A ship propulsion motor control method based on back electromotive force estimation

Abstract

The present application relates to the field of ship propulsion device control, more particularly, the present application relates to a ship propulsion motor control method based on back electromotive force estimation, the method comprises: calculating the rate of change difference based on the stator multiphase current data and the original torque instruction, and extracting the wave disturbance characteristic value representing the transient impact intensity of sea waves; the first ratio of the sum of the characteristic value, the rated phase voltage and the anti-zero constant is used to dynamically widen the cutoff frequency, and the initial estimated phase angle and the actual electrical angular velocity are obtained in combination with the frequency; the first ratio is used to compensate the phase compensation angle to offset the hardware delay error, and the final rotor electrical angle is obtained by addition; the first ratio is used to construct the damping coefficient to attenuate and soften the original torque instruction, and the final control torque with flexible buffering capacity is obtained and converted into the bottom driving signal. The present application eliminates the observation phase lag phenomenon in severe sea conditions, realizes the intelligent flexible unloading mechanism, and prevents the rigid fracture of the transmission shaft system.

Description

Technical Field

[0001] This invention relates to the field of ship propulsion system control. More specifically, this invention relates to a ship propulsion motor control method based on back electromotive force estimation. Background Technology

[0002] The mainstream implementation path of ship propulsion control systems typically uses a traditional back EMF estimation model as the core sensing unit, paired with a fixed-bandwidth low-pass filter to form a signal processing module. This is supplemented by an outer-loop rigid torque direct output control strategy, forming a closed-loop architecture of observation-filtering-command execution. Its core objective is to achieve accurate extraction of the motor's back EMF and real-time calculation of the rotor position. This conventional technical solution is usually based on idealized physical assumptions, namely, that the ship's propeller always operates in a stable and uniform hydrodynamic environment, and that the mechanical load characteristics of the ship's transmission shaft system exhibit a continuous and gradual change trend, without sudden, strongly coupled operational disturbances.

[0003] However, in actual maritime operations, ships inevitably encounter severe sea conditions such as strong winds and giant waves. When waves continuously act on the propeller blades in a periodic, strong impact, the originally stable microscopic physical operating conditions change. During the dynamic process of instantaneous wave impact, the mechanical impedance of the ship's transmission shaft system exhibits nonlinear abrupt changes, and the cross-axis current of the propulsion motor also experiences severe and irreversible transient fluctuations, forming a strong disturbance response of mechanical-electrical coupling. At this time, the inherent defects of conventional control systems become fully apparent: on the one hand, the controller fails to perceive the abrupt changes in the operating conditions and still adheres to a fixed bandwidth filtering strategy based on a smooth load preset. This signal smoothing mechanism, which is reasonable under stable operating conditions, becomes a key problem of blindly blocking the effective components of the high-frequency real back electromotive force when facing the motor output signal after high-frequency impact, resulting in distortion of the core signal required for position calculation; on the other hand, the rigid full torque command output by the outer loop forms a direct mechanical confrontation with the physical shaft system that is instantaneously blocked by giant waves, causing a strong conflict between the command and the load, further exacerbating the instability of the system operation.

[0004] In summary, the core flaw of existing conventional control schemes lies in their control models being built upon a fixed parameter system. These fixed parameters can only adapt to idealized, stable operating conditions and are ill-suited to the wide-frequency, strongly coupled dynamic load transient processes induced by wave impacts. This mismatch between parameters and operating conditions not only leads to significant physical phase lag in the output signal of traditional back-EMF estimation models, directly affecting the accuracy of rotor position calculations, but also causes a series of engineering hazards in the mechanical execution stage, including increased reactive excitation losses in the motor, intensified torsional vibration of the transmission shaft, and accelerated fatigue damage to the mechanical structure. These severely restrict the system's operational reliability and control performance under complex sea conditions. Therefore, it is urgently necessary to introduce key physical quantities that can accurately characterize transient impact intensity, load change rate, and coupling characteristics to implement real-time dynamic adaptive adjustments to the control parameters, thereby overcoming the performance bottleneck of conventional schemes under harsh sea conditions. Summary of the Invention

[0005] To address the problem in existing technologies where fixed model parameters cannot adapt to dynamic load transients, leading to observer phase lag and mechanical resistance, this invention proposes a ship propulsion motor control method based on back EMF estimation. This method includes the following steps: Collect stator multiphase current data and original torque command, calculate the difference between the cross-axis current change rate of the stator multiphase current data and the change rate of the original torque command, and obtain wave disturbance characteristic value used to characterize the transient impact intensity of ocean waves; Collect stator voltage data, calculate the first ratio of the wave interference characteristic value to the rated phase voltage calibration constant, use the first ratio to broaden the base cutoff frequency calibration value to obtain the dynamic cutoff frequency for adapting to high-frequency disturbances, and combine the dynamic cutoff frequency observation to obtain the initial estimated phase angle and the actual electric angular velocity of the motor characterizing the base flux linkage position. Calculate the angular velocity correction coefficient mapped from the first ratio, and multiply the angular velocity correction coefficient by the actual electric angular velocity of the motor to obtain the phase compensation angle used to offset hardware delay errors; The initial estimated phase angle is added to the phase compensation angle to obtain the final rotor electrical angle that aligns with the true magnetic field of the rotor. The damping coefficient constructed based on the first ratio is multiplied by the original torque command to obtain the final control torque with flexible buffering capability. The final rotor electrical angle and the final control torque are converted into drive signals and applied to the power switching device to drive the motor.

[0006] This invention extracts wave interference characteristic values ​​to perceive the transient impact intensity of external ocean waves in real time. It then uses these wave interference characteristic values ​​to dynamically calculate and broaden the basic cutoff frequency calibration value, effectively preventing the high-frequency real back electromotive force signal from being filtered out by a fixed bandwidth, thus ensuring the accuracy of the initial estimated phase angle calculation. Simultaneously, it compensates for angular deviations caused by hardware communication delays by calculating a phase compensation angle, and uses a damping coefficient to soften the original torque command, generating a final control torque with flexible buffering capabilities. This method effectively reduces the mechanical stress concentration experienced by the drive shaft system under extreme sea conditions, improving the overall operational stability of the ship's propulsion control system.

