An electromechanical parameter collaborative identification method and system for cross-medium propulsion

Abstract

The application discloses a kind of electrical mechanical parameter collaborative identification method and system for cross-medium propulsion, the method includes: the parameter of cross-medium propulsion permanent magnet synchronous motor is initialized and runs basic magnetic field orientation control and position sensor observation, obtains mechanical angular velocity estimation and load torque observation value;The running medium state of cross-medium propulsion permanent magnet synchronous motor is identified, and electrical identification enabling signal and mechanical update gate signal are generated;Adaptive double signal injection is used, and the electrical parameter of cross-medium propulsion permanent magnet synchronous motor is identified on line, and updated electrical parameter is obtained;Cross-medium propulsion permanent magnet synchronous motor is identified and speed loop self-tuning, and electrical mechanical parameter collaborative identification result is obtained.The application can improve the electrical mechanical parameter collaborative identification precision under cross-medium environment.The application is a kind of electrical mechanical parameter collaborative identification method and system for cross-medium propulsion, and can be widely applied in parameter identification technical field.

Description

Technical Field

[0001] This invention relates to the field of parameter identification technology, and in particular to a method and system for collaborative identification of electromechanical parameters for cross-medium propulsion. Background Technology

[0002] With the development of cross-medium unmanned aerial vehicles (UAVs), amphibious vehicles, and underwater propulsion platforms, permanent magnet synchronous motors (PMSMs) are widely used in propulsion systems due to their high efficiency, high power density, and fast dynamic response. Existing PMSM propulsion motors typically employ field-oriented control (FOC), speed loop PI control, and sensorless control based on back electromotive force. While these methods achieve good control performance in single air or single underwater environments, they still have the following inherent drawbacks in water-air medium switching, semi-submersible propulsion, and transient water entry / exit: Existing PMSM control methods typically use fixed motor parameters for FOC modeling and speed loop design. However, during cross-medium operation, electrical and mechanical parameters change with the medium state and load conditions, causing a mismatch between the speed loop PI parameters and the actual controlled object, thereby reducing current decoupling accuracy, torque response performance, and speed control stability.

[0003] Existing methods generally do not establish a correlation between media state identification and scheduling, and cannot determine whether electrical and mechanical parameters are suitable for updating based on different media states. Therefore, problems such as improper identification timing and parameter update distortion are prone to occur during media switching. Existing mechanical parameter identification methods also often continuously update rotational inertia and damping throughout the entire operation, making it difficult to avoid the contamination of slowly changing mechanical parameter estimates by transient shocks.

[0004] Meanwhile, existing electrical parameter identification methods mostly employ fixed-amplitude signal injection, which can easily introduce additional current ripple, torque fluctuations, and thrust disturbances during underwater heavy loads or water-to-air switching, affecting the stability of sensorless observation and current control. Existing mechanical parameter identification methods also frequently update rotational inertia and damping throughout the entire operation, making it difficult to avoid the contamination of slowly changing mechanical parameter estimates by transient shocks. Summary of the Invention

[0005] To address the aforementioned technical problems, the present invention aims to provide a method and system for collaborative identification of electromechanical parameters in cross-medium propulsion, which can improve the accuracy of collaborative identification of electromechanical parameters in cross-medium environments.

[0006] The first technical solution adopted in this invention is: a method for collaborative identification of electromechanical parameters for cross-medium propulsion, comprising the following steps: Initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system and run basic field orientation control and sensorless observation to obtain the estimated mechanical angular velocity and the observed load torque. Based on the load torque observation and mechanical angular velocity estimation, the operating medium state of the cross-medium propulsion permanent magnet synchronous motor is identified, and an electrical identification enable signal and a mechanical update gating signal are generated. Based on the electrical identification enable signal, an adaptive dual-signal injection method is used to perform online identification of electrical parameters for a cross-medium propulsion permanent magnet synchronous motor, and the updated electrical parameters are obtained. Based on the mechanical update gating signal and the updated electrical parameters, mechanical parameter identification and speed loop self-tuning are performed on the cross-medium propulsion permanent magnet synchronous motor, and the load torque feedforward compensation is output synchronously to obtain the electrical and mechanical parameter collaborative identification results.

[0007] Furthermore, the step of initializing the parameters of the cross-medium propulsion permanent magnet synchronous motor system and running basic field-oriented control and sensorless observation to obtain the estimated mechanical angular velocity and load torque observation specifically includes: Initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system, including the initial values ​​of electrical parameter calibration, mechanical parameter calibration, and speed loop control parameters; By default, the cross-medium propulsion permanent magnet synchronous motor system is in the medium transition state, and the updates of electrical parameters, mechanical slow variables, and speed loop slow variables are disabled, and the initial gating state is set. Based on the parameters of the initial cross-medium propulsion permanent magnet synchronous motor system, the cross-medium propulsion permanent magnet synchronous motor is controlled to perform basic field orientation control. The three-phase current and bus voltage are collected, and the d-axis and q-axis currents and corresponding voltages are obtained through Clarke / Park transformation. Based on the sensorless control method, a back EMF observation model is constructed to estimate the d-axis and q-axis currents and corresponding voltages, thereby obtaining the initial electrical parameters, initial mechanical parameters and mechanical angular velocity estimates. Based on the initial electrical parameters, initial mechanical parameters, and estimated mechanical angular velocity, LIS-MAI load torque pre-observation is performed to obtain the load torque observation value.

[0008] Furthermore, the step of performing LIS-MAI load torque pre-observation based on initial electrical parameters, initial mechanical parameters, and estimated mechanical angular velocity to obtain the observed load torque specifically includes: Based on the initial electrical and mechanical parameters, the current torque constant estimate of the cross-medium propulsion permanent magnet synchronous motor is determined, and the electromagnetic torque estimate is calculated in conjunction with the q-axis current. Define the internal velocity observations and mechanical angular velocity estimates of LIS-MAI, and obtain the mechanical side observation errors; Define the load torque change rate and determine the adaptive robust gain; By employing a hyperbolic tangent smoothing function, combined with adaptive robust gain, mechanical side observation error, and electromagnetic torque estimation, the robust compensation term of LIS-MAI is determined, and the LIS-MAI velocity observation equation is constructed. Based on the LIS-MAI velocity observation equation, the load torque observation value of the cross-medium propulsion permanent magnet synchronous motor is updated in real time to obtain the load torque observation value.

