable is the second state signal.
  • the second layer controller acquiring the second status signal further includes: Obtain the simulated state variable according to the first state signal and the mechanism model; convert the first state The signal and the simulated state variable are input into the big data model to obtain the expected state variable as the second state signal; wherein the control reference signal based on which the second layer controller generates the control reference signal is The second status signal output by the big data model.
  • Another aspect of the present invention also provides a computer-readable storage medium that stores a computer program.
  • the computer program is executed by a processor, the multi-layer control method for power electronic products as described in any of the above embodiments is implemented. A step of.
  • power electronics and their controllers can have richer functions and advantages based on existing external controllers without additional hardware costs. performance.
  • the present invention can be applied in the vehicle field, and can realize the function and performance optimization of vehicle-mounted power electronic products and their controllers through the vehicle electronic and electrical architecture and intelligent network connection technology without increasing the hardware cost of the entire vehicle, and improve the overall performance of the vehicle. vehicle performance, providing more benefits for the entire vehicle.
  • Figure 1 shows a schematic structural diagram of a multi-layer control system for power electronic products provided by one aspect of the present invention.
  • FIG. 2 shows a schematic structural diagram of a preferred embodiment of the multi-layer control system for power electronic products provided by one aspect of the present invention.
  • FIG. 3 shows a schematic flowchart of a multi-layer control method for power electronic products provided by one aspect of the present invention.
  • the present invention provides a multi-layer controller for power electronic products. Control systems and methods.
  • the multi-layer control system for power electronic products at least includes a first-layer controller 100 and a second-layer controller 200.
  • the first-layer controller 100 and the second-layer controller 200 are The controllers 200 have communication connections with each other. There are electrical connections and/or communication connections between the first-layer controller 100 and the control objects (ie, power electronic products, including but not limited to electric shafts, chargers, DCDC modules, etc.).
  • the first layer controller 100 can obtain the first status signal detected by the control object based on the sensor, and output the first status signal to the second layer controller 200 .
  • the first state signal is an electrical signal sampled based on actual physical sensors, including but not limited to temperature, voltage, current, etc.
  • the first-layer controller 100 has common control and computing capabilities and is mainly used to run software processes with high real-time performance, low computing power requirements, low data storage, and simple control algorithms. It can output control signals of control objects, usually PWM. duty cycle signal.
  • the second-layer controller 200 is mainly used to run software processes with medium real-time performance, high computing power requirements, low data storage, and intelligent optimization algorithms.
  • the second-layer controller 200 can, based on the first status signal and the model of simulated power electronic products, Obtain a second status signal of the power electronic product in addition to the first status signal.
  • the second state signal refers to system state variables that are difficult to observe or measure through physical sensors, including but not limited to variables such as the temperature inside the control object that are difficult to characterize. In the second-layer controller 200, these state quantities that are difficult to obtain through observation or sensor measurement can be obtained by simulating the model of the power electronic product.
  • the second layer controller 200 can optimize certain targets based on the second status signal. According to the control target, the control reference value of the first-layer controller 100 is calculated and output to the first-layer controller 100 .
  • the multi-layer control system also includes a third-layer controller 300.
  • the third-layer controller 300 has communication connections with the first-layer controller 100 and the second-layer controller 200 and can receive The historical first state signal and historical control signal output by the first layer controller 100 can receive the historical second state signal and historical control reference signal output by the second layer controller 200, and based on these historical signals, identify and optimize the second
  • the model parameters of the model in the layer controller 200 are used to enable the second layer controller to update the model, so that the state quantity of the controlled product can be better obtained based on model simulation.
  • the third-layer controller 300 is used to run software processes with slow real-time performance, high computing power requirements, massive data storage, and intelligent optimization algorithms, and is used to optimize or adjust model parameters involved in the second-layer controller. Through this distributed multi-layer control architecture, the functions of power electronic products can be enriched and the performance of power electronic products can be improved.
  • the first-layer controller 100 can be considered as a power electronic vehicle-mounted controller
  • the second-layer controller 200 can be considered as an on-board computing unit
  • the third-layer controller 300 can be considered as a cloud corresponding to the vehicle. controller.
  • FIG. 2 Please further understand a preferred multi-layer control system provided by one aspect of the present invention in conjunction with FIG. 2 .
  • a first layer controller 100 a second layer controller 200 and a third layer controller 300 are included.
  • the first layer controller 100 is used to realize the most basic control functions of the power electronic controller
  • the second layer controller 200 is used to achieve optimal optimization control of certain goals
  • the third layer controller 300 is used to optimize or adjust multiple layers. Control the model parameters involved in the second layer of the system. It should be noted that the third layer controller 300 is used to process massive data, similar to a cloud controller, but is slightly different from the current cloud.
  • the third layer controller 300 in the present invention is only oriented to a single power electronics Products, that is, each power electronic product has its own independent multi-layer control system first layer, second layer and third layer.
  • the first layer controller 100 further includes various internal sub-modules, such as data sampling 110, control algorithm 120, drive generation 130 and component diagnosis 140.
  • the second layer controller 200 further includes: an intelligent algorithm 210, a mechanism model 220 and a data model 230.
  • the third layer controller further includes: data storage 310, identification algorithm 320, parameter transmission 330, identification model 340 and data diagnosis 350.
  • the transmitted data includes the electrical signals related to the actual physical sensor sampling based on the first layer of the multi-layer control system. Including temperature, voltage, current, etc.
  • the signal received by the control object is the control signal output to the controller object from the first layer of the multi-layer control system.
  • it is generally a PWM duty cycle signal.
  • the first layer of the multi-layer control system sends the software sampling signal, software control signal and control-related intermediate quantities to the third layer of the multi-layer control system.
  • the first layer of the multi-layer control system sends software sampling signals, software control signals and control-related intermediate quantities to the second layer of the multi-layer control system.
  • the second layer of the multi-layer control system is based on the model and the data sent by 100->200 data. It can observe or estimate system state variables that cannot be measured by actual sensors, such as the temperature inside the control object, which is difficult to characterize. variables, etc., and send the observed system state variables to the third layer of the multi-layer control system.
  • the third layer of the multi-layer control system records the 100->300 data transmission and 200->300 data transmission of the entire life cycle through data storage 310, and passes the identification algorithm 320 and identification model based on long-term data.
  • 340 online identification of the model parameters used in the second layer of the multi-layer control system (since power electronic products are physical devices, their characteristics will change over time, so the model parameters in the second layer of the multi-layer control system also need to be changed), And the identified model parameters are sent to the model involved in the second layer of the multi-layer control system through parameter sending 330, and replace the current model parameters.
  • the second layer of the multi-layer control system is based on 100->200 data transmission. Based on the corresponding algorithm scheduling and model estimation and prediction, based on the premise of certain target optimization, the multi-layer control system is calculated through intelligent algorithm 210 The control reference value of the first layer of the control system is passed to the control algorithm 120 of the first layer of the multi-layer control system.
  • the control algorithm 120 can calculate relevant control signals of the control object based on common control and computing capabilities, and output them to the control object. In one embodiment, when applied to a vehicle, these control signals need to be converted into relevant driving signals through the driving generation 130 .
  • the third layer of the multi-layer control system records the 100->300 data transmission and 200->300 data transmission of the entire life cycle, and determines whether the control object is controlled based on long-term data through data diagnosis 350 and identification algorithm 320 Normally, if some parameters or state variables have a huge change trend under the same working conditions, the relevant data will be sent to the first layer of the multi-layer control system, and the component diagnosis 140 will perform relevant diagnosis and response.