[0007] Preferably, the method for calculating the wave interference characteristic value is as follows: The rate of change of the quadrature-axis current is obtained based on the basic sampling step size and the difference between the instantaneous values ​​of the quadrature-axis current at adjacent sampling times. The rate of change of the original torque command is obtained based on the basic sampling step size and the difference between the original torque command at adjacent sampling times. The rate of change of the quadrature-axis current is weighted using the calibration value of the quadrature-axis inductance of the propulsion motor stator to obtain a first product. The rate of change of the original torque command is weighted using the equivalent conversion constant from torque to voltage to obtain a second product. The absolute value of the difference between the first product and the second product is calculated. The absolute value of the difference is then summed with the wave interference characteristic value at the previous moment using the basic low-pass filter attenuation constant, and the sum is determined as the wave interference characteristic value at the current moment.

[0008] This invention accurately isolates the conventional data fluctuations caused by the ship's active maneuvering by extracting the difference between the rate of change of current and torque commands, retaining only the abrupt changes caused by external wave impact. Furthermore, it utilizes a basic low-pass filter attenuation constant to perform a first-order inertial weighted summation to smooth out extremely short-term high-frequency random noise from the sensor, ensuring that the output wave interference characteristic values ​​accurately and stably reflect the continuous impact energy of the waves.

[0009] Preferably, the specific calculation process for obtaining the dynamic cutoff frequency to adapt to high-frequency disturbances by using the first ratio to broaden the basic cutoff frequency calibration value is as follows: Multiply the bandwidth extension sensitivity coefficient by the first ratio at the current moment to obtain the bandwidth extension amount; The basic cutoff frequency calibration value is added to the bandwidth extension value to determine the dynamic cutoff frequency at the current moment.

[0010] This invention employs linear broadening logic that multiplies the bandwidth extension sensitivity coefficient by a first ratio to effectively reduce the underlying computing power overhead of the controller. This allows the dynamic cutoff frequency to be proportionally amplified in real time as the intensity of external wave disturbances increases, ensuring that the observer can quickly open the high-frequency signal channel under high-frequency alternating load conditions and avoiding the truncation of the effective flux linkage position signal by conventional filtering strategies.

[0011] Preferably, the method for obtaining the bandwidth extension sensitivity coefficient includes: In the state of towing wave generation in the water tank, simulated wave impact load is applied to the propeller model, and the peak data of wave disturbance characteristic values ​​are recorded simultaneously. By artificially increasing the cutoff frequency extension, the dynamic deviation between the phase angle estimated by the observer and the actual phase angle is made to converge and stabilize within the preset physical safety allowable error band. The minimum cutoff frequency extension that just meets the convergence condition is recorded. The quotient obtained by dividing the minimum cutoff frequency extension by the ratio of the peak data of the wave interference characteristic value to the rated phase voltage calibration constant is determined as the bandwidth extension sensitivity coefficient.

[0012] Preferably, the specific calculation process for obtaining the phase compensation angle used to offset hardware delay errors by multiplying the angular velocity correction coefficient mapped from the first ratio by the actual electrical angular velocity of the motor is as follows: The power function value is calculated with the natural constant as the base and the opposite of the first ratio at the current moment as the exponent; the difference between the value 1 and the power function value is multiplied by the feedforward lead gain constant to determine the angular velocity correction coefficient at the current moment. The angular velocity correction coefficient is multiplied by the actual electrical angular velocity of the motor at the current moment to obtain the effective transient electrical angular velocity within the delay window; the effective transient electrical angular velocity is multiplied by the inherent hardware communication delay constant to determine the phase compensation angle at the current moment.

[0013] This invention utilizes a natural constant as the base and the opposite of a first ratio to construct an exponential function value, thereby realizing a nonlinear adaptive compensation mechanism. This mechanism can dynamically release compensation weights according to the severity of the impact energy, effectively extracting the lost electrical angle within the inherent hardware communication delay constant window.

[0014] Preferably, the method for obtaining the feedforward lead gain constant includes: When the ship's propulsion motor is unloaded and running at its rated full speed, capture the time difference between the moment when the microprocessor sends the drive signal and the moment when the actual stator current crosses zero. The actual physical lag angle is calculated by multiplying the time difference by the electrical angular velocity at the rated full speed. The product of the electrical angular velocity at the rated full speed and the inherent hardware communication delay constant is calculated as the theoretical nominal lag angle. The actual physical lag angle is divided by the theoretical nominal lag angle, and the resulting ratio is determined as the feedforward lead gain constant.

[0015] Preferably, the specific calculation process for multiplying the damping coefficient constructed based on the first ratio with the original torque command to obtain the final control torque with flexible buffering capability is as follows: The product value is obtained by multiplying the flexible unloading damping coefficient by the first ratio at the current moment, and the negative number of the product value is determined as the attenuation index; The damping coefficient is obtained by calculating using the natural constant as the base and the attenuation index as the exponent; The original torque command at the current moment is multiplied by the damping coefficient to determine the final control torque at the current moment.

[0016] This invention uses the negative of the product of the flexible unloading damping coefficient and the first ratio as the attenuation exponent, and uses this to calculate the power function of the natural constant to obtain the damping coefficient. When the propeller encounters severe physical jamming, this damping coefficient will decrease rapidly, thereby proportionally and dynamically reducing and softening the original torque command issued by the energy management system. This calculation process actively abandons the extremely dangerous full output of rigid torque, and instead outputs a tentative holding torque within a safe range to the actuator.