[0009] Furthermore, the step of identifying the operating medium state of the cross-medium propulsion permanent magnet synchronous motor based on the load torque observation value and the mechanical angular velocity estimate, and generating the electrical identification enable signal and the mechanical update gating signal, specifically includes: Based on the observed load torque and the estimated mechanical angular velocity, construct the characteristic quantities for medium discrimination; Set dual judgment thresholds and continuous stability judgment period, and define the first stable medium judgment function and the second stable medium judgment function. The operating medium state of the cross-medium propulsion permanent magnet synchronous motor is determined based on the first stable medium determination function and the second stable medium determination function. Based on the operating medium state of the cross-medium propulsion permanent magnet synchronous motor, the electrical identification enable signal and the mechanical update gating signal are defined.

[0010] Furthermore, the step of performing online electrical parameter identification of the cross-medium propulsion permanent magnet synchronous motor based on the electrical identification enable signal using adaptive dual-signal injection to obtain updated electrical parameters specifically includes: The adaptive injection intensity coefficient is calculated by combining the operating medium state, load torque observation, and electrical identification enable signal of the cross-medium propulsion permanent magnet synchronous motor. Based on the injection intensity coefficient, the SWPO position offset injection amplitude and the SWC current injection amplitude are generated; Inductance identification is performed, and the SWPO position offset injection amplitude is injected into the estimated rotor electrical angle of the cross-medium propulsion permanent magnet synchronous motor. The q-axis voltage response under positive and negative position offsets is collected respectively to construct the inductance identification error signal and update the stator inductance of the cross-medium propulsion permanent magnet synchronous motor. To perform resistance identification, the amplitude of the SWC current injection is injected into the d-axis of the cross-medium propulsion permanent magnet synchronous motor, the corresponding q-axis current response is collected, a resistance identification error signal is constructed, and the stator resistance of the cross-medium propulsion permanent magnet synchronous motor is updated. The back electromotive force estimate from the SMO output and the electric angular velocity estimate from the PLL output are used to calculate the permanent magnet flux linkage and obtain the updated torque constant. The updated electrical parameters are obtained by combining the updated stator inductance, the updated stator resistance, and the updated torque constant of the cross-medium propulsion permanent magnet synchronous motor.

[0011] Furthermore, the step of calculating the adaptive injection intensity coefficient by combining the operating medium state of the cross-medium propulsion permanent magnet synchronous motor, the load torque observation value, and the electrical identification enable signal specifically includes: The injection intensity is determined by setting the corresponding medium gating gain according to the operating medium state of the cross-medium propulsion permanent magnet synchronous motor; The injection intensity is negatively adjusted based on the load torque observation value, and the electrical identification enable signal is injected into the main switch. When enabled, the calculated injection intensity is output, and when disabled, the injection intensity is set to zero, thus obtaining the adaptive injection intensity coefficient.

[0012] Furthermore, the step of identifying mechanical parameters and self-tuning the speed loop of the cross-medium propulsion permanent magnet synchronous motor based on the mechanical update gating signal and the updated electrical parameters, and synchronously outputting the load torque feedforward compensation to obtain the electromechanical parameter collaborative identification result, specifically includes: Based on the updated electrical parameters, the torque constant and q-axis current are used to calculate the new electromagnetic torque estimate. The load torque observation is set as a fast variable, while the equivalent moment of inertia, equivalent damping, and speed loop PI parameters of the cross-medium propulsion permanent magnet synchronous motor are set as slow variables. The load torque observation is updated in real time using the LIS-MAI observer to obtain the new load torque observation. Define the speed tracking error and load impact suppression factor, and construct an inertia update activator by combining the mechanical update gating signal. When the cross-medium propulsion permanent magnet synchronous motor is in a stable medium state and the speed error exceeds the preset dead zone threshold, the inverse inertia update is activated, and the equivalent rotational inertia is obtained after amplitude limiting. Using the recursive least squares method, based on the new load torque observations and mechanical angular velocity estimates, the original updated value of the equivalent damping is calculated according to the linear damping model. Based on the original updated values ​​of the equivalent moment of inertia, equivalent damping, and updated torque constant, combined with the desired damping ratio and desired natural frequency, the updated values ​​of the velocity loop proportional coefficient and integral coefficient are calculated. Based on the updated values ​​of the proportional and integral coefficients of the speed loop, the output of the speed loop PI controller is determined. The new load torque observation value is converted into the q-axis feedforward compensation current, and then superimposed and limited to determine the final q-axis current setpoint, thus obtaining the electromechanical parameter collaborative identification result.

[0013] The second technical solution adopted in this invention is: a cooperative identification system for electromechanical parameters for cross-medium propulsion, comprising: The first module is used to initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system and run basic magnetic field orientation control and sensorless observation to obtain the estimated mechanical angular velocity and the observed load torque. The second module is used to identify the operating medium state of the cross-medium propulsion permanent magnet synchronous motor based on the load torque observation value and the mechanical angular velocity estimate value, and generate electrical identification enable signal and mechanical update gating signal. The third module is used to perform online electrical parameter identification of the cross-medium propulsion permanent magnet synchronous motor based on the electrical identification enable signal using adaptive dual signal injection, and obtain updated electrical parameters. The fourth module is used to identify mechanical parameters and self-tune the speed loop of the cross-medium propulsion permanent magnet synchronous motor based on the mechanical update gating signal and the updated electrical parameters, and synchronously output the load torque feedforward compensation to obtain the electromechanical parameter collaborative identification result.