  • the intelligent algorithm 210 is used to realize certain intelligent scheduling algorithms for target optimization, such as PSO particle swarm algorithm, simulated annealing algorithm, model prediction algorithm, etc.
  • the mechanism model 220 is a mathematical physical model of a reproducible control object, which is used to estimate relevant system state variables that cannot be measured in the control object.
  • Data model 230 Offline trained data model (big data model such as neural network model). After offline training with a large amount of data, when applying online, after inputting various state variables and sampling values of the control object, the desired results can be output.
  • big data model such as neural network model
  • the intelligent algorithm 210 transfers the information in the data sent from the first layer to the second layer to the mechanism model 220.
  • the mechanism model 220 observes or estimates the state variables inside the control object (contents that are difficult to actually measure). ).
  • the intelligent algorithm 210 sends relevant calling instructions to the mechanism model 220 to call the mechanism model 220 to obtain the second status signal.
  • the intelligent algorithm 210 passes the information sent from the first layer to the second layer to the data model 230.
  • the intelligent algorithm 210 sends relevant calling instructions to the data model 230 to call the data model 230 to obtain the first Two status signals.
  • the mechanism model 220 in response to including both the mechanism model 220 and the data model 230, can be first obtained to obtain the simulated state variables based on the first state signal, and then the data model 230 can be called to obtain the simulated state variables based on the first state signal and the simulated state.
  • the variable gets the expected state variable as the second state signal.
  • Data model 230 is based on 210->220 data sending and 220->230 data sending. It is based on a pre-trained data model (such as a neural network model). The output mechanism model is difficult to calculate or the calculation is inaccurate. data information and send it to the intelligent algorithm 210.
  • the mechanism model 220 is based on the 210->220 data sending, and based on the physical mathematical model, it observes or estimates the state variables inside the control object (content that is difficult to actually measure), and sends them to the data model intelligent algorithm 210 .
  • the intelligent algorithm 210 calculates the control reference of the control algorithm 120 based on the first state signal sent by the data sampling 140 and the second state signal obtained by the mechanism model 220 and/or the data model 230 based on certain target-optimized intelligent scheduling algorithms. Signal.
  • the control target is obtained in the second layer controller 200 . If applied to a vehicle, the control target is sent by the VCU of the vehicle.
  • the control target can be optimized through the intelligent algorithm 210, thereby generating a control reference signal to the control algorithm in the first layer controller 100.
  • the control algorithm 120 calculates the control signal based on the first state signal obtained by data sampling and then generates an output based on the drive. to the control object.
  • an identification algorithm based on a model (including a mechanism model and/or a big data model) needs to be used, so that the second parameter that cannot be obtained through observation or sensor sampling can be obtained based on the model.
  • the state signal is used to optimize the control target through an intelligent algorithm based on the first state signal and the second state signal.
  • the third-layer controller 300 is required to identify and optimize model parameters based on various historical data sent by the first-layer controller and the second-layer controller stored in the data storage through identification models and identification algorithms.
  • the relevant updated model parameters are output to the identification algorithm in the second-layer controller to complete the replacement and update of the model parameters.
  • the hardware performance of the physical device can be simulated through data diagnosis in the third-layer controller, which has changed significantly, it will affect the normal operation of the physical device. If used, it is necessary to promptly feed back parameters with large changes to the first-layer controller.
  • the component diagnosis of the first-layer controller is based on the model parameters identified by the third-layer controller to diagnose the power electronics. The hardware status of the product.
  • This invention utilizes the high computing power characteristics of the second layer in the distributed multi-layer control architecture to provide the first layer in the distributed multi-layer control architecture with corresponding control targets, restriction conditions, control modifications and other related information to improve the overall control software. performance.
  • the third layer in the distributed multi-layer control architecture is used to utilize the characteristics of massive data processing capabilities and intelligent algorithm iterative optimization capabilities to form the first layer in the distributed multi-layer control architecture.
  • Provide corresponding long-term monitoring and control related information and provide corresponding model parameter optimization, algorithm parameter optimization and other related information for the second layer in the distributed multi-layer control architecture to improve the overall control software performance.
  • the present invention can be especially applied in the field of vehicles.
  • the multi-layer architecture can correspond to the controller (first layer) and the vehicle-mounted calculator (second layer) of the vehicle-mounted power electronic products.
  • Layer 3 controllers can also correspond to the cloud.
  • the present invention can utilize resources other than the controller of power electronic products (other controllers on the vehicle, on-board computers, cloud and other resources) to optimize the performance (control accuracy, system power) of the controller (the chip used by the controller remains unchanged). , system efficiency and robustness, etc.), it is a cost-reduction solution that indirectly achieves high control performance of the controller.
  • controllers of current power electronic products are subject to cost constraints, it is difficult for the controller MCU to apply intelligent algorithms and complex models, making it difficult to perform performance optimization control.
  • resources outside the controller can be used to realize the performance optimization control of the controller through data transfer through external communication.
  • the hardware of the controller remains unchanged, that is, the cost remains unchanged. .
  • the traditional PI closed-loop control is defined as the first layer
  • the optimization control is defined as the second layer
  • the parameter identification and optimization of the model used in the optimization control is defined as the third layer. Due to the problem of the current controller using MCU resources, it can only complete the first layer, and it is difficult to complete the second and third layers. Since the second layer has to constantly set the control reference value for the first layer, it has a certain impact on the communication time. There are certain requirements, so the second layer needs to be placed in the car (a controller that can support SOA services). The third layer has no communication time requirements, but requires a large amount of data storage, so it can be in the car or in the cloud.
  • software functions of power electronic products can be enriched, software performance of power electronic products can be improved, performance requirements of software control chips of power electronic products can be reduced, and product costs can be reduced.
  • Another aspect of the present invention also provides a computer-readable storage medium that stores a computer program.
  • the steps of the multi-layer control method for power electronic products as described in any of the above embodiments are implemented.
  • the above-mentioned computer-readable storage medium can also be in the form of a system, that is, it includes multiple computer-readable storage sub-media (corresponding to multiple layers), so as to jointly realize the above-described process through multiple computer-readable storage media. Steps in a multi-layer control approach for power electronics.
  • DSP digital signal processor
  • ASIC application specific integrated circuit
  • FPGA field programmable gate array
  • a general purpose processor may be a microprocessor, but in the alternative the processor may be any conventional processor, controller, microcontroller, or state machine.
  • a processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.
  • the steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of both.
  • Software modules may reside in RAM memory, Flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
  • An example storage medium is coupled to the processor such that the processor can read and write information from/to the storage medium.
  • the storage medium may be integrated into the processor.
  • the processor and storage media can reside in an ASIC.
  • the ASIC can reside in the user terminal.
  • the processor and storage medium may reside as discrete components in the user terminal.
  • the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium.
  • Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. Storage media can be any available media that can be accessed by a computer.
  • such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or may be used to carry or store instructions or data structures in the form of Any other medium that contains program code and can be accessed by a computer.
  • any connection is also properly termed a computer-readable medium.
  • the Software is transmitted from a web site, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL), or wireless technologies such as infrared, radio, and microwave
  • coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave
  • disk and disc include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, among which disk is often reproduced in a magnetic way. data, while discs use lasers to optically reproduce data. Combinations of the above should also be included within the scope of computer-readable media.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

