Graphene-based Q-switched neodymium ruthenium tungstate human eye safety laser system

Abstract

The invention provides a graphene-based Q-switched neodymium ruthenium tungstate human eye safety laser system, and relates to the technical field of laser, and the system comprises a laser diode pumping source control module which is used for controlling the output power and the on-off state of a laser diode pumping source; the optical focusing coupling control module is used for adjusting the focusing position and focal length setting of optical focusing coupling; and the resonant cavity control module is used for solving a final solution by using the whale optimization algorithm and adjusting the positions and angles of an input cavity mirror and an output cavity mirror of the resonant cavity so as to optimize the optical performance of the resonant cavity. The passive Q-switched neodymium-doped potassium ruthenium tungstate laser based on the graphene absorber is comprehensively and accurately controlled by integrating a multifunctional module, the performance of the laser is optimized, the output stability and adaptability are improved, and efficient and reliable laser pulse generation is ensured.

Description

Technical Field

[0001] The invention relates to the field of laser technology, in particular to a graphene-based Q-switched neodymium ruthenium tungstate human eye-safe laser system. Background Art

[0002] Traditional control systems may be relatively complex in design, involving the fine-tuning of multiple parameters, such as the stability of the pump source, the alignment and adjustment of the resonant cavity, the saturation absorption characteristics of the graphene absorber, etc. This complexity may limit the system's adaptability and adjustability when facing different application environments or changes in demand, and requires a long time and professional knowledge to debug and optimize.

[0003] For example, in certain application scenarios, the control system may need to be customized to meet specific laser output requirements, which may increase the difficulty and cost of system implementation.

[0004] In addition, traditional technologies may face stability issues during long-term operation. Environmental factors (such as temperature fluctuations, mechanical vibrations) or aging of components within the system may affect the performance of the laser, resulting in a decrease in the stability of the output laser. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system, which realizes stable pulse output from visible light to mid-infrared bands.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows: First, the graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system includes: A laser diode pump source control module, used to control the output power and switch state of the laser diode pump source; An optical focus coupling control module, used for adjusting the focus position and focal length setting of the optical focus coupling; The resonant cavity control module is used to obtain the final solution using the whale optimization algorithm and adjust the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity to optimize the optical performance of the resonant cavity; A graphene saturable absorber control module, used to monitor and control the saturation absorption characteristics of the graphene saturable absorber through a neural network model to adjust the gain loss of the resonant cavity and generate Q-switched pulses; A temperature control module is used to control the temperature of the laser gain medium and the graphene saturable absorber so that the laser gain medium and the graphene saturable absorber maintain a constant temperature environment during operation; The monitoring and feedback module is used to monitor the output power, pulse width, and pulse repetition frequency parameters of the laser in real time, and perform feedback control based on the monitoring results to optimize the performance of the laser.

[0007] Furthermore, the whale optimization algorithm is used to obtain the final solution, and the positions and angles of the input cavity mirror and the output cavity mirror of the resonant cavity are adjusted to optimize the optical performance of the resonant cavity, including: Set the basic parameters of the whale optimization algorithm, including population size, maximum number of iterations, and search space dimension, and construct a performance evaluation function based on the optical performance indicators of the resonant cavity, including beam quality, stability, and efficiency, to evaluate the performance of the resonant cavity under the current individual; In each iteration, the performance of each individual in the population is evaluated according to the performance evaluation function, and the bubble net attack behavior of the whale is simulated, and the position and angle parameters of the individuals in the population are updated; Repeat the process of evaluation, updating and exploration until the maximum number of iterations is reached, and output the current final solution, that is, the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity; According to the final solution, the actual positions and angles of the input cavity mirror and the output cavity mirror of the laser resonator are adjusted to optimize the optical performance of the resonator.

[0008] Furthermore, an initial population is randomly generated in the search space, wherein each individual represents a set of specific position and angle parameters of the input cavity mirror and the output cavity mirror of the resonant cavity.

[0009] Furthermore, the calculation formula of the performance evaluation function is: ; in, represents the value of the performance evaluation function; , , represents the weight coefficient; Indicates that at a given cavity mirror position and angle Configuration, the square of the fundamental mode spot radius of the beam in the resonant cavity; represents the wavelength of the laser, i.e. a constant; , Indicates the position coordinates of the cavity mirror; represents the angle of the input cavity mirror; Indicates the angle of the output cavity mirror; Indicates that at a given cavity mirror position and angle Configuration, the square of the beam divergence angle; represents the beam waist radius; represents the divergence angle of the light beam; represents the cavity length of the resonant cavity; represents the effective reflectivity; Indicates the change in the mirror angle when the cavity length changes; Indicates a small change in cavity length; Represents the efficiency coefficient related to the cavity mirror position.

[0010] Furthermore, the saturation absorption characteristics of the graphene saturable absorber are monitored and controlled by a neural network model to adjust the gain loss of the resonant cavity and generate Q-switched pulses, including: Acquire data related to the saturation absorption characteristics of the graphene saturable absorber, including transmittance, reflectance, and absorptivity parameters, and use the data related to the saturation absorption characteristics of the graphene saturable absorber as a training set; Based on the training set, a neural network model is constructed, including determining the number of network layers, the number of neurons in each layer, the activation function, the loss function, and the optimizer parameters; The neural network model is trained using the training set, and during the training process, the training set is paired with output data, including gain loss adjustment values ​​of the resonant cavity; By learning the mapping relationship between the training set and the output data, the weights and biases of the neural network model are continuously adjusted through the back propagation algorithm to obtain the trained neural network model; During the operation of the laser, the saturation absorption characteristic parameters of the graphene saturable absorber are monitored in real time, and the real-time monitoring data are input into the trained neural network model to obtain the adjustment value of the resonant cavity gain loss; The gain loss of the resonant cavity is adjusted in real time according to the adjustment value of the gain loss of the resonant cavity, so that the laser outputs and generates Q-switched pulses.

