Method for regulating and controlling stability amplitude of ultrasonic power supply

Abstract

According to the method for regulating and controlling the stability amplitude through the ultrasonic power source, the output of the ultrasonic power source is dynamically adjusted through intelligent control and a feedback mechanism, the amplitude of the transducer is stable under different loads, and the machining precision and efficiency are improved. An equivalent model is constructed through a power supply output characteristic equation and a dynamic response relationship, voltage, current and mechanical information are monitored in real time by combining a high-precision sensor, expected power output is predicted by applying a neural network after deep preprocessing, a power error is calculated by using a PID algorithm, and power supply output is dynamically adjusted. According to the method, through continuous feedback and correction, it is ensured that the amplitude of the system is reasonable and stable under the complex working condition, the stability and applicability of the ultrasonic transducer system are remarkably enhanced, and the method is particularly suitable for ultrasonic machining equipment in the complex load environment.

Description

Technical Field

[0001] The present invention belongs to the field of ultrasonic machining, and in particular relates to a method for regulating the stability amplitude of an ultrasonic power supply. Background Art

[0002] With the rapid development of precision manufacturing toward miniaturization and sophistication, high-end equipment such as micro-electromechanical systems, medical implants, and aerospace components are placing extremely high demands on part machining accuracy. Ultrasonic machining, as an advanced machining method, leverages its unique machining mechanism generated by high-frequency vibration to reduce cutting forces and thermal impact while achieving localized embrittlement of materials and inducing microcrack propagation. This demonstrates significant advantages in machining special materials such as hard and brittle composites. Ultrasonic machining not only helps improve the quality of machined surfaces but also achieves high geometric accuracy in the machining of minute structures, making it a highly sought-after technology in the field of ultra-precision manufacturing.

[0003] However, the core of ultrasonic machining lies in maintaining a stable tool head vibration amplitude, and this vibration stability is directly dependent on the power output stability of the ultrasonic power supply. Currently, traditional ultrasonic power supply systems generally use fixed-frequency resonance and open-loop power control strategies. This can lead to a mismatch between output power and load impedance when faced with dynamic operating conditions such as uneven material hardness, tool wear, and heat accumulation. Over time, this mismatch has led to a gradual attenuation of tool amplitude, which in turn has led to a series of problems such as increased surface roughness and microstructure dimensional tolerances, limiting the stability and consistency of ultrasonic machining technology during long-term continuous machining and high-load machining.

[0004] To address this bottleneck, current technological improvements are primarily focused on improving system response speed and energy efficiency. For example, digital phase-locked loop (DPL) technology and wide-bandgap semiconductor devices are introduced to improve power output and load matching. However, these improvements still face a common challenge: insufficient modeling and control of the complex dynamic coupling characteristics between power supply and load. The lack of real-time and precise energy output control makes it difficult to maintain the ideal vibration amplitude of the tool head under complex working conditions, thus affecting overall machining quality. Summary of the Invention

[0005] To address these technical challenges, a new method for regulating an ultrasonic power supply to achieve stable amplitude output has been developed. This method establishes an equivalent model of the power supply and its operating mechanism, incorporating control laws tailored to practical challenges. This method ensures the power supply can consistently and stably output ultrasonic waves of the desired power. Furthermore, the introduction of a PID algorithm avoids the problem of excessive cumulative model inputs. This method significantly improves the stability of ultrasonic machining, significantly enhancing machining efficiency and part accuracy, effectively ensuring part stability and reliability in daily production.

[0006] Through an intelligent control and feedback mechanism, the present invention dynamically adjusts the output of the ultrasonic power supply to ensure stable amplitude. Thus, under different load conditions, the power supply can still maintain a stable output, providing a new method for regulating the stable amplitude of the ultrasonic power supply, and improving the quality and efficiency of ultrasonic processing. Through real-time monitoring, data processing, and adaptive adjustment, the present invention realizes the stability of amplitude and the reasonable control of power supply power.