一种电力电子产品的多层控制系统及方法。多层控制系统至少包括:第一层控制器(100)和第二层控制器(200)。第一层控制器(100)获取电力电子产品根据传感器检测得到的第一状态信号,发送第一状态信号至第二层控制器(200);以及接收第二层控制器(200)反馈的控制参考信号,基于控制参考信号生成电力电子产品的控制信号。第二层控制器(200)根据第一状态信号和模拟电力电子产品的模型,获取电力电子产品除第一状态信号以外的第二状态信号;以及根据第一状态信号、第二状态信号和电力电子产品的控制目标,生成控制参考信号并输出至第一层控制器(100)。根据多层控制系统及方法,能够借由不同层级的控制器扩宽电力电子产品的控制功能、提高电力电子产品的控制性能。

Description

一种电力电子产品的多层控制系统及方法 技术领域
本发明涉及电力电子产品领域,尤其涉及电力电子产品的多层控制系统及方法。
背景技术
随着电气化、自动化的加速,对于电力电子产品的功能和性能也有着更高的要求。电力电子产品能否发挥最佳功能和性能,与其控制有着密不可分的关系。通常来说,在投入使用时,电力电子产品和控制器是一一对应的。受制于控制器的软、硬件限制,即便电力电子产品在硬件上可以发挥更好的功能和性能,由于对应的控制器缺乏给出优选的控制参数的能力,无法将电力电子产品的功能发挥到最佳。
随着电子产品的不断发展,通常来说,一个产品上除了电力电子产品及其控制器外,还会搭配有额外的控制器,甚至还可以连接云端的控制器。在不考虑替换电力电子产品配套的控制器的情况下(不增加成本),希望能够借助于外部的控制器来辅助电力电子产品的控制器给出优选的控制参数。
这一需求在电动汽车领域也十分普遍。随着电动汽车市场渗透率的不断提高,电动汽车技术得到快速的发展,整车对零部件(电力电子产品)的要求也越来越高,需要具备更多的功能和更好的性能。另一方面,整车电子电器架构和智能网联技术也得到了快速发展,出现了域控制器、区域控制器以及电脑等高算力计算单元,可以作为外部控制器。
有鉴于此,希望能够提供一种电力电子产品的多层控制系统及方法,从而能够在不增加零部件成本的前提下,让电力电子产品及其控制器具备更加丰富的功能和更加优越的性能,同时,也具有非常广泛的应用前景,如电动汽车市场。
发明内容
以下给出一个或多个方面的简要概述以提供对这些方面的基本理解。此概述不是所有构想到的方面的详尽综览,并且既非旨在指认出所有方面的关键性或决定性要素亦非试图界定任何或所有方面的范围。其唯一的目的是要以简化形式给出一个或多个方面的一些概念以为稍后给出的更加详细的描述之序。
如上文所描述的,为了解决现有技术中,电力电子产品的控制器在不增加额外成本的情况下,无法提供更优的控制参数的问题,本发明提供了一种电力电子产品的多层控制系统及方法。
本发明的一方面所提供的电力电子产品的多层控制系统至少包括:第一层控制器和第二层控制器;其中所述第一层控制器被配置为:获取所述电力电子产品 根据传感器检测得到的第一状态信号,发送所述第一状态信号至所述第二层控制器;以及接收所述第二层控制器反馈的控制参考信号,基于所述控制参考信号生成所述电力电子产品的控制信号,并输出至所述电力电子产品;所述第二层控制器被配置为:根据所述第一状态信号和模拟所述电力电子产品的模型,获取所述电力电子产品除所述第一状态信号以外的第二状态信号;以及根据所述第一状态信号、所述第二状态信号和所述电力电子产品的控制目标,生成所述控制参考信号,并输出至所述第一层控制器。
根据本发明的一方面所提供的电力电子产品的多层控制系统,能够在不增加额外的硬件、不增加额外成本的情况下,基于外部的控制器(第二层控制器),辅助电力电子产品的控制器(第一层控制器)能够生成更优化的控制参数,从而改进电力电子产品的控制,丰富电力电子产品的功能、提升电力电子产品的性能。
在上述多层控制系统的一实施例中,可选的,还包括第三层控制器;其中所述第三层控制器被配置为:根据所述第一层控制器反馈的历史第一状态信号和历史控制信号、所述第二层控制器反馈的历史第二状态信号和历史控制参考信号辨识并优化所述模型的模型参数,并输出所述模型参数至所述第二层控制器;所述第二层控制器还被配置为:基于所接收到的模型参数更新所述模型。
在上述多层控制系统的一实施例中,可选的,所述第三层控制器还被配置为:响应于所述模型参数的变化超出预设阈值,将变化超出预设阈值变化的模型参数发送至所述第一层控制器,以使所述第一层控制器基于所述模型参数诊断所述电力电子产品的硬件状态。
在上述多层控制系统的一实施例中,可选的,所述第二层控制器中模拟所述电力电子产品的模型包括:机理模型,和/或,大数据模型;其中所述机理模型为模拟所述电力电子产品的数学物理模型,以基于所述第一状态信号获取所述电力电子产品除所述第一状态信号以外的模拟状态变量为所述第二状态信号;所述大数据模型表征所述电力电子产品的第一状态信号和/或所述模拟状态变量与所述电力电子产品的预期状态变量之间的映射关系,以根据所述第一状态信号和/或所述模拟状态变量输出所述预期状态变量为所述第二状态信号。
在上述多层控制系统的一实施例中,可选的,所述第三层控制器进一步被配置为:辨识并优化所述数学物理模型和/或所述大数据模型的模型参数,并输出所述模型参数至所述第二层控制器。
在上述多层控制系统的一实施例中,可选的,响应于所述模型包括所述机理模型和所述大数据模型,所述第二层控制器进一步被配置为:根据所述第一状态信号和所述机理模型获取所述模拟状态变量;将所述第一状态信号和所述模拟状态变量输入所述大数据模型,以获取所述预期状态变量为所述第二状态信号;以及根据所述大数据模型输出的第二状态信号生成所述控制参考信号。
在上述多层控制系统的一实施例中,可选的,所述多层控制系统应用于车辆, 所述第一层控制器为所述车辆的电力电子产品控制器,所述第二层控制器为所述车辆的车载计算单元,所述第三层控制器为与所述车辆一一对应的云端控制器。
在上述的实施例中,多层控制系统应用在车载领域,由于整车电子电器架构和智能网联技术得到了快速发展,车辆上出现了车载域控制器、车载区域控制器以及车载电脑等高算力计算单元,可以利用整车上的这些控制器作为车载电力电子产品的外部控制器,从而能够在不增加整车及零部件成本的前提下,基于车辆上已经搭载的整车电气架构,丰富电力电子产品的功能、提升电力电子产品的性能,为整车提供更多收益。在上述的实施例中,能够利用控制器产品之外的资源(车上其他控制器、车载电脑、云端等资源),来优化控制器产品(控制器使用芯片保持不变)的性能(控制精度、系统功率、系统效率及鲁棒性等),实现控制器产品高控制性能的降本方案。
