Programmable waveform modulation high-power fiber laser system

Abstract

The invention discloses a programmable waveform modulation high-power optical fiber laser system, particularly relates to the field of high-power optical fiber lasers, and is used for solving the problem of power monitoring and control instability caused by alias distortion under high pulse repetition frequency. Micro-amplitude random time jitter is implanted, sampling and pulse rhythm coupling are scattered through synchronous mapping of a time mark and a power sequence, and turn-back pseudo fluctuation is blocked from the source; the abstract evaluation mechanism fuses energy domain and time domain features, and uses a single determination identifier to drive sampling frequency adaptive migration, so that a monitoring link is continuously in a co-prime fraction ratio operation interval, and alias distortion root causes are eliminated; the sampling label and the judgment identification jointly trigger a feed-forward gain matrix to form a monitoring evaluation adjustment compensation closed loop, power aftershock is reduced in real time, stimulated scattering and thermal lens linkage amplification are inhibited, and light beam quality stability and energy smooth distribution are maintained.

Description

Technical Field

[0001] The present invention relates to the field of high-power fiber lasers, and more particularly to a programmable waveform-modulated high-power fiber laser system. Background Art

[0002] In programmable waveform-modulated high-power fiber laser architectures, pulse repetition frequencies often climb to the megahertz range to meet the dual demands of energy density and waveform flexibility for high-speed processes such as steel welding and additive manufacturing. Using full-bandwidth sampling in the power monitoring link would force the analog-to-digital converter to maintain an excessively high sampling rate, introducing significant data throughput and heat dissipation burdens. Engineering practices typically utilize band-limited sampling, capturing only the characteristic frequency band of the power envelope and then employing a feedback loop to stabilize the output. This strategy is stable within the low to medium modulation frequency range.

[0003] As the modulation frequency continues to increase, the sampling rate and pulse repetition frequency gradually approach an integer multiple relationship. The discrete sampling sequence folds back the power pulsations originally located in the high-frequency region to the low-frequency region, causing alias distortion. The feedback algorithm misinterprets the folded-back component as true slow drift and issues a reverse correction command, amplifying the power oscillations. This oscillation amplification not only prematurely triggers the stimulated scattering threshold and amplifies spectral noise, but also leads to enhanced thermal lensing in the fiber core and uneven machining penetration, seriously compromising the consistency of high-precision machining.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a programmable waveform modulated high-power fiber laser system, which implants slight random time jitter through a preprocessing link around the programmable waveform modulated high-power fiber laser, breaks up the sampling and pulse rhythm coupling through synchronous mapping of the time scale and power sequence, and blocks the return pseudo-fluctuation from the source; the abstract evaluation mechanism integrates the energy domain and time domain characteristics, and drives the adaptive migration of the sampling frequency with a single judgment identifier, so that the monitoring link is continuously in the coprime fraction ratio operating range, eliminating the root cause of alias distortion; the sampling label and the judgment identifier jointly trigger the feedforward gain matrix to form a monitoring, evaluation, adjustment and compensation closed loop, which reduces power aftershock in real time, suppresses stimulated scattering and thermal lens chain amplification, maintains stable beam quality and smooth energy distribution, and solves the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: A programmable waveform-modulated high-power fiber laser system, comprising: Jitter injection module, signature evaluation module, frequency adjustment module and power stabilization module; Jitter injection module: injects slight random time jitter into the power monitoring branch and records the timestamp sequence. It simultaneously captures the power sampling sequence and outputs a synchronously aligned data stream. Feature evaluation module: This module inputs the synchronously aligned data stream into the sparse reconstruction model to separate the return energy component, extracts the sampling and pulse coupling features, performs abstract evaluation, determines the degree of decoupling between the sampling link and the pulse link, and obtains a qualified mark. Frequency adjustment module: reads the qualified mark, adjusts the real-time sampling frequency to a fractional ratio that is coprime with the pulse repetition frequency, and simultaneously updates the sampling frequency label; Power stabilization module: This module sets the feedforward gain matrix based on the sampling frequency label and the pass indicator, and corrects the pump drive signal to keep the output power aftershock within the set upper limit.