[0017] Preferably, the method for obtaining the flexible unloading damping coefficient includes: A sweep frequency excitation torque is applied to both ends of the ship's long shaft system, and the stiffness attenuation damping ratio of the ship's long shaft system is extracted. The maximum safe yield limit torque of the ship's long shaft system is obtained, and the maximum safe yield limit torque is divided by the product of the stiffness attenuation damping ratio and the rated output torque of the propulsion motor. The resulting ratio is determined as the flexible unloading damping coefficient.

[0018] Preferably, the specific operation of converting the final rotor electrical angle and the final control torque into a drive signal includes: The final control torque is used to generate a quadrature axis current reference command, and the final rotor electrical angle is used as the magnetic field orientation reference to perform an inverse coordinate transformation to generate a stator voltage vector command. The stator voltage vector command is sent to the space vector pulse width modulation module and converted into a gate trigger pulse sequence as the driving signal. When the vessel is identified as being in a shallow water state and forced to break free, the bypass switch is triggered, and the original torque command is directly converted into the gate trigger pulse sequence as the final control torque.

[0019] Preferably, the step of obtaining the initial estimated phase angle and the actual electric angular velocity of the motor characterizing the position of the base flux linkage by combining the dynamic cutoff frequency observation includes: Based on the motor calibration parameters, the original back electromotive force observation components of the stator voltage data and stator multiphase current data in the stationary coordinate system are calculated; Based on the dynamic cutoff frequency, the original back EMF observation component is dynamically filtered to obtain the filtered back EMF component. The zero-prevention constant is introduced into the filtered back EMF component for trigonometric function calculation to obtain the initial estimated phase angle; the actual electric angular velocity of the motor is calculated based on the time change rate of the initial estimated phase angle.

[0020] The present invention has the following beneficial effects: This invention achieves objective quantitative perception of the transient impact intensity of external ocean waves by accurately calculating wave interference characteristic values. Based on this perception result, the dynamic cutoff frequency is broadened in real time, breaking the signal cutoff limitation of traditional fixed-parameter filters under high-frequency alternating loads. This ensures the smooth transmission of the effective high-frequency signal of the back electromotive force, thereby significantly improving the overall calculation accuracy of the initial estimated phase angle and the actual electric angular velocity of the motor under forced abrupt changes.

[0021] Secondly, this invention addresses the angular deviation caused by inherent hardware communication delays and power device dead zones by introducing an angular velocity correction coefficient and a phase compensation angle based on eigenvalue dynamic mapping. This mechanism, based on nonlinear exponential mapping, not only dynamically extracts and compensates for lost motion electrical angles within the delay window as impact intensity increases, but also possesses a flexible saturation limiting property, effectively preventing divergent overshoot of control parameters under harsh operating conditions and robustly ensuring high-precision alignment between the stator magnetic field and the actual rotor position.

[0022] Finally, under extreme physical conditions such as the propeller encountering abnormally large waves or getting stuck due to foreign objects, the damping coefficient constructed based on the first ratio and physical calibration parameters can quickly take effect, softening the originally highly destructive rigid torque command into a final control torque with flexible buffering capabilities. This yielding strategy effectively avoids the risk of torsional resonance or rigid breakage of the transmission shaft caused by the drive system forcibly outputting maximum electromagnetic torque. At the same time, by introducing bypass switch logic in the case of forced escape from shallow water, the mechanical protection in harsh sea conditions and the full escape requirements in special grounding crises are taken into account, significantly improving the environmental adaptability, operational reliability, and service life of the ship's propulsion control system. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the steps of a ship propulsion motor control method based on back electromotive force estimation provided in an embodiment of the present invention. Figure 2 This is a diagram illustrating the dynamic extraction effect of wave interference feature values ​​provided in an embodiment of the present invention. Figure 3 This is a comparison diagram of the dynamic cutoff frequency adaptive effect provided in the embodiment of the present invention; Figure 4This is a comparison chart of the effect of feedforward compensation on hysteresis phase error suppression provided in the embodiments of the present invention; Figure 5 This is a comparison chart of the control torque flexible unloading protection effect provided in the embodiments of the present invention. Detailed Implementation

[0024] Please see Figure 1 The diagram illustrates a flowchart of a ship propulsion motor control method based on back electromotive force estimation provided in Embodiment 1. The method includes the following steps: S1: Collect stator multiphase current data and original torque command, calculate the difference between the cross-axis current change rate of the stator multiphase current data and the change rate of the original torque command, and obtain wave disturbance characteristic value used to characterize the transient impact intensity of ocean waves.

[0025] It should be noted that when a ship traverses a severe wave-breaking zone, the violent impact of waves on the propeller directly causes a sudden change in the impedance of the transmission shaft system. Due to the inherent characteristics of this type of impedance transient, a single dimension of speed parameter or absolute current amplitude change is insufficient to effectively distinguish between normal ship maneuvering conditions and abnormal conditions caused by external wave impact. Conventional control methods lack the ability to accurately identify these two types of conditions, easily misinterpreting transient inductive distortions as normal fluctuations, thus generating false action responses that do not meet actual needs. Therefore, the core objective of this step is to accurately isolate the active maneuvering interference component from the impedance abrupt change information and effectively extract wave impact characteristic information.

[0026] Preferably, as an example, stator multiphase current data and original torque command are collected, and the difference between the cross-axis current change rate of the stator multiphase current data and the change rate of the original torque command is calculated to obtain wave disturbance characteristic values ​​used to characterize the transient impact intensity of ocean waves, including: First, stator multiphase current data is acquired by Hall current sensors installed on the output busbar of the ship propulsion inverter, and coordinate transformation is performed to obtain the instantaneous value of the quadrature axis current. At the same time, the original torque command is obtained through the ship energy management system communication bus.