[0007] The implementation steps are as follows: Step 1: Establish an equivalent model based on the output characteristics and dynamic response of the ultrasonic power supply Step 2: Deploy high-precision current, voltage, and strain gauge sensors to realize real-time digital characterization of the power supply output parameters Step 3: Use a filtering and noise reduction algorithm to preprocess the collected data and extract features Step 4: Construct a feedforward neural network model to intelligently predict the expected output power of the ultrasonic power supply Step 5: Calculate the deviation between the actual power supply output power and the expected power Step 6: Use the PID control algorithm to dynamically regulate the output power of the ultrasonic power supply Step 7: Real-time correct the output power of the ultrasonic power supply to achieve the expected amplitude Furthermore, based on the output characteristics and dynamic response of the ultrasonic power supply, the present invention establishes an equivalent circuit model between the ultrasonic power supply and the transducer. By analyzing the load characteristics of the transducer and the electro-acoustic conversion process, a series-parallel resonance circuit model is adopted to characterize the relationship between the power supply and the transducer. This model reflects how the transducer efficiently converts the input electrical energy into mechanical vibration energy and optimizes the energy transfer efficiency through the resonance effect. This step provides a theoretical basis and model support for subsequent intelligent control and dynamic amplitude adjustment.

[0008] Furthermore, to ensure the stable amplitude of the transducer, the present invention deploys a variety of sensors in the system, including voltage sensors, current sensors, and laser displacement sensors. Through the above sensors, the output voltage, output current of the ultrasonic power supply, and the amplitude state of the vibration generated by the transducer are collected in real time. The voltage and current sensors have high-frequency response capabilities and can accurately capture the instantaneous changes at ultrasonic frequencies. The laser displacement sensor uses non-contact high-precision measurement technology to accurately capture the vibration displacement information of the transducer and obtain the accurate amplitude data of the transducer.

[0009] Furthermore, the data collected by the above current, voltage, and amplitude sensors are preprocessed before being transmitted to the central control system. In the present invention, a high-speed data acquisition system (DAQ) is used to synchronously collect amplitude, voltage, and current data, and the sampling frequency is higher than 10 times the ultrasonic working frequency. To ensure the accuracy of the data, signal processing algorithms are used to filter and denoise the collected data, removing high-frequency interference and abnormal data. The processed data is converted into digital signals through an analog-to-digital converter (ADC) and transmitted to the central control unit, providing reliable input data for the subsequent control system.

[0010] Furthermore, through a multi-layer feedforward neural network (MLP), the present invention predicts the expected power of the system based on the data such as voltage, current, and amplitude collected in real time. The neural network model processes complex input data relationships through non-linear mappings of the input layer, hidden layer, and output layer. The model is trained with a large amount of historical data, using the mean square error as the loss function, and adopting the backpropagation algorithm to optimize the network weights and bias terms, finally generating the expected power output under the current load conditions of the system, providing a basis for subsequent power supply regulation.

[0011] Furthermore, the system measures the actual output power of the power supply in real time through sensors and compares it with the expected power predicted by the neural network to calculate the power error. The power error reflects the gap between the current actual output power and the expected output power of the ultrasonic power supply. The calculation formula is the error value . This error value is fed back to the control system as an important reference for subsequent adjustment of the power supply output. By calculating the power error, the system can accurately understand the current power deviation and make timely adjustments.

[0012] Furthermore, based on the calculated power error, the present invention adjusts the power supply output voltage through a PID control algorithm. The PID controller combines the proportional (P), integral (I), and derivative (D) terms to generate different control signals according to the magnitude and change trend of the error, and adjusts the power supply output in real time, making its power gradually approach the expected value predicted by the neural network. The PID control system can quickly respond to power deviations and dynamically adjust to ensure that the amplitude of the ultrasonic transducer remains stable under different working conditions.