[0007] In a preferred embodiment, the jitter injection module includes the following: In the power monitoring branch, an independent slight random time offset is introduced for each sampling period to disrupt the fixed correspondence between the sampling moment and the pulse moment. The slight random time offset is generated through uniform distribution. At the same time, the actual sampling moment of each sampling point is recorded. The actual sampling moment is the sum of the nominal sampling moment and the slight random time offset. The nominal sampling moment is the product of the sampling point number and the sampling period. Furthermore, at each actual sampling moment, a high-speed analog-to-digital converter is used to collect the instantaneous value of the laser output power to form a power sampling sequence. Finally, the actual sampling moment is paired with the corresponding instantaneous power value to generate a synchronously aligned data stream.

[0008] In a preferred embodiment, the feature evaluation module includes the following: The power sampling sequence in the synchronous aligned data stream is sparsely decomposed using a dictionary learning method to separate the true power component and the retrieval energy component, from which the sampling and pulse coupling characteristics are extracted, including the retrieval energy amplitude and jitter dispersion rate.

[0009] In a preferred embodiment, the feature evaluation module further includes the following: By collecting the reentry energy rate amplitude sequence and the jitter dispersion rate sequence within a sliding time window, a joint probability density field is constructed. The Wasserstein distance with the reference decoupling distribution is calculated using the optimal transmission algorithm to obtain the minimum transmission potential energy sequence. Multi-scale empirical mode decomposition is then performed on it to extract representative eigencomponents. After radial normalized kernel mapping, adaptive weighted aggregation of local entropy is used to generate a coupling strength label. The degree of decoupling between the sampling link and the pulse link is judged to obtain a qualified label.

[0010] In a preferred embodiment, the feature evaluation module further includes the following: The impact of alias distortion is quantified by calculating the amplitude of the folded energy rate by performing fast Fourier transform on the folded energy component and combining it with an adaptive frequency band selection method based on spectral entropy.

[0011] In a preferred embodiment, the feature evaluation module further includes the following: The deviation between the actual sampling time and the nominal sampling time is calculated, and the mutual information entropy method is used to calculate the jitter dispersion rate and quantify the randomness of random time jitter.

[0012] In a preferred embodiment, the frequency adjustment module includes the following contents: When the qualification mark is qualified, the sampling frequency adjustment process is started; the current pulse repetition frequency and the current sampling frequency are obtained, and the greatest common divisor of the current pulse repetition frequency and the current sampling frequency is calculated. If the greatest common divisor is 1, the current sampling frequency is kept unchanged; if the greatest common divisor is greater than 1, the adjustment step value is set, and the target sampling frequency is adjusted iteratively until the greatest common divisor of the target sampling frequency and the pulse repetition frequency is 1, while ensuring that the target sampling frequency is within the hardware sampling rate range.

[0013] In a preferred embodiment, the frequency adjustment module further includes the following: A frequency adjustment instruction including a target sampling frequency is generated and sent to a sampling clock generator; the sampling clock generator adjusts the internal clock frequency according to the frequency adjustment instruction so that the real-time sampling frequency matches the target sampling frequency; the adjusted sampling frequency is written into a sampling frequency tag and stored in a register.

[0014] In a preferred embodiment, the power stabilization module includes the following: The sampling frequency label and the qualified flag are read from the register to obtain the adjusted sampling frequency and the coupling status of the sampling frequency and the pulse repetition frequency; when the qualified flag indicates that the coupling characteristic index is less than 1, the reference feedforward gain matrix is ​​adjusted according to the ratio of the sampling frequency to the pulse repetition frequency, combined with the preset reference ratio and the adjustment coefficient, to generate a new feedforward gain matrix.

[0015] In a preferred embodiment, the power stabilization module further includes the following: The new feedforward gain matrix is ​​used to correct the original pump drive signal to generate a corrected pump drive signal, which is sent to the pump source controller to adjust the output power of the pump source. The real-time value of the laser output power is continuously monitored, and the difference between it and the target power value is calculated as the output power aftershock. When the absolute value of the output power aftershock exceeds the preset upper limit, the proportional-integral-differential control algorithm is used to dynamically adjust the feedforward gain matrix until the absolute value of the output power aftershock is less than or equal to the preset upper limit.