[0011] Furthermore, by learning the mapping relationship between the training set and the output data, the weights and biases of the neural network model are continuously adjusted through the back propagation algorithm to obtain the trained neural network model, including: In each training iteration, the training set is forward propagated through the neural network model to obtain the predicted output of the neural network model; Use the loss function to calculate the error between the predicted output of the neural network model and the output data of the training set, that is, the loss value; According to the loss value, the gradient of each parameter in the neural network model is calculated through the back propagation algorithm; Use the optimizer to update the weights and biases of the neural network model based on the gradient of each parameter; Repeat the process of forward propagation, loss calculation, back propagation and weight update until the preset number of training times is reached to obtain the trained neural network model.

[0012] Furthermore, the temperature of the laser gain medium and the graphene saturable absorber is controlled so that the laser gain medium and the graphene saturable absorber maintain a constant temperature environment during operation, including: The current temperature data of the laser gain medium and the graphene saturable absorber are collected in real time by the temperature sensor, and the current temperature data is transmitted to the temperature controller; In the temperature controller, the adjusted heating or cooling power is calculated based on the collected current temperature data and the set target temperature to keep the temperature within a constant range; According to the heating or cooling power adjustment amount, the power of the heating and cooling elements is adjusted to achieve temperature control of the laser gain medium and the graphene saturable absorber; The temperature sensor continuously collects temperature data and continuously adjusts the heating or cooling power according to the real-time temperature data to keep the temperature of the laser gain medium and the graphene saturable absorber within a constant range.

[0013] Furthermore, the output power, pulse width, and pulse repetition frequency parameters of the laser are monitored in real time, and feedback control is performed based on the monitoring results to optimize the performance of the laser, including: The output power, pulse width and pulse repetition frequency parameter data are collected in real time through power sensors, pulse width meters and pulse repetition frequency meters; According to the performance indicators of the laser, set the feedback control target, and formulate the corresponding control strategy according to the feedback control target; According to the control strategy, determine the control parameters to be adjusted, including the pump power of the laser and the parameters of the Q-switching element; Continuously monitor the output power, pulse width, and pulse repetition frequency parameter data of the laser in real time, and compare the real-time monitored parameter data with the set target to calculate the deviation value; According to the control strategy and the deviation value, the determined control parameters are automatically adjusted, including adjusting the pump power and adjusting the parameters of the Q-switching element, so as to optimize the performance of the laser.

[0014] In a second aspect, a computing device includes: one or more processors; The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the system.

[0015] According to a third aspect, a computer-readable storage medium stores a program, and the program implements the system when executed by a processor.

[0016] The above solution of the present invention includes at least the following beneficial effects: The laser diode pump source control module can accurately control the output power and switching state of the pump source, provide a stable and adjustable energy input for the laser, and help to achieve precise control of the laser output. The optical focusing coupling control module ensures effective coupling between the pump light and the laser gain medium by adjusting the focus position and focal length settings, improves the utilization of light energy, and thus enhances the output efficiency and stability of the laser.

[0017] The resonant cavity control module uses the whale optimization algorithm to obtain the final solution and adjust the position and angle of the input cavity mirror and the output cavity mirror to optimize the optical performance of the resonant cavity. This helps to reduce the loss in the cavity, increase the gain of the laser, and thus improve the output power and beam quality of the laser. The graphene saturable absorber control module monitors and controls the saturated absorption characteristics of graphene through a neural network model, which can adjust the gain loss of the resonant cavity in real time and accurately control the generation of Q-switched pulses. This intelligent adjustment method improves the flexibility and adaptability of the laser, enabling it to meet the needs of different application scenarios.

[0018] The temperature control module accurately controls the temperature of the laser gain medium and graphene saturable absorber, so that they maintain a constant temperature environment during operation, which helps to reduce the impact of temperature fluctuations on laser performance and improve the stability and reliability of the laser. The monitoring and feedback module can monitor the key parameters of the laser such as output power, pulse width, pulse repetition frequency in real time, and perform feedback control based on the monitoring results. This real-time monitoring and feedback mechanism helps to promptly detect and correct abnormalities in the operation of the laser, optimize the performance of the laser, and ensure its long-term stable operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of a graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system provided in an embodiment of the present invention.

[0020] Figure 2 The present invention provides a flowchart of a graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system, which monitors and controls the saturation absorption characteristics of a graphene saturable absorber through a neural network model to adjust the gain loss of a resonant cavity and generate Q-switched pulses. DETAILED DESCRIPTION

[0021] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0022] like Figure 1As shown, an embodiment of the present invention proposes a graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system, comprising: A laser diode pump source control module 1, used to control the output power and switch state of the laser diode pump source; An optical focus coupling control module 2, used to adjust the focus position and focal length setting of the optical focus coupling; The resonant cavity control module 3 is used to obtain the final solution by using the whale optimization algorithm, and adjust the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity to optimize the optical performance of the resonant cavity; The graphene saturable absorber control module 4 is used to monitor and control the saturation absorption characteristics of the graphene saturable absorber through a neural network model to adjust the gain loss of the resonant cavity and generate Q-switched pulses; The temperature control module 5 is used to control the temperature of the laser gain medium and the graphene saturable absorber so that the laser gain medium and the graphene saturable absorber maintain a constant temperature environment during operation; The monitoring and feedback module 6 is used to monitor the output power, pulse width, and pulse repetition frequency parameters of the laser in real time, and to perform feedback control according to the monitoring results to optimize the performance of the laser.

[0023] In the embodiment of the present invention, the laser diode pump source control module can accurately control the output power and switch state of the pump source to ensure that the laser energy is input into the laser gain medium, which not only improves the energy utilization efficiency, but also reduces unnecessary energy loss, so that the laser can maintain high efficiency and stability during long-term operation. The optical focusing coupling control module can achieve precise control of the laser beam by adjusting the focus position and focal length setting, which helps to optimize the quality of the laser beam, improve the focusing degree and directionality of the beam, and thus meet the high requirements of different applications for the laser beam.