[0013] Furthermore, the present invention continuously monitors the amplitude and power, and combines the PID control feedback to real-time correct the power supply power output to ensure that the transducer maintains a reasonable amplitude. When the system detects that the amplitude deviates from the set range, it automatically adjusts the power supply power. Through the feedback mechanism and fast response ability, the system can real-time correct the output when the load changes and the working conditions fluctuate, ensuring that the ultrasonic transducer is always in a stable working state.

[0014] Figure 1 Steps for realizing the stable amplitude of the ultrasonic power supply regulation; Figure 2 It is a flowchart for realizing the stable amplitude regulation of the ultrasonic power supply; Figure 3 It is a structural diagram of a power prediction neural network for the power supply to achieve stable amplitude output.

[0015] The present invention will be described in detail below in conjunction with the accompanying drawings and by way of examples. Specific embodiments

[0016] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, not all of them. All other embodiments based on this invention, which are completed by those of ordinary skill in the art without creative efforts, are within the protection scope of this invention.

[0017] Figure 1 The implementation steps for the ultrasonic power supply to regulate the stable amplitude are as follows. The specific steps include: establishing an equivalent model based on the output characteristics and dynamic response of the ultrasonic power supply; deploying high-precision current, voltage, and strain gauge sensors to realize real-time digital characterization of the power supply output parameters; using filtering and noise reduction algorithms to preprocess the collected data and extract features; constructing a feedforward neural network model to intelligently predict the expected output power of the ultrasonic power supply; calculating the deviation between the actual power supply output power and the expected power; using a PID control algorithm to dynamically regulate the output power of the ultrasonic power supply; and real-time correcting the output power of the ultrasonic power supply to achieve the expected amplitude.

[0018] Figure 2 It is the implementation process of the ultrasonic power supply to regulate the stable amplitude. First, key parameters such as the current, voltage, and tool head cutting force of the resonant circuit are collected in real time by sensors. After being converted into digital signals by a high-speed data acquisition system, median filtering and Kalman filtering algorithms are used to eliminate noise interference and improve data accuracy. Subsequently, the central control system calculates the actual output power based on the filtered signal and compares it with the preset expected power value. At the same time, a multi-layer feedforward neural network models and predicts the system behavior, assisting in adjusting the driving frequency and voltage amplitude of the resonant circuit to achieve precise regulation. Finally, through a closed-loop feedback mechanism, the output state is continuously monitored, and a PID controller dynamically calculates the deviation value. Combining proportional, differential, and other adjustment strategies to optimize the response characteristics, and real-time correcting the control parameters to ensure that the amplitude remains stable under load fluctuations or external disturbances. This process integrates sensor technology, intelligent algorithms, and adaptive control to ensure that the ultrasonic power supply output achieves a stable amplitude.

[0019] Figure 3It represents the structural design of a multi-layer feedforward neural network. The inputs are parameters such as voltage, current, and cutting force collected by sensors. After calculation and prediction by the hidden layer of the neural network, the output is the output power of the power supply when a stable amplitude is obtained.

[0020] Step 1: Establish an equivalent model based on the output characteristics and dynamic response of the ultrasonic power supply Through the analysis of the output characteristics of the ultrasonic power supply, the load characteristics of the transducer, and the dynamic response, an equivalent circuit model between the ultrasonic power supply and the transducer is established. This model is used to characterize the working state of the ultrasonic system under different working conditions and provides a basis for subsequent intelligent control. In the ultrasonic transducer system, a series-parallel resonance circuit model is used to characterize the relationship between the power supply and the transducer, describing how the transducer efficiently converts the input electrical energy into ultrasonic vibration energy and maintains or adjusts its output performance under different operating environments or working conditions.

[0021] Among them, the series inductance (L) represents the equivalent inductance of the mechanical vibration system of the transducer and reflects the energy storage characteristics of the transducer at ultrasonic frequencies. The parallel capacitance (C) is used to characterize the capacitance characteristics of the mechanical elastic part of the transducer or the medium and represents the charge storage capacity inside the transducer. The series resistance (R) represents the mechanical losses of the transducer and the energy losses during the electro-acoustic conversion process. This resistance reflects the efficiency problem of the ultrasonic transducer. A larger resistance means more energy losses.