[0014] The beneficial effects of the method and system of this invention are as follows: This invention initializes the parameters of a cross-medium propulsion permanent magnet synchronous motor system and runs basic field-oriented control and sensorless observation to obtain estimated mechanical angular velocity and observed load torque. Then, based on the observed load torque and estimated mechanical angular velocity, the operating medium state of the cross-medium propulsion permanent magnet synchronous motor is identified, generating an electrical identification enable signal and a mechanical update gating signal. Medium discrimination characteristic quantities are constructed using the observed load torque and estimated mechanical angular velocity, and cross-medium state identification is achieved using a dual-threshold continuous and stable determination method. This eliminates the need for additional media sensors such as water pressure, liquid level, or humidity sensors, reducing system hardware complexity and improving state perception capabilities under air, underwater, and water-air switching conditions. Furthermore… Based on the electrical identification enable signal, an adaptive dual-signal injection method is adopted to perform online identification of electrical parameters for the cross-medium propulsion permanent magnet synchronous motor, obtaining updated electrical parameters. An adaptive dual-signal injection method is designed so that the amplitude of SWPO position offset injection and SWC current injection can be automatically adjusted according to the load state. This can reduce current ripple, torque fluctuation and thrust disturbance during underwater heavy load and water-to-air switching processes while ensuring the electrical parameter identification capability. Finally, based on the mechanical update gating signal and the updated electrical parameters, mechanical parameter identification and speed loop self-tuning are performed on the cross-medium propulsion permanent magnet synchronous motor, and the load torque feedforward compensation is output synchronously to obtain the electrical and mechanical parameter collaborative identification results, forming a closed-loop linkage update of electrical parameters, mechanical parameters and controller parameters. Attached Figure Description

[0015] Figure 1 This is a flowchart of the steps of a collaborative identification method for electromechanical parameters for cross-medium propulsion according to the present invention; Figure 2This is a structural block diagram of an electromechanical parameter collaborative identification system for cross-medium propulsion according to the present invention; Figure 3 This is a schematic diagram of the system framework for collaborative identification of electrical and mechanical parameters provided in a specific embodiment of the present invention. Detailed Implementation

[0016] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only for ease of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.

[0017] First, such as Figure 3 As shown, in this embodiment of the invention, initial parameters are first loaded and a basic speed estimate is established; then, the load torque is pre-observed using LIS-MAI; then, the medium state is identified and an identification gating is generated using the load torque and mechanical angular velocity; then, the electrical parameters are updated under a reliable electrical condition; finally, the mechanical slow variable and the speed loop slow variable are updated under a reliable mechanical condition, while maintaining real-time compensation for the load torque fast variable.

[0018] Reference Figure 1 This invention provides a method for collaborative identification of electromechanical parameters for cross-medium propulsion, the method comprising the following steps: S100. Initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system and run basic magnetic field orientation control and sensorless observation to obtain the estimated mechanical angular velocity and the observed load torque. S110. Initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system, including the initial values ​​of electrical parameter calibration, mechanical parameter calibration, and speed loop control parameters. In this embodiment, a power-on initialization, basic FOC, and sensorless speed prediction module are first set up. This module is used to provide the system with operable initial electrical parameters, mechanical parameters, speed loop parameters, and mechanical angular velocity estimates before the online identification closed loop is established, thereby providing basic input for subsequent LIS-MAI load torque pre-observation and cross-medium state identification.

[0019] After the system is powered on or reset, it first reads or loads the initial values ​​of the electrical parameter calibration, mechanical parameter calibration, and speed loop control parameters. The initial values ​​of the electrical parameters include the initial values ​​of stator resistance, stator inductance, permanent magnet flux linkage, and torque constant; the initial values ​​of the mechanical parameters include the initial values ​​of equivalent moment of inertia, inverse moment of inertia, equivalent damping, and load torque; the initial values ​​of the controller parameters include the initial values ​​of the speed loop proportional coefficient and integral coefficient. These initial values ​​can be obtained from the motor nameplate parameters, factory calibration, offline testing, no-load testing, or values ​​memorized from the previous operating condition.

[0020] S120, the default cross-medium propulsion permanent magnet synchronous motor system is in the medium transition state, electrical parameter updates, mechanical slow variable updates and speed loop slow variable updates are disabled, and the initial gating state is set; In this embodiment, the initial parameters can be expressed as: in: Before the initial medium state is reliably determined, the system defaults to a medium transition state and disables electrical parameter updates, slow mechanical variable updates, and slow speed loop variable updates. Its initial gating state can be set as follows: in, This indicates that the system will be temporarily treated as a media transition state during the initial startup phase. This indicates that online identification of electrical parameters is temporarily disabled. This indicates that slow variables such as equivalent moment of inertia, equivalent damping, and velocity loop PI parameters will not be updated for the time being.

[0021] S130. Based on the parameters of the initial cross-medium propulsion permanent magnet synchronous motor system, control the cross-medium propulsion permanent magnet synchronous motor to perform basic field orientation control, collect the three-phase current and bus voltage, and obtain the d-axis and q-axis current and corresponding voltage through Clarke / Park transformation. In this embodiment, the system then runs basic FOC control based on the aforementioned initial calibration values. Three-phase current and bus voltage are acquired and converted using Clarke / Park conversion. shaft and The shaft current is obtained, and the corresponding voltage is obtained by voltage reconstruction or controller output.

[0022] S140. Based on the sensorless control method, a back EMF observation model is constructed to estimate the d-axis and q-axis currents and corresponding voltages, and to obtain the initial electrical parameters, initial mechanical parameters and mechanical angular velocity estimates. In this embodiment, under sensorless control, the SMO+PLL uses the initial stator resistance, initial stator inductance, and initial flux linkage parameters to establish a back EMF observation model, and outputs estimated electrical angle, estimated electrical angular velocity, estimated mechanical angular velocity, and estimated back EMF.

[0023] The output of SMO+PLL can be expressed as: in, This serves as the foundational mechanical speed input for subsequent LIS-MAI load torque pre-observation and cross-medium state identification. For zero-speed or low-speed start-up phases, initial angle and speed information can be provided using initial rotor positioning, open-loop drive, or encoder assistance; once the back electromotive force is observable, the speed estimate output by SMO+PLL is switched to.

[0024] The output includes , or the previous moment , And various initial mechanical parameters. These quantities are fed into the next module to establish load torque observations before the medium condition determination is completed.

[0025] S150. Based on the initial electrical parameters, initial mechanical parameters, and estimated mechanical angular velocity, perform LIS-MAI load torque pre-observation to obtain the load torque observation value.