本发明的另一方面还提供了一种电力电子产品的多层控制方法,所述多层控制方法包括:由第一层控制器根据传感器检测得到电力电子产品的第一状态信号;由第二层控制器根据所述第一状态信号和模拟所述电力电子产品的模型,获取所述电力电子产品除所述第一状态信号以外的第二状态信号;由所述第二层控制器根据所述第一状态信号、所述第二状态信号和所述电力电子产品的控制目标,生成控制参考信号;以及由所述第一层控制器根据所述控制参考信号生成所述电力电子产品的控制信号,并输出至所述电力电子产品。
在上述多层控制方法的一实施例中,可选的,所述多层控制方法还包括:由第三层控制器根据所述第一层控制器反馈的历史第一状态信号和历史控制信号、所述第二层控制器反馈的历史第二状态信号和历史控制参考信号辨识并优化所述模型的模型参数;以及使所述第二层控制器基于所述模型参数更新所述模型。
在上述多层控制方法的一实施例中,可选的,所述多层控制方法还包括:响应于所述模型参数的变化超出预设阈值,由所述第三层控制器将变化超出预设阈值变化的模型参数发送至所述第一层控制器,以使所述第一层控制器基于所述模型参数诊断所述电力电子产品的硬件状态。
在上述多层控制方法的一实施例中,可选的,所述第二层控制器中模拟所述电力电子产品的模型包括:机理模型,和/或,大数据模型;其中所述机理模型为模拟所述电力电子产品的数学物理模型,以基于所述第一状态信号获取所述电力电子产品除所述第一状态信号以外的模拟状态变量为所述第二状态信号;所述大数据模型表征所述电力电子产品的第一状态信号和/或所述模拟状态变量与所述电力电子产品的预期状态变量之间的映射关系,以根据所述第一状态信号和/或所述模拟状态变量输出所述预期状态变量为所述第二状态信号。
在上述多层控制方法的一实施例中,可选的,响应于所述模型包括所述机理模型和所述大数据模型,所述第二层控制器获取所述第二状态信号进一步包括:根据所述第一状态信号和所述机理模型获取所述模拟状态变量;将所述第一状态 信号和所述模拟状态变量输入所述大数据模型,以获取所述预期状态变量为所述第二状态信号;其中所述第二层控制器生成所述控制参考信号所根据的控制参考信号为所述大数据模型输出的第二状态信号。
本发明的另一方面还提供了一种计算机可读存储介质,存储有计算机程序,所述计算机程序被处理器执行时实现如上述任意一项实施例所描述的电力电子产品的多层控制方法的步骤。
根据本发明所提供的电力电子产品的多层控制系统及方法,能够在不额外增加硬件成本的情况下,基于已有的外部控制器,使电力电子及其控制器具有更加丰富的功能和优越的性能。本发明能够应用在车载领域,能够在不增加整车的硬件成本的前提下,通过整车电子电器架构和智能网联技术,实现车载电力电子产品及其控制器的功能、性能优化,提高整车性能,为整车提供更多收益。
附图说明
在结合以下附图阅读本公开的实施例的详细描述之后,能够更好地理解本发明的上述特征和优点。在附图中,各组件不一定是按比例绘制,并且具有类似的相关特性或特征的组件可能具有相同或相近的附图标记。
图1示出了本发明的一方面所提供的电力电子产品的多层控制系统的结构示意图。
图2示出了本发明的一方面所提供的电力电子产品的多层控制系统的一优选实施例的结构示意图。
图3示出了本发明的一方面所提供的电力电子产品的多层控制方法的流程示意图。
具体实施方式
以下结合附图和具体实施例对本发明作详细描述。注意,以下结合附图和具体实施例描述的诸方面仅是示例性的,而不应被理解为对本发明的保护范围进行任何限制。
给出以下描述以使得本领域技术人员能够实施和使用本发明并将其结合到具体应用背景中。各种变型、以及在不同应用中的各种使用对于本领域技术人员将是容易显见的,并且本文定义的一般性原理可适用于较宽范围的实施例。由此,本发明并不限于本文中给出的实施例,而是应被授予与本文中公开的原理和新颖性特征相一致的最广义的范围。
在以下详细描述中,阐述了许多特定细节以提供对本发明的更透彻理解。然而,对于本领域技术人员显而易见的是,本发明的实践可不必局限于这些具体细节。换言之,公知的结构和器件以框图形式示出而没有详细显示,以避免模糊本发明。
请读者注意与本说明书同时提交的且对公众查阅本说明书开放的所有文件及文献,且所有这样的文件及文献的内容以参考方式并入本文。除非另有直接说明,否则本说明书(包含任何所附权利要求、摘要和附图)中所揭示的所有特征皆可由用于达到相同、等效或类似目的的可替代特征来替换。因此,除非另有明确说明,否则所公开的每一个特征仅是一组等效或类似特征的一个示例。
注意,在使用到的情况下,标志左、右、前、后、顶、底、正、反、顺时针和逆时针仅仅是出于方便的目的所使用的,而并不暗示任何具体的固定方向。事实上,它们被用于反映对象的各个部分之间的相对位置和/或方向。此外,术语“第一”、“第二”仅用于描述目的,而不能理解为指示或暗示相对重要性。
注意,在使用到的情况下,进一步地、较优地、更进一步地和更优地是在前述实施例基础上进行另一实施例阐述的简单起头,该进一步地、较优地、更进一步地或更优地后带的内容与前述实施例的结合作为另一实施例的完整构成。在同一实施例后带的若干个进一步地、较优地、更进一步地或更优地设置之间可任意组合的组成又一实施例。
如上文所描述的,为了解决现有技术中,电力电子产品的控制器在不增加额外成本的情况下,无法提供更优的控制参数的问题,本发明提供了一种电力电子产品的多层控制系统及方法。
首先,请结合图1和图2来理解本发明的一方面所提供的电力电子产品的多层控制系统的结构。
如图1所示出的,本发明的一方面所提供的电力电子产品的多层控制系统至少包括第一层控制器100和第二层控制器200,第一层控制器100与第二层控制器200彼此之间具有通讯连接。第一层控制器100与控制对象(即电力电子产品,包括但不限于电轴、充电机、DCDC模块等)之间具有电气连接和/或通讯连接。第一层控制器100能够获取控制对象根据传感器检测得到的第一状态信号,并将第一状态信号输出第二层控制器200。第一状态信号是基于实际的物理传感器采样得到的电气信号,包括但不限于温度、电压、电流等。第一层控制器100具有常见的控制、计算能力,主要用于运行实时性较高、低算力需求、低数据存储、简单控制算法的软件进程,可以输出控制对象的控制信号,一般是PWM占空比信号。
第二层控制器200主要用于运行实时性中等、高算力需求、低数据存储、智能优化算法的软件进程,第二层控制器200根据第一状态信号和模拟电力电子产品的模型,能够获取得到电力电子产品除所述第一状态信号以外的第二状态信号。第二状态信号是指难以通过观测或通过物理传感器无法测量得到的系统状态变量,包括但不限于控制对象内部的温度等难以表征的变量。在第二层控制器200中,可以通过模拟电力电子产品的模型,来获取这些难以通过观测或传感器测量得到的状态量。第二层控制器200能够基于第二状态信号,以某些目标最优为前 提,根据控制目标,计算出第一层控制器100的控制参考值,并输出到第一层控制器100。