[0027] Next, wave disturbance characteristic values ​​are calculated based on the instantaneous value of the quadrature-axis current and the original torque command, specifically satisfying the following relationship:

[0028] In the formula, Represents the current time sequence index. Representing the The wave disturbance characteristic value at each moment, in volts; Representing the The wave disturbance characteristic value at each moment, in volts; This represents the basic low-pass filter attenuation constant, which is dimensionless. For example... Take 0.92; This represents the rated value of the stator quadrature-axis inductance of the propulsion motor, and its unit is Henry. Representing the The instantaneous value of the quadrature-axis current at each moment, in amperes; Representing the The instantaneous value of the quadrature-axis current at each moment, in amperes; This represents the basic sampling step size, and its unit is seconds; The equivalent conversion constant from torque to voltage is represented by the unit (volt-second) / (newton-meter); Representing the The original torque command at each moment, in Newton-meters; Representing the The original torque command at each moment, measured in Newton-meters. Represents the first product, which is based on Faraday's law of electromagnetic induction: The transformation yields, This represents the rate of change of current, and is related to the above formula. Part of For inductance, corresponding to the relationship above. , This represents the second product, which is based on the existing electromagnetic transformation equations. The transformation yields, Indicates electromagnetic torque. Represents the torque constant. To represent current, the specific transformation process is as follows: [The equation is incomplete and requires further context to be fully translated.] Differentiating both sides yields The above Replacing Faraday's law of electromagnetic induction In get ,in The formula for calculating the characteristic value of corresponding wave interference , The formula for calculating the characteristic value of corresponding wave interference .

[0029] Understandably, when a ship encounters a sudden, severe cross wave impacting the propeller, the instantaneous value of the quadrature shaft current will drastically change within a very short time due to the momentary jamming of the drive shaft system. Meanwhile, the initial torque command issued by the energy management system typically maintains the original, gradual thrust. Therefore, the quadrature shaft current variation term in the above formula will... This will be significantly larger than the original torque command change. The absolute value of the difference surged, which in turn drove up the wave interference characteristic value. It grows rapidly.

[0030] Meanwhile, the impact of a real giant wave is a continuous squeezing process carrying enormous mechanical kinetic energy, not a fleeting electromagnetic disturbance. Therefore, under this forced abrupt change condition, the eigenvalue of the previous moment in the relational formula... With decay constant The historical inertia term begins to play a core role in energy continuity. Within several consecutive sampling periods immediately following the impact, the characteristic values ​​from the previous moment... As wave energy continues to be injected, it will exhibit a gradually increasing upward trend; this continuously rising historical state value is weighted by the current surge in the difference term, resulting in a final output wave disturbance characteristic value. It exhibits a stable and interference-resistant increasing trend.

[0031] Since the wave interference characteristic value is derived recursively from the wave interference characteristic value of the previous moment, it is necessary to provide the wave interference characteristic value of the initial moment. Additionally, the above embodiments also involve the calibration value of the stator quadrature-axis inductance of the propulsion motor. and the basic low-pass filter attenuation constant The method for obtaining the above parameters will be explained below.

[0032] The method for obtaining the wave disturbance characteristic values ​​at the initial moment includes: Since the ship's control system does not accumulate transient impact energy due to wave impact during each cold start or during long-term calm cruise in still waters without wave interference, the initial wave interference characteristic value is preset to 0 volts.

[0033] In addition, the stator quadrature axis inductance calibration value of the propulsion motor Methods for obtaining [the information] include: The stator quadrature axis inductance rating of a motor reflects the inherent hardware electromagnetic properties of the motor. It can be used as the stator quadrature axis inductance rating of a propulsion motor by directly reading the rating inductance data from the nameplate of the thrust motor.

[0034] S2: Collect stator voltage data, calculate the first ratio of the wave interference characteristic value to the rated phase voltage calibration constant, use the first ratio to broaden the base cutoff frequency calibration value to obtain the dynamic cutoff frequency for adapting to high-frequency disturbances, and combine the dynamic cutoff frequency observation to obtain the initial estimated phase angle and the actual electric angular velocity of the motor characterizing the base flux linkage position.

[0035] It should be noted that traditional back EMF estimation models often employ a fixed low-pass filter bandwidth. While this method can effectively filter out high-frequency noise under stable operating conditions, in real-world scenarios where high-frequency transients in the back EMF of a motor are induced by giant wave impacts, the fixed narrow bandwidth will filter out the effective high-frequency signal, leading to the loss of useful real back EMF information and severe lag in the observed phase. Therefore, the core objective of this step is to utilize the transient impact characteristics to construct an adaptive bandwidth expansion operator, dynamically opening high-frequency channels and achieving lag-free and accurate phase observation.

[0036] Preferably, as an example, stator voltage data is collected, a first ratio of the wave interference characteristic value to the rated phase voltage calibration constant is calculated, and the first ratio is used to broaden the base cutoff frequency calibration value to obtain a dynamic cutoff frequency for adapting to high-frequency disturbances. Combined with the dynamic cutoff frequency observation, an initial estimated phase angle characterizing the base flux linkage position and the actual electric angular velocity of the motor are obtained, including: First, the dynamic cutoff frequency is calculated based on stator voltage data and wave interference characteristic values, specifically satisfying the following relationship:

[0037] In the formula, Representing the The dynamic cutoff frequency output at each moment, in radians per second; This represents the calibrated base cutoff frequency under steady-state, undisturbed conditions, and its unit is radians per second. The bandwidth extension sensitivity coefficient is expressed in radians per second. Representing the The wave disturbance characteristic value at each moment, in volts; This represents the rated phase voltage calibration constant of the propulsion motor, and its unit is volts. This represents the first ratio.