[0022] Inductance-capacitance equation:

[0023] Among them, is the inductance, is the inductance current, is the input voltage, is the output voltage, is the capacitance, is the load resistance.

[0024] According to the relevant content of high-frequency electronic circuits, at ultrasonic frequencies, the transducer can be equivalent to a resonant circuit. The ultrasonic transducer converts electrical energy into mechanical vibration and maximizes the energy carried by the mechanical vibration with the help of the resonance effect. The resonance frequency is one of the most important characteristics of the transducer system. From the above content, it can be deduced that:

[0025] Among them, is the resonance frequency of the transducer. At the resonance frequency, the energy conversion efficiency of the transducer is the highest, and the output energy of the power supply can be transferred to the load with the least loss.

[0026] Step 2: Deploy high-precision current, voltage, and strain gauge sensors to realize real-time digital characterization of the power supply output parameters In order to ensure that the amplitude of the ultrasonic transducer can work stably under different working conditions, the present invention deploys a variety of sensors in the transducer system to monitor and collect the electrical parameters output by the ultrasonic power supply and the amplitude of the transducer in real time. A high-precision voltage sensor is selected and installed at the output end of the ultrasonic power supply to measure the output voltage of the power supply in real time. It has sufficient frequency response ability to accurately capture the voltage fluctuations at ultrasonic frequencies. The current sensor is installed in series in the output circuit of the ultrasonic power supply to monitor the output current of the ultrasonic power supply. It has a high-frequency response to ensure that the instantaneous current changes at ultrasonic frequencies can be captured. A laser displacement sensor is used to measure the displacement of the transducer vibration in a non-contact manner, and then the amplitude is calculated.

[0027] Step 3: Preprocess the collected data using a filtering and noise reduction algorithm and extract features The present invention uses a high-speed data acquisition system (DAQ) to collect the signals of each sensor. The sampling frequency of the data acquisition system should be higher than 10 times the ultrasonic working frequency to ensure that the complete amplitude, voltage, and current waveforms are captured. The collected amplitude data and power supply output parameters may contain noise and fluctuations. Therefore, signal processing algorithms are used to preprocess the data to remove high-frequency noise and abnormal data, ensuring that the input to the subsequent intelligent control module is accurate and reliable data. The analog signals output by the sensors deployed in Step 2 are converted into digital signals through an analog-to-digital converter (ADC), and the data is transmitted to the central controller through a high-speed data bus. The data acquisition system needs to have multi-channel input functions to ensure that the voltage, current, and amplitude signals can be synchronously collected and transmitted.

[0028] Step 4: Construct a feedforward neural network model to intelligently predict the expected output power of the ultrasonic power supply Using a neural network model to predict the expected power is one of the cores of the present invention. Its purpose is to predict the expected power of the ultrasonic transducer system through intelligent algorithms to ensure that the output power of the ultrasonic power supply can be automatically adjusted under different working conditions and loads to maintain the amplitude stability of the system.

[0029] The present invention uses a multi-layer feedforward neural network (MLP) to predict the expected power. This neural network model performs a non-linear mapping on the input data through the input layer, hidden layer, and output layer structures, and can well handle the complex non-linear characteristics in the ultrasonic power supply system. The input layer receives the voltage , current , amplitude and other electrical parameters collected and preprocessed by the sensors; the hidden layer contains several neurons, and non-linear activation functions are used to improve the expression ability of the model; the output layer is used to output the expected power under the current system , for subsequent control and adjustment. The neural network model is trained through a supervised learning method, and the training set consists of a large amount of historical data, including the input data (voltage, current, amplitude, etc.) of the ultrasonic power supply system under different working conditions and the corresponding actual output power. Through the backpropagation algorithm, the model continuously adjusts the weights and bias terms to minimize the error between the predicted power and the actual power.