[0024] The resonant cavity control module uses the whale optimization algorithm to obtain the final solution and intelligently adjusts the position and angle of the input and output mirrors of the resonant cavity. This intelligent adjustment method can significantly optimize the optical performance of the resonant cavity, improve the output power and stability of the laser, and reduce optical loss and noise. The graphene saturable absorber control module monitors and controls the saturation absorption characteristics of the graphene saturable absorber through a neural network model. This precise control method can accurately adjust the gain loss of the resonant cavity, thereby generating stable Q-switched pulses, which not only helps to improve the pulse quality of the laser, but also broadens the application range of the laser.

[0025] The temperature control module controls the temperature of the laser gain medium and the graphene saturable absorber, so that they maintain a constant temperature environment during operation. This helps to reduce the impact of temperature fluctuations on laser performance and improve the long-term stability and reliability of the laser. The monitoring and feedback module can monitor the key parameters of the laser such as output power, pulse width, pulse repetition frequency in real time, and perform feedback control based on the monitoring results. This real-time monitoring and feedback mechanism can promptly detect and correct problems during the operation of the laser and optimize the performance of the laser.

[0026] In a preferred embodiment of the present invention, the laser diode pump source control module 1 is used to control the output power and switch state of the laser diode pump source; the optical focusing coupling control module 2 is used to adjust the focus position and focal length setting of the optical focusing coupling, which may include: In an embodiment of the present invention, a driving circuit is configured, including a power supply, a current controller and a switching circuit, to achieve precise power supply and control of the laser diode. A control signal is generated by a dedicated control chip, and the signal is used to adjust the output power and switching state of the laser diode. The control signal is amplified and filtered before being transmitted to the driving circuit. A power sensor is installed at the output end of the laser diode pump source to monitor the output power in real time. The monitored power signal is fed back to the control module, compared with the preset value, and the control signal is adjusted to achieve precise control. Calibration is performed regularly to ensure the accuracy and stability of the control module.

[0027] Develop a user-friendly software control interface that allows users to remotely control the output power and switch status of the laser diode through a computer or mobile device. It also provides the function of real-time display of output power, switch status and other information to facilitate user monitoring and adjustment.

[0028] According to the overall design requirements of the laser, the optical focusing coupling system is designed, including optical elements such as lenses and reflectors, the focus position and focal length settings are determined, and the mechanical structure is designed to fix and adjust the position of the optical elements. The structure should have sufficient rigidity and stability to ensure that the optical elements will not be displaced or deformed during the adjustment process. Equipped with precision adjustment mechanisms, such as fine-tuning screws, stepper motors, etc., for precise adjustment of the position and focal length of the optical elements. The adjustment mechanism should have sufficient resolution and repeatability to meet the accuracy requirements of optical focusing coupling. According to the preset focus position and focal length settings, the position of the optical elements is automatically adjusted to achieve automatic control of optical focusing coupling. A beam quality monitor is installed at the output end of the optical focusing coupling system to monitor the quality of the laser beam in real time. The monitored beam quality signal is fed back to the control system, compared with the preset value, and the position and focal length of the optical element are adjusted.

[0029] In a preferred embodiment of the present invention, the whale optimization algorithm is used to obtain a final solution, and the positions and angles of the input cavity mirror and the output cavity mirror of the resonant cavity are adjusted to optimize the optical performance of the resonant cavity, which may include: Set the basic parameters of the whale optimization algorithm, including population size, maximum number of iterations, and search space dimension, and construct a performance evaluation function based on the optical performance indicators of the resonant cavity, including beam quality, stability, and efficiency, to evaluate the performance of the resonant cavity under the current individual; In each iteration, the performance of each individual in the population is evaluated according to the performance evaluation function, and the bubble net attack behavior of the whale is simulated, and the position and angle parameters of the individuals in the population are updated; Repeat the process of evaluation, updating and exploration until the maximum number of iterations is reached, and output the current final solution, that is, the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity; According to the final solution, the actual positions and angles of the input cavity mirror and the output cavity mirror of the laser resonator are adjusted to optimize the optical performance of the resonator.

[0030] In an embodiment of the present invention, the parameters of the whale optimization algorithm are initialized, including the population size (i.e., the number of whale individuals), the maximum number of iterations (the number of cycles of the algorithm), and the search space dimension (corresponding to the adjustable range of the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity). According to the optical performance indicators of the resonant cavity, such as beam quality (such as the focusing degree and directivity of the beam), stability (such as the fluctuation of the output power), and efficiency (such as energy conversion efficiency), a performance evaluation function is defined, which accepts the position and angle of the input cavity mirror and the output cavity mirror as parameters, calculates and returns the resonant cavity performance score under the current configuration.

[0031] In each iteration, traverse each whale individual in the population (i.e., the position and angle configuration of each set of resonant cavity mirrors). For each individual, call the performance evaluation function to calculate its performance score, and record the score for subsequent comparison and update. According to the mechanism of the whale optimization algorithm, simulate the bubble net attack behavior of the whale to update the position and angle parameters of the individual. According to the simulated whale behavior, update the position and angle parameters of each individual in the population. Loop through the evaluation, update, and exploration steps until the maximum number of iterations is reached. In each iteration, record and update the final solution currently found. After the maximum number of iterations is reached, output the final solution currently found, that is, the position and angle of the resonant cavity input cavity mirror and the output cavity mirror. According to the final solution, physically adjust the actual position and angle of the input cavity mirror and the output cavity mirror of the laser resonant cavity.

[0032] Suppose you are optimizing a neodymium-doped potassium ruthenium tungstate The laser's resonant cavity aims to maximize beam quality and stability. The population size of the whale optimization algorithm is set to 50, the maximum number of iterations is 100, and the search space dimensions include the two-dimensional position and angle of the input cavity mirror and the output cavity mirror (a total of 4 dimensions). The performance evaluation function measures the indicators of beam quality and the stability of output power to calculate the score. In each iteration, the algorithm evaluates 50 individuals (i.e., 50 sets of mirror configurations) and then simulates the behavior of whales to update these configurations. After 100 iterations, the algorithm found a final set of mirror configurations.