[0038] It is understandable that when a ship encounters large waves at high speed, the propeller experiences an extreme dynamic condition where it is partially out of the water and then plunges back in. The propeller load changes from extremely light to extremely heavy, causing high-frequency oscillations. At this point, the wave interference characteristic value... It will rise sharply, thus causing the first ratio term in the above relationship to... This will significantly increase, thereby driving the dynamic cutoff frequency. The frequency increases proportionally. When the back electromotive force caused by the waves undergoes a severe high-frequency distortion, the filter's bandwidth increases instantaneously, allowing these effective high-frequency signals representing the true flux linkage positions to pass through without loss.

[0039] Subsequently, stator voltage data is acquired using a voltage isolation sensor connected in parallel to the stator winding terminals. A static coordinate transformation, such as a Clarke transform, is then performed on the stator voltage data and stator multiphase current data acquired by the hardware to reduce their dimension and map them to a two-phase orthogonal static coordinate system, thereby extracting the mutually orthogonal coordinates. shaft and The stator voltage component and stator current component of the shaft.

[0040] Based on the stator voltage and stator current components of each axis, and combined with the internal calibration parameters of the motor, the original back electromotive force observation components in the stationary coordinate system are calculated, specifically satisfying the following relationship:

[0041] In the formula, Representing the At that moment The observed component of the original back electromotive force on the axis, in volts; Representing the At that moment The stator voltage component on the shaft, its unit is volt; This represents the stator resistance rating of the propulsion motor, and its unit is ohms. It can be obtained by directly reading the calibration data from the nameplate of the thrust motor. Representing the At that moment The stator current component on the shaft, its unit is ampere; Representing the At that moment Shaft stator current components, the unit of which is ampere; This represents the stator inductance rating of the propulsion motor, and its unit is Henry. This represents the basic sampling step size, measured in seconds. It is calculated in the same way... The original back electromotive force of the axis is observed.

[0042] Subsequently, the dynamic cutoff frequency is used as the core dynamic weight of the first-order discrete low-pass filter to filter the original back EMF observation components. Based on the filtered back EMF observation components, the initial estimated phase angle and the actual electric angular velocity of the motor, representing the position of the base flux linkage, are output through arctangent function and difference operation, specifically satisfying the following relationship:

[0043]

[0044]

[0045] In the formula, and Representing the first At that moment The original and filtered back electromotive force observation components on the axis, in volts. Representing the At that moment The filtered back electromotive force observation component on the axis, its unit is volts. Representing the At that moment The filtered back electromotive force observation component on the axis, its unit is volts. Representing the The dynamic cutoff frequency at any given moment, measured in radians per second. Represents the arctangent mathematical operation function; Representing the The initial estimated phase angle at each moment, in radians; Representing the The initial estimated phase angle at each moment, in radians; Representing the The actual electrical angular velocity of the motor at a given moment, expressed in radians per second. This represents the basic sampling step size, and its unit is seconds. It represents a zero-valued constant, and its unit is volt.

[0046] It is understandable that calculating the original back EMF observation component, filtering the original back EMF observation component, and calculating the initial estimated phase angle and the actual electric angular velocity of the motor based on the filtered back EMF observation component are all existing technologies, and will not be elaborated here.

[0047] It should be added that the above calculation of the filtered back EMF observation component involves the filtered back EMF observation component from the previous time step, therefore it is necessary to set the filtered back EMF observation component at the initial time step. Additionally, the above embodiments also involve bandwidth extension sensitivity coefficients. The method for obtaining the above parameters will be explained below.

[0048] The method for obtaining the filtered back electromotive force observation component at the initial moment includes: Because the rotor has not yet undergone mechanical rotational motion cutting the stator magnetic field lines during each cold start or when the propulsion motor is in a static standby state, no induced electromotive force is generated physically during the ship's propulsion control system's initialization phase. Therefore, during the controller's system initialization phase, the initial time... shaft and The filtered back electromotive force observation components of the axis are all preset to 0 volts.

[0049] Bandwidth extension sensitivity coefficient Methods for obtaining [the information] include: First, in the water tank towing wave-generating laboratory, a real propeller model was connected to a test motor equipped with a high-precision absolute encoder, and it was run in still water to lock the basic cutoff frequency calibration value.

[0050] Next, the wave generator is started to apply simulated wave impact loads with a specific period, and the peak data of the wave disturbance characteristic values ​​at this time are recorded simultaneously.

[0051] Then, under this wave impact condition, due to the high-frequency abrupt change in load, the observer using the base cutoff frequency will exhibit severe phase divergence between the estimated phase angle and the actual phase angle output by the absolute encoder. At this point, the cutoff frequency extension is manually increased step-by-step in the test host computer until the dynamic deviation between the phase angle estimated by the observer and the actual phase angle converges again and stabilizes within the preset physical safety tolerance band; the minimum cutoff frequency extension that just satisfies the convergence condition is recorded. For example, the upper limit of the physical safety tolerance band is 5 degrees, and the lower limit is -5 degrees.

[0052] Finally, the minimum cutoff frequency extension recorded above is divided by the ratio of the peak data of the wave interference characteristic value to the rated phase voltage calibration constant, and the resulting quotient is determined as the bandwidth extension sensitivity coefficient.

[0053] S3: Calculate the angular velocity correction coefficient mapped from the first ratio, and multiply the angular velocity correction coefficient by the actual electric angular velocity of the motor to obtain the phase compensation angle used to offset the hardware delay error.

[0054] It should be noted that the power switching devices inside the frequency converter have inherent hardware dead time, and the microprocessor communication also has inherent delay. From the microprocessor outputting the phase command to the final generation of the actual current, this physical link inevitably produces a fixed time delay. When the speed is stable, the angle deviation caused by this delay is minimal; however, under high-dynamic-change operating conditions, the rotor can rotate through a large electrical angle within a short time delay. If conventional fixed compensation methods are used, it is easy to cause a serious mismatch between the stator magnetic field and the actual rotor position, resulting in reactive excitation, torque oscillation, or even loss of synchronization. To address this, this solution introduces a phase lead compensation strategy that is nonlinearly adaptive to the impact intensity, dynamically offsetting the electrical angle lag error caused by the hardware dead time and transmission delay.