在一优选的实施例中,多层控制系统还包括第三层控制器300,第三层控制器300与第一层控制器100、第二层控制器200之间均具有通讯连接,可以接收第一层控制器100输出的历史第一状态信号和历史控制信号,可以接收第二层控制器200输出的历史第二状态信号和历史控制参考信号,并基于这些历史信号,辨识并优化第二层控制器200中的模型的模型参数,以使第二层控制器更新模型,从而能够更好地基于模型模拟得到被控产品的状态量。第三层控制器300用于运行实时性较慢、较高算力需求、海量数据存储、智能优化算法的软件进程,用于优化或调整第二层控制器中涉及到的模型参数。通过该分布式的多层控制架构,能够丰富电力电子产品的功能及提升电力电子产品的性能。
当应用到车辆上时,第一层控制器100可以认为是电力电子车载控制器,第二层控制器200可以认为是车载计算单元,而第三层控制器300可以认为是与车辆对应的云端控制器。
请进一步结合图2来理解本发明的一方面所提供的一种优选的多层控制系统。如图2所示出的,在该实施例中,包括第一层控制器100、第二层控制器200和第三层控制器300。
第一层控制器100用于实现电力电子控制器最基本的控制功能,第二层控制器200用于实现某些目标最优的优化控制,第三层控制器300用于优化或调整多层控制系统第二层中涉及到的模型参数。需要注意的是,第三层控制器300是用于处理海量数据的,类似于云端控制器,但与当前的云端略有不同,本发明中的第三层控制器300仅面向单一的电力电子产品,即每个电力电子产品都有其独立的多层控制系统第一层、第二层和第三层。
在该优选的实施例中,第一层控制器100进一步包括内部的各个子模块,例如,数据采样110、控制算法120、驱动生成130和部件诊断140。第二层控制器200进一步包括:智能算法210、机理模型220和数据模型230。第三层控制器进一步包括:数据存储310、辨识算法320、参数发送330、辨识模型340和数据诊断350。虽然将其定义为了“模块”,但实际上在控制器的内部可以并不存在物理意义上的模块,而是通过不同的软件来实现各个“模块”的功能。
其中,对于控制对象(即电力电子产品)而言,其需要将传感器采样得到的数据输出给数据采样110,传输的数据包括多层控制系统第一层基于实际的物理传感器采样相关的电气信号,包括温度、电压、电流等。而控制对象接收到的信号则为多层控制系统第一层输出至控制器对象的控制信号,对于电力电子产品而言,一般为PWM占空比信号。
100->300数据发送:多层控制系统第一层将软件采样信号、软件控制信号以及控制相关中间量发送至多层控制系统第三层.
100->200数据发送:多层控制系统第一层将软件采样信号、软件控制信号以及控制相关中间量发送至多层控制系统第二层。
200->300数据发送:多层控制系统第二层基于模型及100->200数据发送的数据,可以观测或估计出实际传感器测量不到的系统状态变量,比如控制对象内部的温度、难以表征的变量等,并将观测的系统状态变量发送至多层控制系统第三层。
300->200数据发送:多层控制系统第三层通过数据存储310记录全生命周期的100->300数据发送和200->300数据发送,并基于长时间尺度数据通过辨识算法320和辨识模型340在线辨识多层控制系统第二层中所用到模型参数(由于电力电子产品作为物理器件,其特性随时间会发生变化,故多层控制系统第二层中的模型参数亦需要进行变化),并将辨识的模型参数通过参数发送330发送至多层控制系统第二层中涉及的模型中,并替代当前模型参数。
200->100数据发送:多层控制系统第二层基于100->200数据发送,基于相应的算法调度及模型估计及预测,以某些目标最优为前提,通过智能算法210计算出多层控制系统第一层的控制参考值,并传递给多层控制系统第一层的控制算法120。控制算法120能够基于常见的控制、计算能力计算出控制对象的相关控制信号,并输出至控制对象。在一实施例中,当应用于车辆时,需要将这些控制信号再经过驱动生成130转化为相关的驱动信号。
300->100数据发送:多层控制系统第三层记录全生命周期的100->300数据发送和200->300数据发送,基于长时间尺度数据通过数据诊断350和辨识算法320判断控制对象是否正常,如某些参数或状态变量在相同的工况下,变化趋势巨大,则将相关数据发送至多层控制系统第一层,由部件诊断140进行相关诊断和响应。
尤其对于第二层控制器200来说,其中的智能算法210用于实现某些目标优化的智能调度算法,诸如PSO粒子群算法、模拟退火算法及模型预测算法等。
机理模型220为可复现控制对象的数学物理模型,用于估计控制对象中,无法测量的相关系统状态变量。
数据模型230:经过离线训练好的数据模型(诸如神经网络模型等大数据模型)。离线经过大量数据训练,在线应用时,将控制对象的各种状态变量和采样值输入后,则可以输出希望得到的结果。
210->220数据发送及调用:智能算法210将第一层发送给第二层的数据中的信息传递给机理模型220,机理模型220观测或估计控制对象内部的状态变量(难以实际测量的内容)。另外,智能算法210向机理模型220发送相关调用指令,以调用机理模型220得到第二状态信号。
220->230数据发送:机理模型220基于输入的数据,观测或估计难以用传感器测量到的系统状态变量,并发送至数据模型230中。
210->230数据发送及调用:智能算法210将第一层发送给第二层的信息传递给数据模型230,另外,智能算法210向数据模型230发送相关调用指令,以调用数据模型230得到第二状态信号。在一实施例中,响应于同时包括机理模型220和数据模型230,可以先通过调用机理模型220,基于第一状态信号获取模拟状态变量,随后调用数据模型230,基于第一状态信号和模拟状态变量获取预期状态变量为第二状态信号。
230->210数据发送:数据模型230基于210->220数据发送和220->230数据发送,基于提前训练好的数据模型(如神经网络模型),输出机理模型难以计算或者计算不准确的相关数据信息,并将其发送至智能算法210中。
220->210数据发送:机理模型220基于210->220数据发送,基于物理数学模型观测或估计控制对象内部的状态变量(难以实际测量的内容),并将其发送至数据模型智能算法210中。
智能算法210则根据数据采样140发送的第一状态信号,以及机理模型220和/或数据模型230获取得到的第二状态信号基于某些目标优化的智能调度算法来计算出控制算法120的控制参考信号。
请结合图3所示出的本发明的另一方面所提供的电力电子产品的多层控制方法的流程示意图来进一步理解本发明所提供的多层控制系统的具体实现过程。
如图3所示出的,响应于开始,在第二层控制器200中获取得到控制目标,若应用在车辆上,该控制目标为车辆的VCU发送的。通过智能算法210能够将控制目标进行优化,从而生成控制参考信号给第一层控制器100中的控制算法,由控制算法120基于数据采样得到的第一状态信号计算出控制信号后基于驱动生成输出至控制对象。