[0026] S151. Determine the estimated value of the current torque constant of the cross-medium propulsion permanent magnet synchronous motor based on the initial electrical parameters and initial mechanical parameters, and calculate the estimated value of the electromagnetic torque in combination with the q-axis current; In this embodiment, cross-medium state identification cannot be performed directly during the initial system startup because the required load torque observation values ​​have not yet been generated. To address this issue, this embodiment sets up a LIS-MAI load torque pre-observation stage before medium state identification. This stage utilizes initial electrical parameters, initial mechanical parameters, and the obtained mechanical angular velocity estimates to pre-observe the load torque, and updates the load torque as a fast mechanical variable in real time.

[0027] First, based on the current torque constant estimate and Calculate the electromagnetic torque estimate using shaft current: In the initial stage of startup Initial calibration value can be taken. During subsequent operation, Updated online by the electrical parameter identification module.

[0028] S152. Define the internal velocity observations and mechanical angular velocity estimates of LIS-MAI, and obtain the mechanical side observation errors; In this embodiment, the internal velocity observation value of LIS-MAI is defined as The estimated mechanical angular velocity is The mechanical side observation error is: S153. Define the load torque change rate and determine the adaptive robust gain; In this embodiment, to reduce chattering near the error zero point, LIS-MAI employs a hyperbolic tangent smoothing robust function: in, For smooth boundary layer parameters. When When it is large, Approaching the sign function, it can maintain strong robust correction capability; when When smaller, The approximate linear function makes the load torque prediction process smoother.

[0029] Further define the load torque change rate: In the initial stage of startup, it can be made Alternatively, the load change rate can be initialized using the memory value from the previous operating condition. To enhance the observation robustness under different load conditions, the adaptive robustness gain is set as follows: in, Basic robust gain, This indicates that the greater the load, the stronger the observation injection. This indicates that the faster the load changes, the stronger the observed injection.

[0030] S154. Using a hyperbolic tangent smoothing function, combined with adaptive robust gain, mechanical side observation error and electromagnetic torque estimation, the robust compensation term of LIS-MAI is determined, and the LIS-MAI velocity observation equation is constructed. In this embodiment, based on the hyperbolic tangent smoothing function and adaptive robust gain, the robust compensation term of LIS-MAI is defined as: in, This is the convergence gain for linear error. The compensation term consists of two parts: the first part... Used to ensure that velocity observation errors converge via linear feedback; Part Two This feature automatically enhances robustness correction under heavy loads or rapid load changes. The compensation allows LIS-MAI to maintain a small injection rate under stable light load conditions and automatically improve load torque tracking capability under heavy underwater loads or sudden changes in cross-medium loads.

[0031] The LIS-MAI velocity observation equation is: in, This is the nominal frictional or damping torque. This is an estimate of the inverse inertia. The initial pre-observation phase is mainly used to establish... ,therefore and You can first take the initial calibration value or the frozen value from the previous moment.

[0032] S155. Based on the LIS-MAI velocity observation equation, the load torque observation value of the cross-medium propulsion permanent magnet synchronous motor is updated in real time to obtain the load torque observation value.

[0033] In this embodiment, the load torque observation value is updated in real time according to the fast variable: in, This is the gain for load torque observation. This update is not subject to mechanical update gating. Freeze. In other words, during system startup, medium switching, or stable operation, LIS-MAI maintains real-time load torque monitoring capabilities and generates q-axis feedforward compensation current. Through this pre-observation phase, the system obtains the necessary information for subsequent cross-media state identification. and .

[0034] S200: Based on the load torque observation value and the mechanical angular velocity estimate, the operating medium state of the cross-medium propulsion permanent magnet synchronous motor is identified, and an electrical identification enable signal and a mechanical update gating signal are generated. S210. Based on the observed load torque and the estimated mechanical angular velocity, construct the characteristic quantities for medium discrimination. In this embodiment, after obtaining the estimated mechanical angular velocity value and LIS-MAI load torque observations Then, the system enters the cross-medium state recognition and identification scheduling stage. It does not rely on external water pressure, liquid level or humidity sensors, but uses the relationship between load torque and mechanical angular velocity to identify the medium state of the thruster, and generates electrical identification enable signal and mechanical update gating signal according to the medium state.

[0035] After obtaining the observed load torque and estimated mechanical angular velocity, construct the medium discrimination characteristic quantity: in, This is used to avoid the denominator being zero under low-speed operating conditions. At the same rotational speed, the air medium load is smaller. Smaller; larger underwater medium load, Larger load torque; drastic changes occur during water-to-air switching or partial immersion. They are often in the transition zone or cannot continuously and stably meet the threshold conditions.

[0036] S220. Set dual judgment thresholds and continuous stable judgment period, and define the first stable medium judgment function and the second stable medium judgment function. In this embodiment, two thresholds are set. And set a continuous judgment period. Define the first stable medium determination function and the second stable medium determination function: in, This is an indicator function that takes the value 1 when the condition is true and 0 when the condition is false.

[0037] S230. Determine the operating medium state of the cross-medium propulsion permanent magnet synchronous motor according to the first stable medium determination function and the second stable medium determination function; In this embodiment, the medium state variable is further defined: in, Indicates the first stable medium state. Indicates the second stable medium state. This indicates the medium transition state. For a water-air cross-medium propulsion system, the first stable medium state corresponds to the air stable state, the second stable medium state corresponds to the underwater stable state, and the medium transition state corresponds to the water ingress, water egress, or semi-submerged state.

[0038] It should be noted that this occurs when the system has just started up and has not yet accumulated enough continuous judgment cycles. Previously, the medium was in a transitional state, i.e. Electrical parameters, mechanical slow variables, and velocity loop slow variables all retain their initial values ​​or values ​​from the previous moment. Only when... The system enters the corresponding stable medium state only when the air stability threshold or the underwater stability threshold is continuously met.

[0039] S240. Define the electrical identification enable signal and the mechanical update gate signal according to the operating medium state of the cross-medium propulsion permanent magnet synchronous motor.

[0040] In this embodiment, after obtaining the medium state, an electrical identification enable signal is generated. and mechanically updated gate control signals The rate of change of mechanical angular velocity is defined as: The electrical identification enable signal is defined as: in, To allow for a threshold rate of change of speed for electrical identification, for Shaft current limit, and This refers to the allowable range of DC bus voltage. When, it allows the identification of electrical parameters such as inductance, resistance, and flux linkage; when At that time, electrical identification injection is turned off and electrical parameter updates are frozen.