[0016] The technical effects and advantages of the programmable waveform modulated high-power fiber laser system of the present invention are as follows: The present invention focuses on the pre-processing link of a programmable waveform-modulated high-power fiber laser, implants a slight random time jitter, and breaks up the coupling of sampling and pulse rhythm through synchronous mapping of time scale and power sequence, thus blocking the return pseudo-fluctuation at the source. The abstract evaluation mechanism integrates the energy domain and time domain characteristics, and drives the adaptive migration of sampling frequency with a single judgment mark, so that the monitoring link is continuously in the coprime fraction ratio operation range, eliminating the root cause of alias distortion. The sampling label and the judgment mark jointly trigger the feedforward gain matrix to form a monitoring, evaluation, adjustment and compensation closed loop, reducing power aftershocks in real time, suppressing stimulated scattering and thermal lens chain amplification, and maintaining stable beam quality and smooth energy distribution. This improves the consistency of the molten pool depth, makes it difficult for spatter and microcracks to accumulate, expands the overall processing window width, and allows the equipment to maintain accurate output and high reliability for a long time under high repetition rate operation, giving industrial precision manufacturing greater flexibility and quality predictability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic structural diagram of a programmable waveform modulated high-power fiber laser system of the present invention; Figure 2 A schematic flow chart of an abstract evaluation process of a programmable waveform-modulated high-power fiber laser system according to the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Example 1: Figure 1 The present invention provides a programmable waveform modulated high-power fiber laser system, comprising: a jitter injection module, a characteristic evaluation module, a frequency adjustment module and a power stabilization module; Jitter injection module: Injects slight random time jitter into the power monitoring branch and records the timestamp sequence. It simultaneously captures the power sampling sequence and outputs a synchronously aligned data stream.

[0020] Feature evaluation module: Input the synchronous aligned data stream into the sparse reconstruction model to separate the return energy component, extract the sampling and pulse coupling features to perform abstract evaluation, determine the degree of decoupling between the sampling link and the pulse link, and obtain a qualified mark.

[0021] Frequency adjustment module: reads the qualified mark, adjusts the real-time sampling frequency to a fractional ratio that is coprime with the pulse repetition frequency, and simultaneously updates the sampling frequency label.

[0022] Power stabilization module: This module sets the feedforward gain matrix based on the sampling frequency label and the pass indicator, and corrects the pump drive signal to keep the output power aftershock within the set upper limit.

[0023] In modern industrial precision manufacturing, high-speed processes such as steel plate welding and additive manufacturing, programmable waveform-modulated high-power fiber laser systems have become a core technology, offering high energy density and flexible waveform control. These systems meet the dual requirements of processing speed and precision by increasing the pulse repetition frequency (PRF) to the megahertz (MHz) range. However, high PRF presents a challenge in power monitoring. Traditional full-bandwidth sampling methods require the analog-to-digital converter (ADC) to operate at extremely high sampling rates, resulting in a surge in data throughput and exacerbated heat dissipation issues. To address this, engineering practices often employ band-limited sampling techniques, capturing only the characteristic frequency band of the power envelope and stabilizing the output through a feedback loop. However, when the modulation frequency approaches or exceeds an integer multiple of the sampling rate, the discrete sampling sequence folds high-frequency power pulsations back to the low-frequency region, resulting in aliasing distortion. This distortion is misinterpreted by the feedback algorithm as slow drift, causing amplified power oscillations, which in turn triggers stimulated scattering, enhances thermal lensing, and ultimately compromises process consistency. To solve this problem, the present invention proposes a programmable waveform-modulated high-power fiber laser system, in which a jitter injection module injects slight random time jitter to break the coupling between sampling and pulse rhythm from the source.

[0024] The jitter injection module includes the following: S1.1, inject slight random time jitter: In the sampling clock of the power monitoring branch, an independent random time offset is introduced for each sampling period to disrupt the fixed correspondence between the sampling moment and the pulse moment.