而在智能算法210对控制目标进行优化的过程中,需要使用到基于模型(包括机理模型和/或大数据模型)的辨识算法,从而能够基于模型获取得到无法通过观察或者传感器采样得到的第二状态信号,以基于第一状态信号和第二状态信号通过智能算法来对控制目标进行优化。
由于电力电子产品作为物理器件,其性能参数在经过一段时间的使用之后会有变化,需要将物理器件的参数变化体现在模拟电力电子产品的模型上。因此需要第三层控制器300通过辨识模型和辨识算法,基于数据存储中存储的第一层控制器和第二层控制器发送的各类历史数据对模型参数进行辨识与优化。当需要对模型参数进行更新时,将相关的更新后的模型参数输出至第二层控制器中的辨识算法,以完成模型参数的替换更新。
另一方面,由于电力电子产品是实际存在的物理器件,若能够在第三层控制器中通过数据诊断发现模拟得到该物理器件的硬件性能发生了较大的变化,会影响到物理器件的正常使用,则需要及时将变化较大的参数反馈给第一层控制器,由第一层控制器的部件诊断基于第三层控制器辨识得到的模型参数诊断电力电子 产品的硬件状态。
至此,已经描述了本发明所提供的电力电子产品的多层控制系统和方法的具体实现方法。本发明利用分布式多层控制架构中第二层的高算力特性,为分布式多层控制架构中第一层提供相应的控制目标、限制条件、控制修正等相关信息,以提升整体控制软件性能。在优选的具有第三层控制器的实施例中,利用分布式多层控制架构中第三层的海量数据处理能力与智能化算法迭代优化能力特点,为分布式多层控制架构中第一层提供相应的长时间尺度监控及控制相关信息,为分布式多层控制架构中第二层提供相应的模型参数优化、算法参数优化等相关信息,以提升整体控制软件性能。
本发明尤其能够应用在车辆领域,多层架构可以对应于车载电力电子产品的控制器(第一层)、车载计算器(第二层)。第三层控制器也可以与云端对应。本发明能够利用电力电子产品的控制器之外的资源(车上其他控制器、车载电脑、云端等资源),来优化控制器(控制器使用芯片保持不变)的性能(控制精度、系统功率、系统效率及鲁棒性等),间接实现控制器高控制性能的降本方案。
由于当前电力电子产品的控制器受成本制约,控制器MCU难以进行智能算法及复杂模型的应用,故难以进行性能优化控制。基于本发明提出的多层控制架构,可以利用控制器之外的资源,通过外部通信进行数据传递的方式,来实现该控制器的性能优化控制,同时该控制器硬件不变,即成本不变。
具体的,当前电力电子产品需要提升性能(比如效率),以满足更高的产业目标,而目前的现状难以满足,软件提升系统性能的手段即是通过寻优控制,让系统运行于某些性能(诸如效率最高等)较优的区域,为了实时地让系统运行于性能较优的区域,需要使用一些智能算法、机理模型、数据模型来做在线寻优控制,其中机理模型和数据模型中的参数是离线获取,但随着使用时间的推移,控制系统中可能出现一些老化现象,故需要基于该控制器长时间尺度的信息,在线辨识模型参数,并替换机理或数据模型中的参数。故综上所述,将传统PI闭环控制定义为第一层,将寻优控制定义为第二层,将寻优控制中使用模型的参数辨识优化定义为第三层。由于当前控制器使用MCU资源的问题,只能完成第一层,难以完成第二层和第三层,由于第二层要不停的给第一层设定控制参考值,所以对通讯时间有一定要求,故第二层需要放在车内(可以支持SOA服务的控制器),第三层对通讯时间没有要求,但需要大量的数据存储,故可以在车内,也可以在云端。
根据本发明,能够丰富电力电子产品软件功能,提升电力电子产品软件性能,能够降低电力电子产品软件控制芯片性能要求,实现产品降本。
本发明的另一方面还提供了一种计算机可读存储介质,存储有计算机程序,计算机程序被处理器执行时实现如上文任意一项实施例所描述的电力电子产品的多层控制方法的步骤,具体请参考上文的描述,在此不再赘述。另外,可以理解 的是,上述的计算机可读存储介质亦可以是系统形式,即包括有多个计算机可读存储子介质(对应于多层),以通过多个计算机可读存储介质共同实现上文所描述的电力电子产品的多层控制方法的步骤。
结合本文所公开的实施例描述的各种解说性逻辑模块、和电路可用通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或其它可编程逻辑器件、分立的门或晶体管逻辑、分立的硬件组件、或其设计成执行本文所描述功能的任何组合来实现或执行。通用处理器可以是微处理器,但在替换方案中,该处理器可以是任何常规的处理器、控制器、微控制器、或状态机。处理器还可以被实现为计算设备的组合,例如DSP与微处理器的组合、多个微处理器、与DSP核心协作的一个或多个微处理器、或任何其他此类配置。
结合本文中公开的实施例描述的方法或算法的步骤可直接在硬件中、在由处理器执行的软件模块中、或在这两者的组合中体现。软件模块可驻留在RAM存储器、闪存、ROM存储器、EPROM存储器、EEPROM存储器、寄存器、硬盘、可移动盘、CD-ROM、或本领域中所知的任何其他形式的存储介质中。示例性存储介质耦合到处理器以使得该处理器能从/向该存储介质读取和写入信息。在替换方案中,存储介质可以被整合到处理器。处理器和存储介质可驻留在ASIC中。ASIC可驻留在用户终端中。在替换方案中,处理器和存储介质可作为分立组件驻留在用户终端中。
在一个或多个示例性实施例中,所描述的功能可在硬件、软件、固件或其任何组合中实现。如果在软件中实现为计算机程序产品,则各功能可以作为一条或更多条指令或代码存储在计算机可读介质上或藉其进行传送。计算机可读介质包括计算机存储介质和通信介质两者,其包括促成计算机程序从一地向另一地转移的任何介质。存储介质可以是能被计算机访问的任何可用介质。作为示例而非限定,这样的计算机可读介质可包括RAM、ROM、EEPROM、CD-ROM或其它光盘存储、磁盘存储或其它磁存储设备、或能被用来携带或存储指令或数据结构形式的合意程序代码且能被计算机访问的任何其它介质。任何连接也被正当地称为计算机可读介质。例如,如果软件是使用同轴电缆、光纤电缆、双绞线、数字订户线(DSL)、或诸如红外、无线电、以及微波之类的无线技术从web网站、服务器、或其它远程源传送而来,则该同轴电缆、光纤电缆、双绞线、DSL、或诸如红外、无线电、以及微波之类的无线技术就被包括在介质的定义之中。如本文中所使用的盘(disk)和碟(disc)包括压缩碟(CD)、激光碟、光碟、数字多用碟(DVD)、软盘和蓝光碟,其中盘(disk)往往以磁的方式再现数据,而碟(disc)用激光以光学方式再现数据。上述的组合也应被包括在计算机可读介质的范围内。
提供之前的描述是为了使本领域中的任何技术人员均能够实践本文中所描述的各种方面。但是应该理解,本发明的保护范围应当以所附权利要求书为准,而 不应被限定于以上所解说实施例的具体结构和组件。本领域技术人员在本发明的精神和范围内,可以对各实施例进行各种变动和修改,这些变动和修改也落在本发明的保护范围之内。