[0041] The mechanical update gating adopts a minimal implementation: when When the system is in a stable medium state, it allows for slow updates of the equivalent moment of inertia, equivalent damping, and velocity loop variables; when At this time, it indicates that the system is in a medium transition state, freezing the slow mechanical variables and the slow velocity loop variables. It should be noted that... LIS-MAI continues to output without freezing load torque observations. This ensures that the system still has the ability to observe load torque and provide feedforward compensation during the water-to-air switching transient.

[0042] This module outputs , , , and .in, The data is sent to the electrical parameter identification module. The data is fed into the mechanical parameter identification and speed loop self-tuning modules. Thus, the system forms a sequential logic of "first establishing speed and load observations, then identifying the medium state, and finally scheduling parameter updates."

[0043] S300: Based on the electrical identification enable signal, an adaptive dual-signal injection is used to perform online identification of electrical parameters for the cross-medium propulsion permanent magnet synchronous motor, and the updated electrical parameters are obtained. S310. Calculate the adaptive injection intensity coefficient by combining the operating medium state of the cross-medium propulsion permanent magnet synchronous motor, the load torque observation value, and the electrical identification enable signal. Specifically, the injection intensity is determined by setting the corresponding medium gating gain according to the operating medium state of the cross-medium propulsion permanent magnet synchronous motor; the injection intensity is negatively adjusted in combination with the load torque observation value, and the electrical identification enable signal is injected into the main switch. When enabled, the calculated injection intensity is output, and when disabled, the injection intensity is set to zero, thus obtaining the adaptive injection intensity coefficient.

[0044] In this embodiment, after completing cross-medium state identification and scheduling, the system uses the electrical identification enable signal. The electrical parameter identification is performed based on the medium stability determination result. Unlike the fixed amplitude injection method, this embodiment calculates the injection intensity based on the medium state, load torque, and electrical identification enable signal, so that electrical identification is only performed under stable, observable conditions that do not significantly disturb the propulsion control.

[0045] First, construct the dielectric-gated gain based on the dielectric stability criterion function: in, The injection gain corresponding to the first stable medium state, The injection gain corresponds to the second stable medium state, and is typically taken as... When the system is in a medium transition state, and ,therefore .

[0046] Calculate the adaptive injection strength by combining load torque observations and electrical identification enable signals: in, This is the load torque suppression coefficient. This is the reference for rated torque or normalized torque. This expression indicates that the greater the load torque, the smaller the injection intensity; the injection intensity is zero when the system is in a medium transition state or when electrical identification is disabled.

[0047] S320. Generate the SWPO position offset injection amplitude and SWC current injection amplitude based on the injection intensity coefficient. In this embodiment, the SWPO position offset injection amplitude and SWC current injection amplitude are generated based on the adaptive injection intensity: when At that time, both of the above injection amplitudes are zero, that is, the electrical parameter identification injection is turned off.

[0048] S330. Perform inductance identification, inject SWPO position offset injection amplitude into the estimated rotor electrical angle of the cross-medium propulsion permanent magnet synchronous motor, collect the q-axis voltage response under positive and negative position offsets respectively, construct the inductance identification error signal, and update the stator inductance of the cross-medium propulsion permanent magnet synchronous motor. In this embodiment, during inductance identification, a SWPO square wave position offset is injected into the estimated rotor electrical angle: in Next, samples were collected under positive and negative position offsets. Axis voltage response and Construct an inductor identification error signal: And the stator inductance is updated using the LMS method: in, For inductor refresh rate, This refers to the physical limiting range of the inductor.

[0049] S340. Perform resistance identification, inject SWC current injection amplitude into the d-axis of the cross-medium propulsion permanent magnet synchronous motor, collect the corresponding q-axis current response, construct the resistance identification error signal, and update the stator resistance of the cross-medium propulsion permanent magnet synchronous motor. In resistance identification, this embodiment refers to... Shaft-injected SWC current disturbance, with two current levels set: Collect the corresponding Axis current response and Constructing a resistor identification error signal: And update the stator resistance: in, For the resistance update rate, [ [This refers to the physical limiting range of the resistor.]

[0050] All of the above updates are by Control: When electrical identification is disabled, the electrical parameters retain their previous values ​​or initial calibration values.

[0051] S350: Obtain the back electromotive force estimate from the SMO output and the electric angular velocity estimate from the PLL output, calculate the permanent magnet flux linkage, and obtain the updated torque constant. S360. By combining the updated stator inductance, the updated stator resistance, and the updated torque constant of the cross-medium propulsion permanent magnet synchronous motor, the updated electrical parameters are obtained.

[0052] In this embodiment, the system calculates the permanent magnet flux linkage using the back electromotive force estimate output by the SMO and the electric angular velocity estimate output by the PLL, and further obtains the torque constant: Electrical parameter identification module output , , and .in, and Feedback is sent to the SMO+PLL sensorless observer and the current loop decoupling compensation module. Used for calculation , The data is fed into the next module for mechanical-side electromagnetic torque estimation, speed loop self-tuning, and load torque feedforward compensation. Thus, the electrical parameter results positively correct the mechanical torque input, while the mechanical-side load state negatively adjusts the electrical injection intensity, forming an electrical-mechanical collaborative identification link.

[0053] S400: Based on the mechanical update gate signal and the updated electrical parameters, the mechanical parameters of the cross-medium propulsion permanent magnet synchronous motor are identified and the speed loop is self-tuned. The load torque feedforward compensation is output synchronously to obtain the electrical and mechanical parameter collaborative identification results.

[0054] S410. Calculate a new electromagnetic torque estimate based on the torque constant and q-axis current in the updated electrical parameters. Set the load torque observation as a fast variable and set the equivalent moment of inertia, equivalent damping, and speed loop PI parameters of the cross-medium propulsion permanent magnet synchronous motor as slow variables. Use the LIS-MAI observer to update the load torque observation in real time to obtain a new load torque observation. In this embodiment, after obtaining the estimated torque constant value through electrical parameter identification, the system enters the mechanical parameter identification, speed loop self-tuning, and disturbance compensation stage. This stage uses the load torque observation value output by LIS-MAI as a fast variable, and the equivalent moment of inertia, equivalent damping, and speed loop PI parameters as slow variables, with mechanical update gating. Unified scheduling of whether slow variables are updated.