[0033] The mirror configuration of the resonant cavity is optimized by the whale optimization algorithm, which improves the beam quality and stability of the laser. The optimized resonant cavity can use the pump light more effectively and improve the energy conversion efficiency of the laser. The automated optimization process reduces the experimental cost and time of manually adjusting the mirror configuration. The algorithm can quickly find the final solution, allowing the laser to be put into use more quickly. The optimized resonant cavity configuration can be adjusted according to different application requirements, enhancing the adaptability and flexibility of the laser. The laser can work stably under a wider range of conditions, improving its practical value. The application of the whale optimization algorithm in the optimization of laser resonant cavities provides new ideas and methods for the development of laser technology.

[0034] In another preferred embodiment of the present invention, an initial population is randomly generated in the search space, wherein each individual represents a set of specific position and angle parameters of the input cavity mirror and the output cavity mirror of the resonant cavity, which may include: In an embodiment of the present invention, first, the adjustable range of the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity is clarified, which constitutes the search space. The position parameters include the lateral and longitudinal positions of the mirror in the resonant cavity, and the angle parameters include the tilt angle or rotation angle of the mirror. According to the complexity of the problem and the computing resources, the population size of the whale optimization algorithm, that is, the number of individuals to be generated, is determined, and the random number generation algorithm is customized to ensure that the seed of the random number generator is random or changes according to a certain rule to increase the diversity of the search. Within the defined position range, a random number generator is used to generate the lateral and longitudinal position coordinates of the mirror, and within the defined angle range, a random number generator is used to generate the tilt angle or rotation angle of the mirror, and the generated position and angle parameters are combined into a complete individual, representing a set of mirror configurations of the resonant cavity.

[0035] In a preferred embodiment of the present invention, the calculation formula of the performance evaluation function is: ; in, represents the value of the performance evaluation function; , , represents the weight coefficient; Indicates that at a given cavity mirror position and angle Configuration, the square of the fundamental mode spot radius of the beam in the resonant cavity; represents the wavelength of the laser, i.e. a constant; , Indicates the position coordinates of the cavity mirror; represents the angle of the input cavity mirror; Indicates the angle of the output cavity mirror; Indicates that at a given cavity mirror position and angle Configuration, the square of the beam divergence angle; represents the beam waist radius; represents the divergence angle of the light beam; represents the cavity length of the resonant cavity; represents the effective reflectivity; Indicates the change in the mirror angle when the cavity length changes; Indicates a small change in cavity length; Represents the efficiency coefficient related to the cavity mirror position.

[0036] In the embodiment of the present invention, the weight coefficient is initialized , , , laser wavelength , Laparoscope position , (Assuming the cavity mirror is initially at the center of the resonant cavity), the angle , (Assuming the cavity mirrors are initially parallel.) Based on the given cavity mirror positions and angle configurations, the fundamental mode spot radius of the beam in the resonant cavity is calculated using a simulation method. , assuming that by calculation , then calculate and store , as a component of the performance evaluation function.

[0037] Similarly, the beam divergence angle is calculated based on the cavity mirror configuration , assuming that by calculation , then calculate and store , as another component of the performance evaluation function. For a given cavity length and angle configuration , , calculate the change in the mirror angle when the cavity length changes . Assume that through analysis and calculation (This is an assumed linear rate of change).

[0038] According to the position of the laparoscope , , calculate the coefficient related to efficiency , assuming that through simulation (Assuming that the efficiency is as high as 90% when the cavity mirror position is (0, 0), the specific value needs to be determined based on the analysis of the energy conversion efficiency in the resonant cavity. , , , , and the given weight coefficient , , , calculation performance evaluation function In each iteration of the whale optimization algorithm, the performance evaluation function is used to evaluate the performance of each individual in the current population. According to the evaluation results, the position and angle parameters of the individuals in the population are updated. After reaching the maximum number of iterations, the final solution currently found is output, that is, the optimal cavity mirror configuration of the resonant cavity (including the final value).

[0039] By optimizing the performance evaluation function, the final cavity mirror configuration can be found to improve the beam quality, stability and efficiency of the resonant cavity, which helps to improve the overall performance of the laser and meet a wider range of application needs. The flexibility of the performance evaluation function enables the Whale Optimization Algorithm to adapt to different resonant cavity designs and laser parameters, which means that the algorithm can be widely used in various types of lasers and promote the continuous advancement of laser technology. By automating the optimization process, the experimental cost and time of manually adjusting the cavity mirror configuration can be reduced, which helps to speed up the research and development and production of lasers and reduce manufacturing costs.

[0040] In a preferred embodiment of the present invention, monitoring and controlling the saturation absorption characteristics of the graphene saturable absorber by a neural network model to adjust the gain loss of the resonant cavity and generate Q-switched pulses may include: Acquire data related to the saturation absorption characteristics of the graphene saturable absorber, including transmittance, reflectance, and absorptivity parameters, and use the data related to the saturation absorption characteristics of the graphene saturable absorber as a training set; Based on the training set, a neural network model is constructed, including determining the number of network layers, the number of neurons in each layer, the activation function, the loss function, and the optimizer parameters; The neural network model is trained using the training set, and during the training process, the training set is paired with output data, including gain loss adjustment values ​​of the resonant cavity; By learning the mapping relationship between the training set and the output data, the weights and biases of the neural network model are continuously adjusted through the back propagation algorithm to obtain the trained neural network model; During the operation of the laser, the saturation absorption characteristic parameters of the graphene saturable absorber are monitored in real time, and the real-time monitoring data are input into the trained neural network model to obtain the adjustment value of the resonant cavity gain loss; The gain loss of the resonant cavity is adjusted in real time according to the adjustment value of the gain loss of the resonant cavity, so that the laser outputs and generates Q-switched pulses.