Claims (13)

  1. 一种电力电子产品的多层控制系统,其特征在于,所述多层控制系统至少包括:第一层控制器和第二层控制器;其中
    所述第一层控制器被配置为:获取所述电力电子产品根据传感器检测得到的第一状态信号,发送所述第一状态信号至所述第二层控制器;以及
    接收所述第二层控制器反馈的控制参考信号,基于所述控制参考信号生成所述电力电子产品的控制信号,并输出至所述电力电子产品;
    所述第二层控制器被配置为:根据所述第一状态信号和模拟所述电力电子产品的模型,获取所述电力电子产品除所述第一状态信号以外的第二状态信号;以及
    根据所述第一状态信号、所述第二状态信号和所述电力电子产品的控制目标,生成所述控制参考信号,并输出至所述第一层控制器。
  2. 如权利要求1所述的多层控制系统,其特征在于,还包括第三层控制器;其中
    所述第三层控制器被配置为:根据所述第一层控制器反馈的历史第一状态信号和历史控制信号、所述第二层控制器反馈的历史第二状态信号和历史控制参考信号辨识并优化所述模型的模型参数,并输出所述模型参数至所述第二层控制器;
    所述第二层控制器还被配置为:基于所接收到的模型参数更新所述模型。
  3. 如权利要求2所述的多层控制系统,其特征在于,所述第三层控制器还被配置为:
    响应于所述模型参数的变化超出预设阈值,将变化超出预设阈值变化的模型参数发送至所述第一层控制器,以使所述第一层控制器基于所述模型参数诊断所述电力电子产品的硬件状态。
  4. 如权利要求1或2所述的多层控制系统,其特征在于,所述第二层控制器中模拟所述电力电子产品的模型包括:机理模型,和/或,大数据模型;其中
    所述机理模型为模拟所述电力电子产品的数学物理模型,以基于所述第一状态信号获取所述电力电子产品除所述第一状态信号以外的模拟状态变量为所述第二状态信号;
    所述大数据模型表征所述电力电子产品的第一状态信号和/或所述模拟状态变量与所述电力电子产品的预期状态变量之间的映射关系,以根据所述第一状态信号和/或所述模拟状态变量输出所述预期状态变量为所述第二状态信号。
  5. 如权利要求4所述的多层控制系统,其特征在于,所述第三层控制器进一步被配置为:
    辨识并优化所述数学物理模型和/或所述大数据模型的模型参数,并输出所述模型参数至所述第二层控制器。
  6. 如权利要求4所述的多层控制系统,其特征在于,响应于所述模型包括所述机理模型和所述大数据模型,所述第二层控制器进一步被配置为:
    根据所述第一状态信号和所述机理模型获取所述模拟状态变量;
    将所述第一状态信号和所述模拟状态变量输入所述大数据模型,以获取所述预期状态变量为所述第二状态信号;以及
    根据所述大数据模型输出的第二状态信号生成所述控制参考信号。
  7. 如权利要求4所述的多层控制系统,其特征在于,所述多层控制系统应用于车辆,所述第一层控制器为所述车辆的电力电子产品控制器,所述第二层控制器为所述车辆的车载计算单元,所述第三层控制器为与所述车辆一一对应的云端控制器。
  8. 一种电力电子产品的多层控制方法,其特征在于,所述多层控制方法包括:
    由第一层控制器根据传感器检测得到电力电子产品的第一状态信号;
    由第二层控制器根据所述第一状态信号和模拟所述电力电子产品的模型,获取所述电力电子产品除所述第一状态信号以外的第二状态信号;
    由所述第二层控制器根据所述第一状态信号、所述第二状态信号和所述电力电子产品的控制目标,生成控制参考信号;以及
    由所述第一层控制器根据所述控制参考信号生成所述电力电子产品的控制信号,并输出至所述电力电子产品。
  9. 如权利要求8所述的多层控制方法,其特征在于,所述多层控制方法还包括:
    由第三层控制器根据所述第一层控制器反馈的历史第一状态信号和历史控制信号、所述第二层控制器反馈的历史第二状态信号和历史控制参考信号辨识并优化所述模型的模型参数;以及
    使所述第二层控制器基于所述模型参数更新所述模型。
  10. 如权利要求9所述的多层控制方法,其特征在于,所述多层控制方法还包括:
    响应于所述模型参数的变化超出预设阈值,由所述第三层控制器将变化超出 预设阈值变化的模型参数发送至所述第一层控制器,以使所述第一层控制器基于所述模型参数诊断所述电力电子产品的硬件状态。
  11. 如权利要求8或9所述的多层控制方法,其特征在于,所述第二层控制器中模拟所述电力电子产品的模型包括:机理模型,和/或,大数据模型;其中
    所述机理模型为模拟所述电力电子产品的数学物理模型,以基于所述第一状态信号获取所述电力电子产品除所述第一状态信号以外的模拟状态变量为所述第二状态信号;
    所述大数据模型表征所述电力电子产品的第一状态信号和/或所述模拟状态变量与所述电力电子产品的预期状态变量之间的映射关系,以根据所述第一状态信号和/或所述模拟状态变量输出所述预期状态变量为所述第二状态信号。
  12. 如权利要求11所述的多层控制方法,其特征在于,响应于所述模型包括所述机理模型和所述大数据模型,所述第二层控制器获取所述第二状态信号进一步包括:
    根据所述第一状态信号和所述机理模型获取所述模拟状态变量;
    将所述第一状态信号和所述模拟状态变量输入所述大数据模型,以获取所述预期状态变量为所述第二状态信号;其中
    所述第二层控制器生成所述控制参考信号所根据的控制参考信号为所述大数据模型输出的第二状态信号。
  13. 一种计算机可读存储介质,存储有计算机程序,其特征在于,所述计算机程序被处理器执行时实现如权利要求8-12中任意一项所述的电力电子产品的多层控制方法的步骤。
PCT/CN2023/103039 2022-07-20 2023-06-28 一种电力电子产品的多层控制系统及方法 WO2024016966A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210862341.4 2022-07-20
CN202210862341.4A CN115291545A (zh) 2022-07-20 2022-07-20 一种电力电子产品的多层控制系统及方法