[0055] Preferably, as an example, the angular velocity correction coefficient mapped by the first ratio is calculated, and the angular velocity correction coefficient is multiplied by the actual electrical angular velocity of the motor to obtain the phase compensation angle used to compensate for hardware delay errors, including:

[0056] In the formula, Representing the The phase compensation angle at each moment, in radians; This represents the feedforward lead gain constant, which is dimensionless. Represents the natural constant; represent The wave disturbance characteristic value at each moment, in volts; The rated phase voltage calibration constant of the propulsion motor is represented by volts; Representing the The actual electric angular velocity of the motor at a given moment, expressed in radians per second. This represents the inherent hardware communication delay constant of the power module of the drive inverter, and its unit is seconds. This is the angular velocity correction factor. The effective transient electric angular velocity within the delay window.

[0057] It is understandable that when a ship encounters a sudden reverse swell that forces the propulsion motor speed down in a very short time, the wave disturbance characteristic value, representing the transient impact intensity, will be affected. It will rise sharply. At this point, the angular velocity correction coefficient in the equation... As the angular velocity increases, multiplying the angular velocity correction coefficient by the actual electrical angular velocity of the motor distorted by the forced impact at the current moment allows for accurate prediction of the rotor's true effective transient electrical angular velocity during the hardware delay period. By multiplying this effective transient electrical angular velocity by the inherent hardware communication delay constant, the electrical angle that the rotor should have rotated within the physical dead zone and delay window but was lost due to hardware lag can be accurately calculated and used as a feedforward leading phase compensation angle to compensate the system.

[0058] It should be added that the feedforward lead gain constant Obtained through the following calibration procedure: First, when the ship's propulsion motor is unloaded and running at its rated full speed, the moment when the microprocessor sends the gate drive signal and the moment when the corresponding phase's actual stator current crosses zero are captured using a high-frequency oscilloscope. The difference between the two moments is then used to obtain the actual physical delay time. Next, the physical delay time is multiplied by the electrical angular velocity at the rated full speed state to calculate the actual physical lag angle; Finally, the product of the electrical angular velocity at rated full speed and the inherent hardware communication delay constant is calculated as the theoretical nominal lag angle. The actual physical lag angle is divided by the theoretical nominal lag angle, and the resulting ratio is determined as the feedforward lead gain constant.

[0059] S4: The initial estimated phase angle is added to the phase compensation angle to obtain the final rotor electrical angle that aligns with the true magnetic field of the rotor. The damping coefficient constructed based on the first ratio is multiplied by the original torque command to obtain the final control torque with flexible buffering capability. The final rotor electrical angle and the final control torque are converted into drive signals and applied to the power switching device to drive the motor.

[0060] It should be noted that traditional ship speed control systems typically use rigid torque output commands on the outer loop, with the core objective of strictly maintaining the set speed. If a large wave jams the propeller during navigation, causing a sudden drop in speed, the conventional drive system will still forcibly output maximum electromagnetic torque to counteract wave resistance. This continuous high-load, rigid resistance condition easily leads to significant stress concentration on slender transmission shafts, potentially inducing irreversible torsional resonance or even shaft breakage. Therefore, this solution introduces a closed-loop flexible buffering mechanism at the end, dynamically softening the torque command and translating it into smooth physical control actions, thus mitigating mechanical damage caused by extreme impacts at the source.

[0061] Preferably, as an example, the initial estimated phase angle is added to the phase compensation angle to obtain the final rotor electrical angle aligned with the true magnetic field of the rotor; the damping coefficient constructed based on the first ratio is multiplied by the original torque command to obtain the final control torque with flexible buffering capability; the final rotor electrical angle and the final control torque are converted into drive signals and applied to power switching devices to drive the motor, including: First, determine the final rotor electrical angle and the final control torque:

[0062]

[0063] In the formula, Representing the The final rotor electrical angle at each moment, in radians; Representing the The initial estimated phase angle at each moment, in radians; Representing the The phase compensation angle at each moment, in radians; Representing the The final control torque that provides flexible buffering capability at any given moment, measured in Newton-meters; Representing the The raw torque command issued by the energy management system at any given moment, measured in Newton-meters; Represents the natural constant; This represents the flexible unloading damping coefficient, which has a dimensionless unit. Representing the The wave disturbance characteristic value at each moment, in volts; This represents the rated phase voltage calibration constant, and its unit is volts. is the damping coefficient.

[0064] It is understandable that when a ship encounters abnormally large waves that cause the propeller to become stuck due to foreign objects or ice, resulting in a momentary drop in rotational speed, the wave disturbance characteristic value, which represents the transient impact intensity of the waves, is affected. A sharp increase will occur, reaching a peak value. At this point, the nonlinear damping coefficient in the relationship will... It shrinks rapidly and proportionally, approaching zero; simultaneously, the original torque command at the front end of the relation... Originally attempting to output maximum rigid electromagnetic torque to overcome wave resistance, it was working in conjunction with a nonlinear damping coefficient. Multiplication can achieve the following: the more severe the external mechanical obstruction and the more violent the wave impact, the greater the softening weight released by the nonlinear damping coefficient, enabling the control system to actively and decisively suppress the originally dangerous rigid full torque command, and instead output an extremely soft, safe, pressure-holding, and exploratory torque to the actuator.

[0065] It should be added that the flexible unloading damping coefficient Obtained through the following calibration procedure: First, high-frequency torque strain gauges are installed at both ends of the ship's long shaft system, and a sweep frequency excitation torque is applied to the long shaft system through a torsional vibration exciter. Next, the torque transmission amplitude-frequency response curves at both ends of the shaft system are recorded and plotted. The resonance peak point of the curve is extracted to determine the first-order torsional resonance frequency of the ship's long shaft system. The corresponding stiffness attenuation damping ratio is calculated using the half-power bandwidth method. Finally, the maximum safe yield limit torque in the mechanical manual of the long shaft system is obtained, and the maximum safe yield limit torque is divided by the product of the stiffness attenuation damping ratio and the rated output torque of the propulsion motor. The resulting ratio is determined as the flexible unloading damping coefficient.