[0025] First, the reciprocal of the sampling frequency is determined as the sampling period. For example, if the sampling frequency is set to 10 million times per second, the sampling period is 100 nanoseconds. Next, a random time offset is generated for each sampling point, with its amplitude range set to 1% to 5% of the sampling period. Taking a sampling period of 100 nanoseconds as an example, if the proportional coefficient is selected as 0.03, the range of the random time offset is -3 nanoseconds to +3 nanoseconds. The value of the random time offset is generated using a uniform distribution, that is, any value within the set range is randomly selected with equal probability. Ultimately, the actual sampling time of each sampling point is determined by adding this random time offset to the nominal sampling time, where the nominal sampling time is the sampling point number multiplied by the sampling period.

[0026] The purpose of introducing slight random time jitter is to break the integer multiple relationship between the sampling rate and the pulse repetition frequency, thereby preventing aliasing distortion caused by high-frequency power pulsations folding back into the low-frequency region. This perturbation method can block data distortion caused by fixed coupling at the source, providing a more accurate raw data foundation for subsequent processing. At the same time, the amplitude of the random time offset is limited to a small range, ensuring that it does not significantly affect sampling accuracy and effectively decouples the correlation between the sampling time and the pulse rhythm.

[0027] S1.2, record timestamp sequence: The actual sampling moment is recorded for each sampling point to generate a timestamp sequence, ensuring the accurate correspondence between the time information and the power data.

[0028] First, the nominal sampling time of each sampling point is calculated as the sampling point number multiplied by the sampling period. For example, if the sampling period is 100 nanoseconds, the nominal sampling time of the tenth sampling point is 1000 nanoseconds. Then, the random time offset of each sampling point is added to its nominal sampling time to obtain the actual sampling time. For example, if the random time offset of the tenth sampling point is 2 nanoseconds, its actual sampling time is 1002 nanoseconds. The actual sampling times of all sampling points are recorded sequentially to form a timestamp sequence. The timestamp sequence requires the use of a high-precision time counter with a time resolution better than the magnitude of the random time offset, for example, a resolution of 1 nanosecond or higher.

[0029] S1.3, capture power sampling sequence: On the basis of injecting slight random time jitter, a high-speed analog-to-digital converter is used to collect the instantaneous value of the laser output power at each actual sampling moment to form a power sampling sequence.

[0030] At each actual sampling moment recorded in the timestamp sequence, the analog-to-digital converter (ADC) is triggered to perform power sampling. For example, if the actual sampling moment at a particular sampling point is 1002 nanoseconds, the instantaneous value of the laser output power is acquired at this moment. The instantaneous power values ​​at all sampling points are recorded sequentially, generating a power sampling sequence that matches the length of the timestamp sequence. The sampling accuracy of the ADC must meet the requirements for power monitoring and accurately reflect the characteristic changes in the power envelope. For example, a 12-bit or higher accuracy converter should be used.

[0031] S1.4, output synchronization alignment data stream: The timestamp sequence is integrated with the power sampling sequence to generate a synchronously aligned data stream containing the sampling moments and the corresponding power values, which serves as the input for subsequent steps.

[0032] Each actual sampling moment in the timestamp sequence is paired with its corresponding instantaneous power value in the power sampling sequence to form a data pair. For example, if the actual sampling moment of a sampling point is 1002 nanoseconds and the corresponding instantaneous power value is a certain value, a data pair is generated containing this moment and this power value. All data pairs are arranged in sampling order to form a synchronized and aligned data stream and stored in a digital format for subsequent processing, such as as a two-dimensional array or data file.

[0033] In high-pulse repetition frequency applications, band-limited sampling is prone to aliasing distortion due to the relationship between the sampling rate and the pulse repetition frequency, causing the power monitoring data to deviate from the true value and affecting the stability of the laser system. The jitter injection module injects a small amount of random temporal jitter to disrupt the fixed coupling between the sampling time and the pulse time, mitigating the impact of aliasing distortion at the source and providing a cleaner and more reliable data foundation for subsequent processing steps.