Publications (1)

Publication Number Publication Date
WO2024016966A1 true WO2024016966A1 (zh) 2024-01-25

Family

ID=83823772

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/103039 WO2024016966A1 (zh) 2022-07-20 2023-06-28 一种电力电子产品的多层控制系统及方法

Country Status (2)

Country Link
CN (1) CN115291545A (zh)
WO (1) WO2024016966A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115291545A (zh) * 2022-07-20 2022-11-04 联合汽车电子有限公司 一种电力电子产品的多层控制系统及方法
CN115426397B (zh) * 2022-11-07 2023-03-24 蔚来汽车科技(安徽)有限公司 用于微控制器mcu的面向服务系统、电子电器系统

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104471852A (zh) * 2012-07-17 2015-03-25 Abb研究有限公司 多端hvdc控制
CN105939028A (zh) * 2015-03-03 2016-09-14 通用电气公司 用于增强发电厂发电单元的控制的方法和系统
CN107069973A (zh) * 2017-06-09 2017-08-18 浙江群力电气有限公司 一种电力网络监控系统
CN110912230A (zh) * 2019-12-04 2020-03-24 中南大学 基于寻优预测的多模块充电变流器控制装置及方法
CN112947061A (zh) * 2019-11-26 2021-06-11 东京毅力科创株式会社 控制系统、控制方法、控制程序和处理系统
CN113511150A (zh) * 2020-04-10 2021-10-19 采埃孚汽车科技(上海)有限公司 一种集成的电子控制单元及车辆的控制方法
WO2022147299A1 (en) * 2020-12-31 2022-07-07 Concentric Power, Inc. Adaptive power for ai/ml systems and methods in a microgrid
CN115291545A (zh) * 2022-07-20 2022-11-04 联合汽车电子有限公司 一种电力电子产品的多层控制系统及方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104471852A (zh) * 2012-07-17 2015-03-25 Abb研究有限公司 多端hvdc控制
CN105939028A (zh) * 2015-03-03 2016-09-14 通用电气公司 用于增强发电厂发电单元的控制的方法和系统
CN107069973A (zh) * 2017-06-09 2017-08-18 浙江群力电气有限公司 一种电力网络监控系统
CN112947061A (zh) * 2019-11-26 2021-06-11 东京毅力科创株式会社 控制系统、控制方法、控制程序和处理系统
CN110912230A (zh) * 2019-12-04 2020-03-24 中南大学 基于寻优预测的多模块充电变流器控制装置及方法
CN113511150A (zh) * 2020-04-10 2021-10-19 采埃孚汽车科技(上海)有限公司 一种集成的电子控制单元及车辆的控制方法
WO2022147299A1 (en) * 2020-12-31 2022-07-07 Concentric Power, Inc. Adaptive power for ai/ml systems and methods in a microgrid
CN115291545A (zh) * 2022-07-20 2022-11-04 联合汽车电子有限公司 一种电力电子产品的多层控制系统及方法

Also Published As

Publication number Publication date
CN115291545A (zh) 2022-11-04

Similar Documents

Publication Publication Date Title
WO2024016966A1 (zh) 一种电力电子产品的多层控制系统及方法
WO2021082478A1 (zh) 空调系统能耗预测方法及装置
CN111475384B (zh) 壳体温度的计算方法、装置、存储介质及电子设备
CN104899135A (zh) 软件缺陷预测方法和系统
WO2021131958A1 (ja) 模擬電池構築方法および模擬電池構築装置
TWI692702B (zh) 具品質量測與機構診斷功能之生產系統及其驅動器與方法
CN116193819B (zh) 一种数据中心机房节能控制方法、系统、装置及电子设备
CN116089870A (zh) 基于元学习的小样本条件下工业设备故障预测方法及装置
CN115186555B (zh) 基于数字孪生的干燥设备实况仿真方法及相关设备
AU2021329231B2 (en) Systems and methods to assess and repair data using data quality indicators
US20220156171A1 (en) Temperature prediction system and method for predicting a temperature of a chip of a pcie card of a server
CN109960147A (zh) 冷源控制参数的确定方法及装置、设备及存储介质
CN113487086A (zh) 设备剩余使用寿命预测方法、装置、计算机设备和介质
WO2024060539A1 (zh) 动力电池的年度碳排放量估算方法及装置
JP7214283B1 (ja) ニューラルネットワーク基盤の負荷推定モデルを含む積算電力装置で負荷を推定する方法及びこれを利用した積算電力装置
Zhang et al. State of charge estimation for lithium-ion batteries using simple recurrent units and unscented kalman filter
CN115718478A (zh) 基于sac深度强化学习的svg参数优化辨识方法
CN114757111A (zh) 一种基于数字孪生技术的智能装备健康管理方法
CN110955286B (zh) 一种禽蛋监测方法及装置
Wang et al. IUP: an intelligent utility prediction scheme for solid-state fermentation in 5G IoT
CN117458011B (zh) 一种储能系统性能评估模型以及性能评估方法
CN117236448B (zh) 一种基于时序知识图谱的雷达意图推理及其模型训练方法
CN117171551B (zh) 大规模工业设备数据分析与智能管理方法
TWI756933B (zh) 伺服器的PCIe晶片的溫度預測系統及方法
CN117545257A (zh) 一种机房空调节能控制方法、系统、设备及介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23842032

Country of ref document: EP

Kind code of ref document: A1