[0066] Next, the final rotor electrical angle and the final control torque are converted into drive signals, which are applied to the power switching devices to drive the motor, specifically including: The final control torque is used to generate a quadrature axis current reference command, and the final rotor electrical angle is used as a magnetic field orientation reference to perform an inverse coordinate transformation to generate a stator voltage vector command. The stator voltage vector command is sent to the space vector pulse width modulation module to be converted into a PWM gate trigger pulse sequence as a drive signal closed loop to act on the IGBT power switching device inside the ship propulsion frequency converter.

[0067] Furthermore, considering the extremely special scenario where a vessel runs aground in shallow water and needs to forcibly extricate itself at full power regardless of mechanical losses, this embodiment incorporates an extrication takeover procedure in parallel within the control loop. When the system detects that the vessel is in a state of forced extrication in shallow water, it triggers a bypass switch, directly using the original torque command issued by the energy management system as the final control torque. Following the aforementioned execution logic of inverse coordinate transformation and space vector pulse width modulation, this torque is converted into the gate trigger pulse sequence to forcibly drive the motor. Thus, through the above operations, the control logic achieves full-condition coverage of both flexible protection in harsh sea conditions and rigid extrication from shallow water grounding, successfully completing the spatiotemporal dynamic closed loop of the system's core parameters.

[0068] To demonstrate the effectiveness of the solution, relevant experiments were conducted. Below are the images obtained from the experiments: Figure 2 The image shows the dynamic extraction effect of wave interference feature values, with the solid line representing the trajectory of wave interference feature value changes. The image reveals that within the interval from the 2nd to the 3rd second on the horizontal axis, the ship encountered powerful impacts from severe breaking waves, and the solid line exhibits a significant and smooth peak in this interval. In contrast, the noise level remains extremely low in the stable navigation area without impact. This clearly demonstrates that the differential feature extraction analysis of this invention can accurately and sensitively capture the transient impact energy caused by external ocean waves.

[0069] Figure 3 The graph shows the comparison of the effects of dynamic cutoff frequency adaptation. The dashed line represents the constant base cutoff frequency calibration value used in traditional fixed-bandwidth observation methods; the solid line represents the dynamic cutoff frequency trajectory proposed in this invention. The graph shows that in the transient range of encountering wave impact, the dashed line representing the traditional method remains a rigid horizontal straight line, reflecting the severe inadequacy of the traditional method to handle high-frequency physical disturbances. The solid line, however, expands proportionally with the increase in wave impact intensity. This strongly demonstrates that this invention can dynamically open high-frequency signal channels using eigenvalues, effectively avoiding the defect of the effective component of the high-frequency true back electromotive force being blindly filtered out by a fixed low-pass filter, ensuring undistorted observations under harsh sea conditions from the signal source.

[0070] Figure 4The graph shows a comparison of the effects of feedforward compensation on suppressing hysteresis phase error. The dashed line represents the actual physical hysteresis angle caused by hardware dead zones and communication delays under traditional rigid delay accumulation; the solid line represents the final phase deviation output by this invention. The graph shows that in a stable navigation area without wave impact, the two curves coincide. When faced with a sudden change in operating conditions caused by a sudden wave in the middle of the horizontal axis, the dashed line representing the traditional method shows an extremely sharp error spike, indicating that the injected current has seriously deviated from the actual rotor position. In contrast, the solid line, throughout the entire wave impact cycle, not only does not diverge but is steadily suppressed and maintained near its original value. This directly verifies that the nonlinear exponential feedforward mechanism of this invention can accurately predict and dynamically compensate for the additional hysteresis angle amplified by sudden changes in operating conditions, eliminating the hidden dangers of dynamic reactive excitation and loss of synchronization.

[0071] Figure 5 The graph shows a comparison of the protective effects of flexible torque unloading. The dashed line represents the original torque command blindly output under traditional rigid outer loop control; the solid line represents the final control torque actually issued to the power device after the introduction of the exponential damping attenuation mechanism in this invention. The graph shows that when encountering strong wave impact, the dashed line representing the traditional method rises sharply, attempting to forcibly resist the waves with an extremely dangerous peak rigid torque. In contrast, the solid line representing this invention not only does not rise during this period but also shows a significant downward collapse, smoothly recovering to the cruising torque after the impact energy dissipates. This strongly demonstrates that this invention possesses an intelligent closed-loop flexible buffering and yielding mechanism, protecting the vulnerable mechanical shaft system by actively applying downward electromagnetic torque.

[0072] This concludes the embodiment.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A ship propulsion motor control method based on back electromotive force estimation, characterized in that, include: The process involves collecting stator multiphase current data and raw torque commands. The rate of change of the quadrature-axis current is obtained based on the basic sampling step size and the difference between instantaneous values ​​of the quadrature-axis current at adjacent sampling times. The rate of change of the raw torque command is also obtained based on the basic sampling step size and the difference between raw torque commands at adjacent sampling times. A first product is obtained by weighting the rate of change of the quadrature-axis current using the stator quadrature-axis inductance calibration value of the propulsion motor. A second product is obtained by weighting the rate of change of the raw torque command using the equivalent conversion constant from torque to voltage. The absolute value of the difference between the first product and the second product is calculated, and a first-order inertial weighted sum is performed on the absolute value of the difference and the wave interference characteristic value at the previous moment using the basic low-pass filter attenuation constant to determine the wave interference characteristic value at the current moment. Stator voltage data is collected, and a first ratio of the wave interference characteristic value to the rated phase voltage calibration constant is calculated. The first ratio is used to broaden the basic cutoff frequency calibration value to obtain a dynamic cutoff frequency for adapting to high-frequency disturbances. Combined with observations of the dynamic cutoff frequency, an initial estimated phase angle characterizing the position of the basic flux linkage and the actual electric angular velocity of the motor are obtained. Calculate the angular velocity correction coefficient mapped by the first ratio, multiply the angular velocity correction coefficient by the actual electrical angular velocity of the motor to obtain the phase compensation angle used to offset hardware delay errors; add the initial estimated phase angle to the phase compensation angle to obtain the final rotor electrical angle aligned with the true magnetic field of the rotor; multiply the flexible unloading damping coefficient by the first ratio at the current moment to obtain the product value; and determine the negative of the product value as the attenuation index. The damping coefficient is calculated using the natural constant as the base and the attenuation index as the exponent. The damping coefficient is multiplied by the original torque command to obtain the final control torque with flexible buffering capability. The final rotor electrical angle and the final control torque are converted into drive signals and applied to the power switching device to drive the motor.