[0055] First, based on the estimated torque constant and Calculate the electromagnetic torque estimate using shaft current: LIS-MAI continues to update the load torque in real time according to the aforementioned hyperbolic tangent smoothing function, adaptive robust gain, and velocity observation equation. It should be noted that this load torque observation process is not affected by... It is frozen, thus remaining effective in real time in both stable and transitional media states.

[0056] S420. Define the speed tracking error and load impact suppression factor, and construct an inertia update activator by combining the mechanical update gating signal. When the cross-medium propulsion permanent magnet synchronous motor is in a stable medium state and the speed error exceeds the preset dead zone threshold, the inverse inertia update is activated. After the amplitude limiting process, the equivalent rotational inertia is obtained. In this embodiment, to avoid contaminating the equivalent moment of inertia estimation due to load shocks, LIS-MAI employs a load shock suppression-type inertia update activator. Velocity tracking error is defined as follows: in, It can be obtained from a sensorless observer, SMO+PLL or encoder. It characterizes the deviation between the internal velocity prediction of LIS-MAI and the actual mechanical angular velocity estimate, and serves as a common error signal for load torque observation and inverse inertia update.

[0057] Define the load shock suppression factor and the inertia update activator: in, This is the load impact suppression coefficient. As a normalized reference for the rate of change of load torque, To update the base gain for inertia, This is the speed error dead zone threshold. When the system is in a stable medium state and the speed error exceeds the dead zone, inertia updates are allowed; when the system is in a medium transition state or the load torque changes drastically, inertia updates are disabled or weakened.

[0058] The inverse inertia update gain and the inverse inertia primitive update can be written as: Subsequently, only the inverse inertia is limited, and the equivalent moment of inertia is obtained from the limited inverse inertia. In this way, the system treats load abrupt changes as fast disturbances and compensates for them in real time, rather than immediately interpreting them as changes in the equivalent moment of inertia.

[0059] S430. Using the recursive least squares method, based on the new load torque observation and mechanical angular velocity estimate, the original updated value of the equivalent damping is calculated according to the linear damping model. In this embodiment, the equivalent damping is extracted from the load torque using minimum RLS. To reduce implementation complexity, a minimum linear damping model is adopted: make , Calculate the original updated values ​​of RLS gain and damping. Since the equivalent damping is a mechanically slow variable, its actual update is controlled by mechanical update gating. The damping update under mechanical gating is: Covariance synchronization uses gated updates: when At that time, RLS updates the equivalent damping and covariance based on the current load torque and mechanical angular velocity; when At this time, the equivalent damping and covariance remain at the values ​​of the previous moment, avoiding misidentification of transient load changes in the water inlet, outlet, or semi-submerged states as damping changes. This gating only applies to slow mechanical variables such as equivalent damping and equivalent moment of inertia, and does not freeze the load torque observations.

[0060] S440. Based on the original updated values ​​of the equivalent moment of inertia, equivalent damping, and updated torque constant, combined with the desired damping ratio and desired natural frequency, the updated values ​​of the velocity loop proportional coefficient and integral coefficient are calculated. In this embodiment, the mechanical parameters are identified , and electrical parameters identified This is further used for speed loop self-tuning. The speed loop object is approximated by a first-order mechanical model, and the new speed loop proportional and integral coefficients are calculated according to the desired second-order closed-loop model: in, For the desired damping ratio, The desired natural frequency is given. Although the proportional and integral coefficients of the speed loop participate in real-time speed control, their parameters are themselves slow variables in the speed loop. If the speed loop PI parameters are updated directly based on the instantaneous identification values ​​during water-to-air switching or load shock processes, it can easily cause sudden changes in the controller parameters.

[0061] S450. Based on the updated values ​​of the proportional and integral coefficients of the speed loop, the output of the speed loop PI controller is determined. The new load torque observation value is converted into the q-axis feedforward compensation current, and then superimposed and limited to determine the final q-axis current setpoint, thus obtaining the electromechanical parameter collaborative identification result.

[0062] In this embodiment, the mechanical update gate control is also used to determine whether the slow variable of the speed loop is updated. when When this occurs, it indicates that the system is in a stable medium state, and the mechanical parameter identification results have high reliability. The velocity loop PI parameters are based on the current... , and Reconfiguration; when When the system is in a state of water-air switching, semi-submersion, or load impact, the speed loop PI parameter remains at the value of the previous moment to avoid sudden changes in the controller's slow variable due to transient disturbances.

[0063] It should be noted that freezing the speed loop parameters does not mean that the speed loop control is turned off. During medium transition states, the speed error still participates in PI control, and the load torque obtained from real-time LIS-MAI observations is still used. Shaft feedforward compensation. Finally. The shaft current is given as follows: Therefore, this embodiment forms a control method that combines gated updates of mechanical slow variables with real-time compensation for fast disturbances: under stable medium conditions, the equivalent moment of inertia, equivalent damping, and velocity loop PI parameters are updated based on the electromechanical collaborative identification results; under medium transition conditions, the above-mentioned slow variables are frozen, but the load torque feedforward compensation remains effective in real time. This design can avoid abrupt changes in mechanical and velocity loop parameters caused by water-to-air switching shocks, while using load torque observations to quickly correct electromagnetic torque requirements, thereby reducing velocity drops and current surges during water ingress, water effluent, and semi-submersion conditions.

[0064] Therefore, the embodiments of the present invention have the following distinguishing technical features compared to the prior art: 1) A motor identification method with cross-medium capability is proposed, which integrates medium state identification, electrical parameter identification, sensorless observation, mechanical parameter identification, speed loop self-tuning, disturbance compensation and FOC into the control framework, so that electrical side parameters, mechanical side parameters and controller parameters are updated collaboratively according to air, water-air switching and underwater state.

[0065] 2) An adaptive dual-signal injection electrical parameter identification method is proposed, which constructs the injection intensity coefficient based on the load disturbance estimate, mechanical angular acceleration, q-axis current fluctuation, and immersion state change. The SWPO position offset and SWC current injection amplitude are adjusted online; injection is enhanced in the air stable state, injection is reduced in the water stable state, and injection is turned off transiently when switching between water and air.