[0041] In an embodiment of the present invention, transmittance, reflectance, and absorptance sensors are deployed near the graphene saturable absorber to monitor these parameters in real time. Read data from the sensor regularly (e.g., once per second), record the corresponding timestamp, and perform preprocessing operations such as denoising and normalization on the collected data to ensure data quality. Organize the preprocessed data into a training set according to the time series, and each sample contains transmittance, reflectance, absorptance parameters and corresponding timestamps. According to the complexity of the problem and the amount of data, select the appropriate number of network layers (e.g., 3-5 layers), and select the appropriate number of neurons in each layer, which gradually decreases from the input layer to the output layer. Select activation functions such as ReLU, Sigmoid, or Tanh, and select according to the characteristics of the network layer. Select loss functions such as mean square error (MSE) or cross entropy, and select according to the output type (regression or classification).

[0042] Pair the transmittance, reflectivity, and absorptivity parameters in the training set with the gain loss adjustment value of the resonant cavity to form training samples. Use the training samples to train the neural network model, and continuously adjust the weights and biases of the model through the back propagation algorithm. During the training process, monitor the change trend of the loss function to ensure model convergence. Use the validation set to verify the trained model, evaluate the generalization ability of the model, and save the verified model as a file.

[0043] During the operation of the laser, the transmittance, reflectivity and absorptivity parameters of the graphene saturable absorber are monitored in real time, and the real-time monitored data are input into the trained neural network model. According to the output value of the model (the adjustment value of the resonant cavity gain loss), the gain loss of the resonant cavity is adjusted in real time.

[0044] Suppose there is a laser that contains a graphene saturable absorber. It is hoped that the saturation absorption characteristics of the absorber can be monitored and controlled through a neural network model to generate Q-switched pulses.

[0045] Transmittance, reflectance, and absorptance sensors were deployed, and data was read once per second. After preprocessing, a training set containing 1,000 samples was obtained. A neural network model with 4 layers was constructed, in which the input layer had 3 neurons (corresponding to transmittance, reflectance, and absorptance), the hidden layer had 64 neurons, and the output layer had 1 neuron (corresponding to the adjustment value of the resonant cavity gain loss). ReLU was selected as the activation function, MSE as the loss function, and Adam as the optimizer. The model was trained using the training set, and the trend of the loss function was monitored. After 100 epochs of training, the model converged and achieved good performance. The model was verified using the validation set, and it was confirmed that the model had good generalization ability. Then, the model was saved as a file. During the operation of the laser, the transmittance, reflectance, and absorptance parameters of the graphene saturable absorber were monitored in real time, and these data were input into the trained neural network model. The model output the adjustment value of the resonant cavity gain loss, and the resonant cavity gain loss was adjusted in real time according to this value, thereby successfully generating Q-switched pulses.

[0046] Through real-time monitoring and adjustment of the neural network model, the saturation absorption characteristics of the graphene saturable absorber can be controlled more accurately, thereby improving the stability and output performance of the laser, which helps to produce higher quality Q-switched pulses and meet the needs of a wider range of applications. The automated monitoring and adjustment process reduces the need for manual intervention and improves the automation of the laser, which helps to reduce operating costs and improve production efficiency. The neural network model is able to learn and adapt to different graphene saturable absorbers and laser configurations, which means that this method can be widely used in different types of lasers, promoting the continuous advancement of laser technology.

[0047] In another preferred embodiment of the present invention, by learning the mapping relationship between the training set and the output data, the weights and biases of the neural network model are continuously adjusted by the back propagation algorithm to obtain the trained neural network model, which may include: In each training iteration, the training set is forward propagated through the neural network model to obtain the predicted output of the neural network model; Use the loss function to calculate the error between the predicted output of the neural network model and the output data of the training set, that is, the loss value; According to the loss value, the gradient of each parameter in the neural network model is calculated through the back propagation algorithm; Use the optimizer to update the weights and biases of the neural network model based on the gradient of each parameter; Repeat the process of forward propagation, loss calculation, back propagation and weight update until the preset number of training times is reached to obtain the trained neural network model.

[0048] In an embodiment of the present invention, a suitable neural network structure (such as a feedforward neural network, a convolutional neural network, a recurrent neural network, etc.) is selected according to the task requirements (such as classification, regression, etc.). The weights and biases of the neural network are initialized, which is done by randomly assigning small values. In each training iteration, a set of samples (i.e., input data) in the training set is passed to the input layer of the neural network model. The input data is transmitted through the connection between the neural network layers, and each layer performs linear transformation and nonlinear activation (such as ReLU) on the input data according to the weights and biases. After calculations of all layers, the neural network model outputs the prediction results.

[0049] Use a predefined loss function (such as mean squared error MSE) to calculate the error between the predicted output of the neural network model and the output data of the training set (i.e., the true label). The loss function receives the predicted output and the true label as input, and returns a scalar value as the loss value, which represents the degree of mismatch between the model prediction and the true situation. According to the loss value, the gradient of each parameter in the neural network model is calculated by the back propagation algorithm. Back propagation starts from the output layer and calculates the gradient layer by layer through the chain rule until the input layer. The gradient represents the rate of change of the loss function relative to each parameter (weight and bias), which is the key to updating the parameters during the optimization process. Use an optimizer (such as stochastic gradient descent SGD, Adam, etc.) to update the weights and biases of the neural network model according to the gradient of each parameter. The optimizer receives the gradient as input and calculates the new weight and bias values ​​according to its internal algorithm (such as learning rate, momentum, etc.). The updated weights and biases are used for the forward propagation of the next training iteration.

[0050] Repeat the process of forward propagation, loss calculation, back propagation and weight update until the preset number of training times (i.e. Epoch number) is reached. During the entire training process, the performance of the neural network model gradually improves, and finally the trained neural network model is obtained.

[0051] Suppose there is a binary classification task, and a feedforward neural network model is used for training. A feedforward neural network is initialized with an input layer, a hidden layer, and an output layer. The weights and biases are initialized by randomly assigning small values. The first sample in the training set (such as a two-dimensional vector) is passed to the input layer.