[0034] The jitter injection module captures the power sampling sequence by injecting a small amount of random time jitter into the power monitoring branch and recording the timestamp sequence, generating a synchronized data stream. This synchronized data stream breaks the fixed coupling between the sampling instant and the pulse rhythm, providing a foundation for mitigating aliasing distortion in subsequent processing. The feature evaluation module receives the synchronized data stream, analyzes its characteristics, separates the return energy component, extracts the sampling and pulse coupling characteristics, and generates a coupling strength indicator to guide subsequent sampling frequency adjustments and ensure process stability.

[0035] The feature assessment module includes the following: S2.1, input the synchronized aligned data stream into the sparse reconstruction model: The synchronously aligned data stream contains the actual sampling time and the corresponding instantaneous power value. The synchronously aligned data stream is first input into the sparse reconstruction model to decompose the power sampling sequence into the real power component and the return energy component.

[0036] The sparse reconstruction model uses a dictionary learning approach to represent the power sampling sequence using a sparse combination of a set of basis functions. Basis functions are divided into two categories: the first category represents low-frequency variations in the true power envelope, while the second category represents high-frequency ripples caused by aliasing distortion. The decomposition process is achieved by optimizing an objective function designed to separate the folded-back energy components. A sparsity penalty parameter is used to control the sparsity of the decomposition results, ensuring accurate separation of the true power and folded-back energy components.

[0037] For example, the implementation can be as follows: The input data is the synchronous aligned data stream generated by the jitter injection module ,in is the actual sampling time, is the corresponding instantaneous power value, is the sampling point number. The sparse reconstruction method based on dictionary learning is used to convert the power sampling sequence Decompose into a sparse combination of basis functions. The basis functions include the low-frequency components of the true power envelope and the high-frequency components caused by foldback. The foldback energy components are separated by optimizing the following objective function: ; : True power low-frequency basis function, representing the slowly varying component of the power envelope.

[0038] : Folded high-frequency basis function, representing the high-frequency ripple caused by alias distortion.

[0039] : Sparse coefficient of the real power component.

[0040] : Sparsity coefficient of the reentrant energy component.

[0041] : Sparsity penalty parameter that controls the sparsity of the decomposition.

[0042] After optimization, the return energy component is , the real power component is .

[0043] The sparse reconstruction model effectively separates the different frequency components in a power sampling sequence, making it suitable for processing signals containing high-frequency foldback distortion. By separating the foldback energy component from the true power component, the impact of alias distortion can be more accurately identified and quantified. This separation provides a reliable data foundation for subsequent feature extraction, improving analysis accuracy.

[0044] S2.2-1, Extract sampling and pulse coupling characteristics - reentry energy amplitude: After separating the reentrant energy component, the next step is to extract the reentrant energy rate amplitude from the sampling and pulse coupling characteristics to quantify the energy intensity of the reentrant energy component. First, a fast Fourier transform is performed on the reentrant energy component to obtain a frequency domain spectrum. Next, an adaptive frequency band selection method based on spectral entropy is used to dynamically determine the reentrant frequency band. The specific process is as follows: the normalized power density of the frequency domain spectrum is calculated, and then the spectral entropy is calculated based on the normalized power density to assess the uniformity of the frequency domain spectral distribution. The rate of change of the spectral entropy is used to detect significant change points in the frequency domain spectral distribution and determine the boundaries of the reentrant frequency band. Finally, the frequency domain spectral energy is accumulated within the adaptively determined reentrant frequency band to obtain the reentrant energy rate amplitude.

[0045] For example, the foldback energy amplitude can be obtained as follows: Definition: Reentrant energy amplitude Quantifying the reentrant energy component The energy intensity reflects the influence of alias distortion.

[0046] Calculation method: Reentrant energy component Perform a fast Fourier transform (FFT) to obtain the frequency domain spectrum .

[0047] An adaptive frequency band selection method based on spectral entropy is introduced to dynamically determine the foldback frequency band.

[0048] Specific steps: 1. Calculate spectral power density: Compute the power spectral density and normalize it to a probability distribution: ; : Normalized power spectral density, which represents the probability distribution of frequency domain energy.