[0066] 3) A cross-medium mechanical side parameter identification method based on LIS-MAI+RLS is proposed, which estimates the load torque using LIS-MAI. and equivalent moment of inertia The equivalent damping is extracted from the load torque and speed data by RLS. And freeze or enable the slow-changing parameter update based on the medium state.

[0067] 4) An electromechanical collaborative control output mechanism is proposed: a bidirectional coupling relationship is established between electrical parameter identification and mechanical parameter identification, so that the load torque, speed and medium state of the mechanical side participate in the enabling and injection regulation of electrical parameter identification. At the same time, the parameters such as flux linkage and torque constant obtained by electrical side identification correct the electromagnetic torque input of the mechanical side, and together they act on the identification and control method of speed loop self-tuning and disturbance compensation.

[0068] In summary, the present invention has the following advantages over the prior art: 1) By constructing medium discrimination characteristic quantities through load torque observation values ​​and mechanical angular velocity estimation values, and using dual threshold continuous stability judgment to realize cross-medium state recognition, without relying on additional medium sensors such as water pressure, liquid level or humidity, the system hardware complexity is reduced and the state perception capability under air, underwater and water-air switching conditions is improved.

[0069] 2) By using the medium state as the upper-level scheduling variable, the electrical parameter identification enablement, mechanical slow variable update and speed loop parameter update are uniformly controlled, so that the system can update parameters under stable medium state and freeze slow variable parameters under medium transition state, so as to avoid the contamination of parameter identification results by transient impacts such as water ingress, water effluent and semi-submersion.

[0070] 3) An adaptive dual-signal injection method is designed to automatically adjust the amplitudes of SWPO position offset injection and SWC current injection according to the load condition. Compared with the fixed injection amplitude method, this embodiment can reduce current ripple, torque fluctuation and thrust disturbance during underwater heavy load and water-to-air switching processes while ensuring the identification capability of electrical parameters.

[0071] 4) Form a closed-loop coordination between electrical and mechanical parameter identification: The mechanical load torque and medium state reversely adjust the electrical parameter identification enable and injection intensity; the flux linkage and torque constant obtained by online identification on the electrical side are used to correct the electromagnetic torque estimation and serve as inputs for mechanical parameter identification, speed loop self-tuning and load feedforward compensation, thereby forming a closed-loop linkage update of electrical parameters, mechanical parameters and controller parameters.

[0072] Reference Figure 2 A collaborative identification system for electromechanical parameters for cross-medium propulsion, comprising: The first module 201 is used to initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system and run basic magnetic field orientation control and sensorless observation to obtain the estimated mechanical angular velocity and the observed load torque. The second module 202 is used to identify the operating medium state of the cross-medium propulsion permanent magnet synchronous motor based on the load torque observation value and the mechanical angular velocity estimate, and generate an electrical identification enable signal and a mechanical update gating signal. The third module 203 is used to perform online electrical parameter identification of the cross-medium propulsion permanent magnet synchronous motor based on the electrical identification enable signal using adaptive dual signal injection, and obtain updated electrical parameters. The fourth module 204 is used to identify mechanical parameters and self-tune the speed loop of the cross-medium propulsion permanent magnet synchronous motor based on the mechanical update gating signal and the updated electrical parameters, and synchronously output the load torque feedforward compensation amount to obtain the electromechanical parameter collaborative identification result.

[0073] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0074] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this is not intended to limit the scope of the embodiments of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A method for cooperative identification of electromechanical parameters for cross-medium propulsion, characterized in that, Includes the following steps: Initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system and run basic field orientation control and sensorless observation to obtain the estimated mechanical angular velocity and the observed load torque. Based on the load torque observation and mechanical angular velocity estimation, the operating medium state of the cross-medium propulsion permanent magnet synchronous motor is identified, and an electrical identification enable signal and a mechanical update gating signal are generated. Based on the electrical identification enable signal, an adaptive dual-signal injection method is used to perform online identification of electrical parameters for a cross-medium propulsion permanent magnet synchronous motor, and the updated electrical parameters are obtained. Based on the mechanical update gating signal and the updated electrical parameters, mechanical parameter identification and speed loop self-tuning are performed on the cross-medium propulsion permanent magnet synchronous motor, and the load torque feedforward compensation is output synchronously to obtain the electromechanical parameter collaborative identification result.

2. The method for cooperative identification of electromechanical parameters for cross-medium propulsion according to claim 1, characterized in that, The step of initializing the parameters of the cross-medium propulsion permanent magnet synchronous motor system and running basic field-oriented control and sensorless observation to obtain the estimated mechanical angular velocity and observed load torque specifically includes: Initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system, including the initial values ​​of electrical parameter calibration, mechanical parameter calibration, and speed loop control parameters; By default, the cross-medium propulsion permanent magnet synchronous motor system is in the medium transition state, and the updates of electrical parameters, mechanical slow variables, and speed loop slow variables are disabled, and the initial gating state is set. Based on the parameters of the initial cross-medium propulsion permanent magnet synchronous motor system, the cross-medium propulsion permanent magnet synchronous motor is controlled to perform basic field orientation control. The three-phase current and bus voltage are collected, and the d-axis and q-axis currents and corresponding voltages are obtained through Clarke / Park transformation. Based on the sensorless control method, a back EMF observation model is constructed to estimate the d-axis and q-axis currents and corresponding voltages, thereby obtaining the initial electrical parameters, initial mechanical parameters and mechanical angular velocity estimates. Based on the initial electrical parameters, initial mechanical parameters, and estimated mechanical angular velocity, LIS-MAI load torque pre-observation is performed to obtain the load torque observation value.

3. The method for cooperative identification of electromechanical parameters for cross-medium propulsion according to claim 2, characterized in that, The step of performing LIS-MAI load torque pre-observation based on initial electrical parameters, initial mechanical parameters, and estimated mechanical angular velocity to obtain the load torque observation value specifically includes: Based on the initial electrical and mechanical parameters, the current torque constant estimate of the cross-medium propulsion permanent magnet synchronous motor is determined, and the electromagnetic torque estimate is calculated in conjunction with the q-axis current. Define the internal velocity observations and mechanical angular velocity estimates of LIS-MAI, and obtain the mechanical side observation errors; Define the load torque change rate and determine the adaptive robust gain; By employing a hyperbolic tangent smoothing function, combined with adaptive robust gain, mechanical side observation error, and electromagnetic torque estimation, the robust compensation term of LIS-MAI is determined, and the LIS-MAI velocity observation equation is constructed. Based on the LIS-MAI velocity observation equation, the load torque observation value of the cross-medium propulsion permanent magnet synchronous motor is updated in real time to obtain the load torque observation value.