[0030] During model training, the mean squared error is used as the loss function, and by minimizing the loss function, the model can gradually learn the non-linear relationship between the input data and the power output.

[0031] It is defined as follows:

[0032] Among them, is the expected power predicted by the neural network, is the actual power, is the number of samples, represents the parameters such as the weights and biases of the neural network, and represent the new and old weights respectively, is the learning rate, is the gradient of the loss function with respect to the weights.

[0033] The trained neural network model is used for real-time prediction. First, the voltage , current , amplitude and other data collected by the sensor in real time are input into the neural network. The input data needs to be normalized to ensure the consistency of the data range and improve the prediction accuracy; further, the neural network calculates the output layer by layer through the forward propagation process; further, after the non-linear mapping of the hidden layer, the neural network finally generates the expected power in the output layer for power regulation of the power supply. The activation values of the neurons in each layer can be obtained by the following formula:

[0034] Among them, is the weight matrix of the th layer, is the bias vector, is the activation function, is the th layer's activation value.

[0035] Step Five: Calculate the deviation between the actual power output of the power supply and the expected power The present invention calculates the error value between the expected power predicted by the neural network and the actual output power, providing a reference for subsequent power output regulation of the power supply. The system measures the actual output power of the power supply in real time through sensors, and the actual power calculation formula is:

[0036] Where and are the power supply output voltage and current data collected by the voltage and current sensors deployed in step two, respectively.

[0037] The expected power of the power supply is predicted by the neural network model in step four, and the actual power is measured by the sensor. The error value calculation formula is:

[0038] This error value reflects the deviation between the current power output of the system and the ideal value, and will be fed back to the control system as the basis for the subsequent PID controller to adjust the power output of the power supply. The system continuously monitors the change of the error and dynamically adjusts the output power to reduce the error and ensure the stability of the amplitude. By calculating the power error, the present invention can understand the power deviation in real time and provide data support for accurately regulating the power output of the power supply.

[0039] Step six: Dynamically regulate the output power of the ultrasonic power supply using the PID control algorithm In the present invention, based on the power error value calculated in step five, the PID control algorithm is used to adjust the output signal of the power supply to ensure the stability of the amplitude of the ultrasonic transducer. The power error value calculated in step five is introduced into the PID controller. The PID controller returns a control signal for adjusting the power supply output according to the size and change trend of . The PID controller adaptively regulates and returns through three control coefficients: proportional ( ), integral ( ), and derivative ( ):

[0040] Among them, P (Proportion) represents adjusting the voltage output according to the current error value to quickly respond to the system deviation. I (Integral) represents accumulating past errors to correct long-existing deviations. D (Derivative) represents predicting future trends based on the error change rate and making adjustments in advance to avoid overshoot. The PID controller adjusts the power output signal of the power supply in real time according to the calculated control signal, making the actual power gradually tend to the expected power predicted by the neural network, thereby reducing the power error and stabilizing the amplitude of the transducer. The control system continuously collects new data and adjusts the output signal of the PID controller in real time according to the new error value to ensure that the system can still maintain stable power and amplitude under dynamic working conditions. Through the PID control algorithm, the present invention can dynamically adjust the output voltage of the power supply, achieve stable amplitude and precise power control, and ensure the efficient operation of the system.

[0041] Step 7: Real-time correct the output power of the ultrasonic power supply to achieve the expected amplitude The present invention ensures that the ultrasonic transducer maintains a reasonable amplitude under different working conditions through continuous feedback and real-time power correction. In each control cycle, the power data and amplitude data collected by the sensor in real time are continuously fed back to the central control system. By monitoring the amplitude change in real time, the system can determine whether the current output meets the design requirements. If it is detected that the amplitude deviates from the set range, the system automatically adjusts the power output of the power supply according to the power error and the amplitude change trend. The power correction is driven by the PID controller, combined with the output signal in Step 6, to further optimize the power output of the power supply and ensure that the amplitude tends to be stable. Through the fast response mechanism, the system can real-time correct the power output of the power supply when the load changes, the environment fluctuates or the working conditions suddenly change, ensuring that the amplitude of the ultrasonic transducer always remains within a reasonable range and avoiding the situation of too large or too small amplitude. When the system operates for a long time or the working conditions change greatly, the control system will automatically adjust the PID parameters or retrain the neural network (Step 4) to enhance the adaptive ability and ensure that the amplitude stability is further optimized.