[0052] The input layer passes the sample to the hidden layer, and the hidden layer performs linear transformation and nonlinear activation (such as ReLU) according to the weights and biases. The output of the hidden layer is then passed to the output layer, and the output layer performs linear transformation and Sigmoid activation according to the weights and biases to obtain the predicted probability. The cross entropy loss function is used to calculate the error between the predicted probability and the true label (0 or 1). Assuming the true label is 1 and the predicted probability is 0.8, the loss value is small; if the predicted probability is 0.2, the loss value is large.

[0053] Based on the loss value, the gradient of each parameter in the neural network model is calculated by the back propagation algorithm. The gradient represents the rate of change of the loss function relative to each parameter. For example, if the gradient of a weight parameter is 0.05, it means that the loss function increases as the weight parameter increases by 0.05. The stochastic gradient descent optimizer is used to update the weights and biases of the neural network model. Assuming the learning rate is 0.1, a weight parameter will be subtracted from its gradient multiplied by the learning rate (i.e. 0.05). 0.1=0.005), repeat the above process, perform forward propagation, loss calculation, back propagation and weight update for each sample in the training set. After multiple iterations, the performance of the neural network model gradually improves, and finally the trained model is obtained.

[0054] Through repeated training and weight adjustment, the neural network model can gradually learn the mapping relationship in the training set, thereby improving the prediction accuracy. The trained neural network model can not only accurately predict the samples in the training set, but also make reasonable predictions for unseen samples, showing good generalization ability. The neural network model can handle complex nonlinear relationships and high-dimensional data, and is suitable for various complex machine learning tasks, such as image recognition, natural language processing, etc. It can automatically complete the process of forward propagation, loss calculation, back propagation and weight update without manual intervention, which improves the training efficiency.

[0055] In a preferred embodiment of the present invention, the temperature of the laser gain medium and the graphene saturable absorber is controlled so that the temperature of the laser gain medium and the graphene saturable absorber is kept within a constant range during operation, which may include: The current temperature data of the laser gain medium and the graphene saturable absorber are collected in real time by the temperature sensor, and the current temperature data is transmitted to the temperature controller; In the temperature controller, the adjusted heating or cooling power is calculated based on the collected current temperature data and the set target temperature to keep the temperature within a constant range; According to the heating or cooling power adjustment amount, the power of the heating and cooling elements is adjusted to achieve temperature control of the laser gain medium and the graphene saturable absorber; The temperature sensor continuously collects temperature data and continuously adjusts the heating or cooling power according to the real-time temperature data to keep the temperature of the laser gain medium and the graphene saturable absorber within a constant range.

[0056] In an embodiment of the present invention, the temperature sensor is designed to accurately and in real time monitor the surface or internal temperature of the laser gain medium and the graphene saturable absorber. The sensor converts the temperature into an electrical signal through physical contact (such as a thermistor, a thermocouple), and the converted electrical signal is converted into a digital signal through an analog-to-digital converter (ADC). The digital temperature signal is transmitted to the temperature controller through a wired (such as I2C) communication interface. The temperature data received by the temperature controller is stored in a memory or register for further processing.

[0057] The temperature controller runs a control algorithm (such as PID control) inside, which receives the current temperature data and the set target temperature as input. Based on the deviation between the current temperature and the target temperature, as well as the rate of change of the deviation (if the algorithm takes dynamic characteristics into account), it calculates the heating or cooling power that needs to be adjusted. ;in, It is The heating or cooling power that needs to be adjusted at each sampling moment; is the proportional gain coefficient; It is Temperature deviation at each sampling moment; is the integral gain coefficient; is the index of the sampling moment; is the index variable to be summed; It is Temperature deviation at each sampling moment; is the sampling time interval; is the differential gain coefficient; It is The temperature deviation at each sampling moment. The calculation results are output in the form of digital signals for controlling heating and cooling elements. The temperature controller controls the power of heating elements (such as resistors, heating sheets) and cooling elements (such as fans, liquid cooling pumps) through digital output signals (such as PWM signals). The heating element increases or decreases the heat output according to the received signal, while the cooling element increases or decreases the cooling effect accordingly. In this way, the temperature of the laser gain medium and the graphene saturable absorber are adjusted to close to the target temperature. The temperature sensor continuously collects temperature data and transmits the new data to the temperature controller. The temperature controller continuously updates the calculation results of the heating or cooling power based on the new data, and adjusts the power of the heating and cooling elements accordingly. This process forms a closed-loop control system to ensure that the temperature of the laser gain medium and the graphene saturable absorber remains within a constant range.

[0058] Suppose there is a laser consisting of a laser gain medium and a graphene saturable absorber, and both components need to be temperature controlled.

[0059] The temperature sensor is installed near the laser gain medium and the graphene saturable absorber to monitor the temperature in real time. The temperature controller is configured to receive the temperature data from the sensor and set the target temperature to 25°C. The temperature sensor collects temperature data every 1 second and transmits the data to the temperature controller through the I2C interface. Assume that the current temperature of the laser gain medium is 24.5°C and the temperature of the graphene saturable absorber is 25.2°C. After receiving the temperature data, the temperature controller runs the PID control algorithm. The algorithm calculates that the power of the heating element needs to be fine-tuned to compensate for the temperature deviation while keeping the power of the cooling element unchanged (because the overall temperature is close to the target). The temperature controller increases the power of the heating element through a PWM signal to make it output more heat. The temperature sensor continues to collect temperature data and transmits it to the temperature controller. The temperature controller continuously adjusts the power of the heating and cooling elements based on the new data to ensure that the temperature of the laser gain medium and the graphene saturable absorber remains around 25°C.