4. The method for cooperative identification of electromechanical parameters for cross-medium propulsion according to claim 3, characterized in that, The step of identifying the operating medium state of the cross-medium propulsion permanent magnet synchronous motor based on the load torque observation value and the mechanical angular velocity estimate, and generating the electrical identification enable signal and the mechanical update gating signal, specifically includes: Based on the observed load torque and the estimated mechanical angular velocity, construct the characteristic quantities for medium discrimination; Set dual judgment thresholds and continuous stability judgment period, and define the first stable medium judgment function and the second stable medium judgment function. The operating medium state of the cross-medium propulsion permanent magnet synchronous motor is determined based on the first stable medium determination function and the second stable medium determination function. Based on the operating medium state of the cross-medium propulsion permanent magnet synchronous motor, the electrical identification enable signal and the mechanical update gating signal are defined.

5. The method for cooperative identification of electromechanical parameters for cross-medium propulsion according to claim 4, characterized in that, The step of performing online electrical parameter identification of the cross-medium propulsion permanent magnet synchronous motor based on the electrical identification enable signal using adaptive dual-signal injection to obtain updated electrical parameters specifically includes: The adaptive injection intensity coefficient is calculated by combining the operating medium state, load torque observation, and electrical identification enable signal of the cross-medium propulsion permanent magnet synchronous motor. Based on the injection intensity coefficient, the SWPO position offset injection amplitude and the SWC current injection amplitude are generated; Inductance identification is performed, and the SWPO position offset injection amplitude is injected into the estimated rotor electrical angle of the cross-medium propulsion permanent magnet synchronous motor. The q-axis voltage response under positive and negative position offsets is collected respectively to construct the inductance identification error signal and update the stator inductance of the cross-medium propulsion permanent magnet synchronous motor. To perform resistance identification, the amplitude of the SWC current injection is injected into the d-axis of the cross-medium propulsion permanent magnet synchronous motor, the corresponding q-axis current response is collected, a resistance identification error signal is constructed, and the stator resistance of the cross-medium propulsion permanent magnet synchronous motor is updated. The back electromotive force estimate from the SMO output and the electric angular velocity estimate from the PLL output are used to calculate the permanent magnet flux linkage and obtain the updated torque constant. The updated electrical parameters are obtained by combining the updated stator inductance, the updated stator resistance, and the updated torque constant of the cross-medium propulsion permanent magnet synchronous motor.

6. The method for cooperative identification of electromechanical parameters for cross-medium propulsion according to claim 5, characterized in that, The step of calculating the adaptive injection intensity coefficient by combining the operating medium state, load torque observation, and electrical identification enable signal of the cross-medium propulsion permanent magnet synchronous motor specifically includes: The injection intensity is determined by setting the corresponding medium gating gain according to the operating medium state of the cross-medium propulsion permanent magnet synchronous motor; The injection intensity is negatively adjusted based on the load torque observation value, and the electrical identification enable signal is injected into the main switch. When enabled, the calculated injection intensity is output, and when disabled, the injection intensity is set to zero, thus obtaining the adaptive injection intensity coefficient.

7. The method for cooperative identification of electromechanical parameters for cross-medium propulsion according to claim 6, characterized in that, The step of identifying mechanical parameters and self-tuning the speed loop of a cross-medium propulsion permanent magnet synchronous motor based on the mechanical update gating signal and the updated electrical parameters, and synchronously outputting the load torque feedforward compensation to obtain the electromechanical parameter collaborative identification result, specifically includes: Based on the updated electrical parameters, the torque constant and q-axis current are used to calculate the new electromagnetic torque estimate. The load torque observation is set as a fast variable, while the equivalent moment of inertia, equivalent damping, and speed loop PI parameters of the cross-medium propulsion permanent magnet synchronous motor are set as slow variables. The load torque observation is updated in real time using the LIS-MAI observer to obtain the new load torque observation. Define the speed tracking error and load impact suppression factor, and construct an inertia update activator by combining the mechanical update gating signal. When the cross-medium propulsion permanent magnet synchronous motor is in a stable medium state and the speed error exceeds the preset dead zone threshold, the inverse inertia update is activated, and the equivalent rotational inertia is obtained after amplitude limiting. Using the recursive least squares method, based on the new load torque observations and mechanical angular velocity estimates, the original updated value of the equivalent damping is calculated according to the linear damping model. Based on the original updated values ​​of the equivalent moment of inertia, equivalent damping, and updated torque constant, combined with the desired damping ratio and desired natural frequency, the updated values ​​of the velocity loop proportional coefficient and integral coefficient are calculated. Based on the updated values ​​of the proportional and integral coefficients of the speed loop, the output of the speed loop PI controller is determined. The new load torque observation value is converted into the q-axis feedforward compensation current, and then superimposed and limited to determine the final q-axis current setpoint, thus obtaining the electromechanical parameter collaborative identification result.

8. A cooperative identification system for electromechanical parameters for cross-medium propulsion, characterized in that, Includes the following modules: The first module is used to initialize the parameters of the cross-medium propulsion permanent magnet synchronous motor system and run basic magnetic field orientation control and sensorless observation to obtain the estimated mechanical angular velocity and the observed load torque. The second module is used to identify the operating medium state of the cross-medium propulsion permanent magnet synchronous motor based on the load torque observation value and the mechanical angular velocity estimate value, and generate electrical identification enable signal and mechanical update gating signal. The third module is used to perform online electrical parameter identification of the cross-medium propulsion permanent magnet synchronous motor based on the electrical identification enable signal using adaptive dual signal injection, and obtain updated electrical parameters. The fourth module is used to identify mechanical parameters and self-tune the speed loop of the cross-medium propulsion permanent magnet synchronous motor based on the mechanical update gating signal and the updated electrical parameters, and synchronously output the load torque feedforward compensation to obtain the electromechanical parameter collaborative identification result.

Images