[0042] Through this real-time correction mechanism, the present invention can maintain a reasonable amplitude under complex working conditions and ensure the processing effect of the ultrasonic transducer and the system stability.

[0043] The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, in terms of specific implementation manners and application scopes, for those of ordinary skill in the art, there will also be changes according to the concept of the present invention. Generally speaking, the content of this description cannot be understood as a limitation of this invention.

Claims

1. A method for regulating an ultrasonic power supply to obtain a stable amplitude output, characterized in that Including the following steps: Establish an equivalent model based on the output characteristics and dynamic response of the ultrasonic power supply; Deploy high-precision current, voltage, and strain gauge sensors to realize real-time digital characterization of the power supply output parameters; Use filtering and noise reduction algorithms to preprocess the collected data and extract features; Construct a feedforward neural network model to intelligently predict the expected output power of the ultrasonic power supply; Calculate the deviation between the actual power supply output power and the expected power; Use the PID control algorithm to dynamically regulate the output power of the ultrasonic power supply; Real-time correct the output power of the ultrasonic power supply to achieve the expected amplitude.

2. The method according to claim 1, characterized in that, In the step of establishing an equivalent model based on the output characteristics and dynamic response of the ultrasonic power supply, a series-parallel resonance circuit model is used to characterize the relationship between the power supply and the transducer. This model reflects that the transducer efficiently converts the input electrical energy into mechanical vibration energy and optimizes the energy transfer efficiency through the resonance effect.

3. The method according to claim 1, wherein In the step of deploying high-precision current, voltage, and strain gauge sensors to realize real-time digital characterization of the power supply output parameters, the selected voltage sensor is installed at the output end of the ultrasonic power supply to measure the output voltage of the power supply in real time; the current sensor is serially installed in the output circuit of the ultrasonic power supply to monitor the output current of the ultrasonic power supply; a laser displacement sensor is used to measure the displacement of the transducer vibration in a non-contact manner, and then the amplitude is calculated.

4. The method according to claim 1, wherein In the step of constructing a feedforward neural network model to intelligently predict the expected output power of the ultrasonic power supply, a multi-layer feedforward neural network (MLP) is used to predict the expected power. The input layer receives electrical parameters such as voltage, current, and amplitude collected and preprocessed by the sensors; the hidden layer contains several neurons, and a non-linear activation function is used to improve the expression ability of the model; the output layer is used to output the expected power under the current system for subsequent control and adjustment. The neural network model is trained by a supervised learning method, uses the mean square error as the loss function, and uses the backpropagation algorithm to optimize the network weights and bias terms.

5. The method according to claim 1, wherein In the step of using the PID control algorithm to dynamically regulate the output power of the ultrasonic power supply, the PID controller combines the proportional (P), integral (I), and derivative (D) terms to generate different control signals according to the error size and change trend, and adjusts the power supply output in real time to make its power gradually approach the expected value predicted by the neural network, ensuring that the amplitude of the ultrasonic transducer remains stable under different working conditions.

6. The method according to claim 1, characterized in that, In the step of real-time correcting the output power of the ultrasonic power supply to achieve the expected amplitude, by continuously monitoring the amplitude and power, and combining the PID control feedback, the power supply power output is corrected in real time to ensure that the transducer maintains a reasonable amplitude. When the system detects that the amplitude deviates from the set range, the power supply power is automatically adjusted. Through the feedback mechanism and fast response ability, the system can correct the output in real time when the load changes and the working conditions fluctuate, ensuring that the ultrasonic transducer is always in a stable working state.

Images