[0060] Through precise temperature control, the operating temperature of the laser gain medium and graphene saturable absorber is kept within a constant range, improving the stability of the entire laser system. A stable temperature environment helps reduce the thermal stress of the laser gain medium and graphene saturable absorber, thereby extending their service life. Precise temperature control helps maintain the gain characteristics of the laser gain medium and the saturation characteristics of the graphene saturable absorber, thereby improving the output performance and stability of the laser. The closed-loop control system can automatically respond to temperature changes and environmental interference, ensuring that the laser system can operate reliably under various conditions. Precise temperature control technology provides a basis for further optimization and innovation of laser systems, and helps promote the development and application of related technologies.

[0061] In a preferred embodiment of the present invention, real-time monitoring of the output power, pulse width, and pulse repetition frequency parameters of the laser, and performing feedback control based on the monitoring results to optimize the performance of the laser may include: The output power, pulse width and pulse repetition frequency parameter data are collected in real time through power sensors, pulse width meters and pulse repetition frequency meters; According to the performance indicators of the laser, set the feedback control target, and formulate the corresponding control strategy according to the feedback control target; According to the control strategy, determine the control parameters to be adjusted, including the pump power of the laser and the parameters of the Q-switching element; Continuously monitor the output power, pulse width, and pulse repetition frequency parameter data of the laser in real time, and compare the real-time monitored parameter data with the set target to calculate the deviation value; According to the control strategy and the deviation value, the determined control parameters are automatically adjusted, including adjusting the pump power and adjusting the parameters of the Q-switching element, so as to optimize the performance of the laser.

[0062] In the embodiment of the present invention, the power sensor is precisely calibrated and placed at the output end of the laser to measure the output power of the laser in real time. The pulse width meter captures the rising and falling edges of the laser pulse through a high-speed photodetector to calculate the pulse width. The pulse repetition frequency meter determines the pulse repetition frequency by counting the number of laser pulses per unit time. These three instruments convert the measured data into electrical signals and convert them into digital signals through an analog-to-digital converter (ADC).

[0063] According to the design specifications and application requirements of the laser, the target values ​​of output power, pulse width and pulse repetition frequency are set. Based on these target values, a set of feedback control strategies are formulated, which include PID control, fuzzy control, adaptive control, etc., to adjust the parameters of the laser according to the deviation between the real-time measurement value and the target value. According to the control strategy, the control parameters that need to be adjusted are determined, which include the pump power of the laser (used to control the gain of the laser) and the parameters of the Q-switched element (such as the refractive index and transmittance of the Q-switched crystal, which are used to control the formation of laser pulses). A mathematical model between these parameters and the laser performance is established to enable precise adjustment during the feedback control process.

[0064] Power sensors, pulse width meters, and pulse repetition frequency meters continuously monitor the output parameters of the laser in real time and transmit new data to the machine's control unit. The control unit compares the parameter data monitored in real time with the target value and calculates the deviation value, which reflects the difference between the current performance of the laser and the target performance. The control unit automatically adjusts the control parameters of the laser according to the control strategy and the calculated deviation value. For example, if the output power is lower than the target value, the control unit may increase the pump power to increase the gain of the laser; if the pulse width does not meet the requirements, the parameters of the Q-switched element may be adjusted to optimize the formation of the laser pulse. The adjusted control parameters are applied to the laser in real time to optimize its performance.

[0065] Suppose there is a pulsed laser used for precision machining, and its output power, pulse width, and pulse repetition frequency need to be monitored and optimized in real time.

[0066] The power sensor, pulse width meter and pulse repetition frequency meter are used to monitor the output parameters of the laser in real time. According to the processing requirements, the target values ​​of output power of 10W, pulse width of 100ns and pulse repetition frequency of 1kHz are set. The power sensor, pulse width meter and pulse repetition frequency meter collect data once a second and transmit the data to the control unit of the machine through the wired interface. Assume that the current output power is 9.5W, the pulse width is 105ns and the pulse repetition frequency is 980Hz. After receiving the data, the control unit calculates that the output power deviation is -0.5W, the pulse width deviation is +5ns, and the pulse repetition frequency deviation is -20Hz. According to these deviation values ​​and the preset control strategy (such as PID control), the control unit formulates an adjustment plan. The control unit compensates for the deviation of the output power by increasing the pump power, and fine-tunes the parameters of the Q-switched element to optimize the pulse width and pulse repetition frequency. The adjusted parameters are applied to the laser in real time, so that its output performance gradually approaches the target value. The power sensor, pulse width meter and pulse repetition frequency meter continue to monitor the output parameters of the laser in real time and transmit the new data to the control unit, which continuously adjusts the control parameters based on the new data.

[0067] Through real-time monitoring and feedback control, the laser output power, pulse width and pulse repetition frequency are kept within a constant range, which improves the stability of the laser. Stable laser output parameters help improve the quality of laser processing, such as reducing processing errors and improving processing accuracy. Accurate control reduces damage to the laser due to overload or underload, thereby extending its service life. Stable laser output and optimized processing parameters improve production efficiency and reduce downtime caused by unstable laser performance.

[0068] The embodiment of the present invention further provides a computing device, including: a processor, a memory storing a computer program, and when the computer program is executed by the processor, the system described above is executed. All implementation methods in the above system embodiment are applicable to this embodiment and can achieve the same technical effect.

[0069] The embodiment of the present invention further provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the system described above. All implementations in the above system embodiment are applicable to this embodiment and can achieve the same technical effect.

[0070] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system, characterized in that: include: A laser diode pump source control module, used to control the output power and switch state of the laser diode pump source; An optical focus coupling control module, used for adjusting the focus position and focal length setting of the optical focus coupling; The resonant cavity control module is used to obtain the final solution using the whale optimization algorithm and adjust the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity to optimize the optical performance of the resonant cavity; A graphene saturable absorber control module, used to monitor and control the saturation absorption characteristics of the graphene saturable absorber through a neural network model to adjust the gain loss of the resonant cavity and generate Q-switched pulses; A temperature control module is used to control the temperature of the laser gain medium and the graphene saturable absorber so that the laser gain medium and the graphene saturable absorber maintain a constant temperature environment during operation; The monitoring and feedback module is used to monitor the output power, pulse width, and pulse repetition frequency parameters of the laser in real time, and perform feedback control based on the monitoring results to optimize the performance of the laser.

2. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 1, characterized in that: The whale optimization algorithm is used to obtain the final solution and adjust the position and angle of the input and output mirrors of the resonant cavity to optimize the optical performance of the resonant cavity, including: Set the basic parameters of the whale optimization algorithm, including population size, maximum number of iterations, and search space dimension, and construct a performance evaluation function based on the optical performance indicators of the resonant cavity, including beam quality, stability, and efficiency, to evaluate the performance of the resonant cavity under the current individual; In each iteration, the performance of each individual in the population is evaluated according to the performance evaluation function, and the bubble net attack behavior of the whale is simulated, and the position and angle parameters of the individuals in the population are updated; Repeat the process of evaluation, updating and exploration until the maximum number of iterations is reached, and output the current final solution, that is, the position and angle of the input cavity mirror and the output cavity mirror of the resonant cavity; According to the final solution, the actual positions and angles of the input cavity mirror and the output cavity mirror of the laser resonator are adjusted to optimize the optical performance of the resonator.

3. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 2, characterized in that: An initial population is randomly generated in the search space, wherein each individual represents a set of specific position and angle parameters of the input cavity mirror and the output cavity mirror of the resonant cavity.

4. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 3, characterized in that: The calculation formula of the performance evaluation function is: ; in, represents the value of the performance evaluation function; , , represents the weight coefficient; Indicates that at a given cavity mirror position and angle Configuration, the square of the fundamental mode spot radius of the beam in the resonant cavity; represents the wavelength of the laser, i.e. a constant; , Indicates the position coordinates of the cavity mirror; represents the angle of the input cavity mirror; Indicates the angle of the output cavity mirror; Indicates that at a given cavity mirror position and angle Configuration, the square of the beam divergence angle; represents the beam waist radius; represents the divergence angle of the light beam; represents the cavity length of the resonant cavity; represents the effective reflectivity; Indicates the change in the mirror angle when the cavity length changes; Indicates a small change in cavity length; Represents the efficiency coefficient related to the cavity mirror position.

5. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 4, characterized in that: The saturation absorption characteristics of the graphene saturable absorber are monitored and controlled by a neural network model to adjust the gain loss of the resonant cavity and generate Q-switched pulses, including: Acquire data related to the saturation absorption characteristics of the graphene saturable absorber, including transmittance, reflectance, and absorptivity parameters, and use the data related to the saturation absorption characteristics of the graphene saturable absorber as a training set; Based on the training set, a neural network model is constructed, including determining the number of network layers, the number of neurons in each layer, the activation function, the loss function, and the optimizer parameters; The neural network model is trained using the training set, and during the training process, the training set is paired with output data, including gain loss adjustment values ​​of the resonant cavity; By learning the mapping relationship between the training set and the output data, the weights and biases of the neural network model are continuously adjusted through the back propagation algorithm to obtain the trained neural network model; During the operation of the laser, the saturation absorption characteristic parameters of the graphene saturable absorber are monitored in real time, and the real-time monitoring data are input into the trained neural network model to obtain the adjustment value of the resonant cavity gain loss; The gain loss of the resonant cavity is adjusted in real time according to the adjustment value of the gain loss of the resonant cavity, so that the laser outputs and generates Q-switched pulses.

6. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 5, characterized in that: By learning the mapping relationship between the training set and the output data, the weights and biases of the neural network model are continuously adjusted through the back propagation algorithm to obtain the trained neural network model, including: In each training iteration, the training set is forward propagated through the neural network model to obtain the predicted output of the neural network model; Use the loss function to calculate the error between the predicted output of the neural network model and the output data of the training set, that is, the loss value; According to the loss value, the gradient of each parameter in the neural network model is calculated through the back propagation algorithm; Use the optimizer to update the weights and biases of the neural network model based on the gradient of each parameter; Repeat the process of forward propagation, loss calculation, back propagation and weight update until the preset number of training times is reached to obtain the trained neural network model.

7. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 6, characterized in that: The temperature of the laser gain medium and the graphene saturable absorber is controlled so that the laser gain medium and the graphene saturable absorber maintain a constant temperature environment during operation, including: The current temperature data of the laser gain medium and the graphene saturable absorber are collected in real time by the temperature sensor, and the current temperature data is transmitted to the temperature controller; In the temperature controller, the adjusted heating or cooling power is calculated based on the collected current temperature data and the set target temperature to keep the temperature within a constant range; According to the heating or cooling power adjustment amount, the power of the heating and cooling elements is adjusted to achieve temperature control of the laser gain medium and the graphene saturable absorber; The temperature sensor continuously collects temperature data and continuously adjusts the heating or cooling power according to the real-time temperature data to keep the temperature of the laser gain medium and the graphene saturable absorber within a constant range.

8. The graphene-based Q-switched neodymium ruthenium tungstate eye-safe laser system according to claim 7, characterized in that: Real-time monitoring of the laser's output power, pulse width, and pulse repetition frequency parameters, and feedback control based on the monitoring results to optimize the laser's performance, including: The output power, pulse width and pulse repetition frequency parameter data are collected in real time through power sensors, pulse width meters and pulse repetition frequency meters; According to the performance indicators of the laser, set the feedback control target, and formulate the corresponding control strategy according to the feedback control target; According to the control strategy, determine the control parameters to be adjusted, including the pump power of the laser and the parameters of the Q-switching element; Continuously monitor the output power, pulse width, and pulse repetition frequency parameter data of the laser in real time, and compare the real-time monitored parameter data with the set target to calculate the deviation value; According to the control strategy and the deviation value, the determined control parameters are automatically adjusted, including adjusting the pump power and adjusting the parameters of the Q-switching element, so as to optimize the performance of the laser.

9. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the system according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, and when the program is executed by a processor, the system according to any one of claims 1 to 8 is implemented.

Images