2. The ship propulsion motor control method based on back electromotive force estimation according to claim 1, characterized in that, The specific calculation process for obtaining the dynamic cutoff frequency to adapt to high-frequency disturbances by widening the basic cutoff frequency calibration value using the first ratio is as follows: multiply the bandwidth extension sensitivity coefficient by the first ratio at the current time to obtain the bandwidth extension amount; add the basic cutoff frequency calibration value to the bandwidth extension amount to determine the dynamic cutoff frequency at the current time.

3. The ship propulsion motor control method based on back electromotive force estimation according to claim 2, characterized in that, The method for obtaining the bandwidth extension sensitivity coefficient includes: applying simulated wave impact load to a propeller model under wave-generating conditions in a water tank, and simultaneously recording the peak data of wave interference characteristic values; artificially increasing the cutoff frequency extension to make the dynamic deviation between the phase angle estimated by the observer and the actual phase angle converge and stabilize within a preset physical safety allowable error band, and recording the minimum cutoff frequency extension that just meets the convergence condition; dividing the minimum cutoff frequency extension by the ratio of the peak data of the wave interference characteristic value to the rated phase voltage calibration constant, and determining the quotient as the bandwidth extension sensitivity coefficient.

4. The ship propulsion motor control method based on back electromotive force estimation according to claim 1, characterized in that, The specific calculation process for obtaining the phase compensation angle used to offset hardware delay errors by multiplying the angular velocity correction coefficient mapped from the first ratio by the actual electrical angular velocity of the motor is as follows: A power function value is calculated using the natural constant as the base and the opposite of the first ratio at the current moment as the exponent; the difference between the value 1 and the power function value is multiplied by the feedforward lead gain constant to determine the angular velocity correction coefficient at the current moment; the angular velocity correction coefficient is multiplied by the actual electrical angular velocity of the motor at the current moment to obtain the effective transient electrical angular velocity within the delay window; the effective transient electrical angular velocity is multiplied by the inherent hardware communication delay constant to determine the phase compensation angle at the current moment.

5. The ship propulsion motor control method based on back electromotive force estimation according to claim 4, characterized in that, The method for obtaining the feedforward lead gain constant includes: capturing the time difference between the moment when the microprocessor sends a drive signal and the moment when the actual stator current crosses zero, when the ship's propulsion motor is unloaded and operating at its rated full speed; multiplying the time difference by the electrical angular velocity at the rated full speed to calculate the actual physical lag angle; calculating the product of the electrical angular velocity at the rated full speed and the inherent hardware communication delay constant as the theoretical nominal lag angle; dividing the actual physical lag angle by the theoretical nominal lag angle; and determining the ratio obtained as the feedforward lead gain constant.

6. The ship propulsion motor control method based on back electromotive force estimation according to claim 1, characterized in that, The specific calculation process for multiplying the damping coefficient by the original torque command to obtain the final control torque with flexible buffering capability is as follows: multiply the original torque command at the current moment by the damping coefficient to determine the final control torque at the current moment.

7. The ship propulsion motor control method based on back electromotive force estimation according to claim 6, characterized in that, The method for obtaining the flexible unloading damping coefficient includes: applying a sweep frequency excitation torque to both ends of the ship's long shaft system to extract the stiffness attenuation damping ratio of the ship's long shaft system; obtaining the maximum safe yield limit torque of the ship's long shaft system; dividing the maximum safe yield limit torque by the product of the stiffness attenuation damping ratio and the rated output torque of the propulsion motor; and determining the obtained ratio as the flexible unloading damping coefficient.

8. The ship propulsion motor control method based on back electromotive force estimation according to claim 1, characterized in that, The specific operation of converting the final rotor electrical angle and the final control torque into a drive signal includes: generating a quadrature-axis current reference command using the final control torque, and performing an inverse coordinate transformation using the final rotor electrical angle as a magnetic field orientation reference to generate a stator voltage vector command; sending the stator voltage vector command into a space vector pulse width modulation module to convert it into a gate trigger pulse sequence as the drive signal; and when the ship is identified as being in a shallow water forcibly extricated state, triggering a bypass switch to directly convert the original torque command into the gate trigger pulse sequence as the final control torque.

9. A ship propulsion motor control method based on back electromotive force estimation according to claim 1, characterized in that, The process of obtaining the initial estimated phase angle and actual electric angular velocity of the motor by combining the observation of the dynamic cutoff frequency includes: calculating the original back EMF observation components of the stator voltage data and stator multiphase current data in the stationary coordinate system based on the motor calibration parameters; dynamically filtering the original back EMF observation components based on the dynamic cutoff frequency to obtain the filtered back EMF components; introducing the anti-zero constant into the filtered back EMF components for trigonometric function calculation to obtain the initial estimated phase angle; and calculating the actual electric angular velocity of the motor based on the time change rate of the initial estimated phase angle.

Images