[0049] 2. Calculate spectral entropy: based on , calculate the spectral entropy : ; : Spectral entropy, reflecting the uniformity of spectrum distribution.

[0050] 3. Calculate the rate of change of spectral entropy: By frequency Derivative to get the rate of change , in practice, the difference approximation is used: ; : The rate of change of spectral entropy, used to detect significant changes in spectrum distribution.

[0051] 4. Determine the return frequency band boundary: Find The local extreme point of position) and confirm the power spectrum There are significant peaks near these points, and the foldback frequency band is defined as .

[0052] : Adaptively determined foldback frequency band.

[0053] 5. Verify and adjust the frequency band: calculate the energy ratio within the frequency band If the percentage is lower than the threshold (e.g. 80%), expand the frequency band and repeat steps 3 and 4 until the requirement is met.

[0054] : The total energy within the foldback frequency band.

[0055] Calculate the foldback energy magnitude: Accumulate the spectrum energy within the adaptively determined frequency band: ; : The reentry energy amplitude, which indicates the intensity of the reentry energy.

[0056] The foldback energy amplitude is a parameter that quantifies the pseudo-power fluctuation energy caused by alias distortion in high-power fiber laser systems. When the sampling rate and the pulse repetition frequency are close to an integer multiple, the power pulsation in the high-frequency region will be folded back to the low-frequency region, forming alias distortion. The size of this parameter directly reflects the severity of the alias distortion: when the amplitude is large, more energy is folded back to the low-frequency region, and the alias distortion is more significant, which may cause the feedback system to misjudge power changes, trigger power oscillations and affect output stability; when the amplitude is small, the foldback energy is small, the alias distortion has a smaller impact, and the system can more accurately monitor and control the power output, thereby maintaining the stability of the processing process. Therefore, the foldback energy amplitude is an important indicator for evaluating the degree of system sampling and pulse coupling, and its size is closely related to the stability of the system and the processing quality.

[0057] S2.2-2, Extracting Sampling and Pulse Coupling Characteristics - Jitter Dispersion Rate: Another step in extracting sampling and pulse coupling characteristics is calculating the jitter dispersion ratio, which quantifies the degree of dispersion of random temporal jitter with respect to the sampling time distribution. This process first calculates the deviation between the actual sampling time and the nominal sampling time, and then analyzes it using the mutual information entropy method. Specifically, the kernel density estimation method is used to obtain the probability density function of the deviation. The entropy value is then calculated based on the probability density function, which is defined as the jitter dispersion ratio.

[0058] The jitter dispersion ratio is a parameter that measures the impact of random time jitter on the discreteness of the sampling moment distribution in high-power fiber laser systems. By injecting a small amount of random time jitter, the fixed relationship between the sampling moment and the pulse moment can be broken, thereby reducing the impact of alias distortion. The magnitude of this parameter reflects the randomness and effectiveness of the random time jitter: when the dispersion ratio is large, the deviation distribution of the sampling moment is more dispersed and more random, which can effectively decouple the fixed relationship between sampling and pulse and significantly reduce alias distortion. When the dispersion ratio is small, the deviation distribution of the sampling moment is more concentrated and lacks randomness. It may not be able to fully eliminate the coupling between sampling and pulse, resulting in the continued existence of alias distortion. By quantifying the effect of jitter injection, the jitter dispersion ratio provides a basis for stable power monitoring of the system at high pulse repetition frequencies.

[0059] S2.3, perform abstract evaluation (e.g. Figure 2 shown): After extracting the reentry energy amplitude and jitter dispersion, an abstract evaluation is performed to generate a coupling strength indicator. The purpose is to evaluate the degree of coupling between the sampling link and the pulse link based on the reentry energy amplitude and jitter dispersion. The specific processing process includes the following sub-steps: The reentry energy rate amplitude sequence and jitter dispersion rate sequence are collected within a sliding time window to construct a joint probability density field. The optimal transmission algorithm is used to calculate the Wasserstein distance between the joint probability density field and the reference decoupling distribution to obtain the minimum transmission potential energy sequence. Multi-scale empirical mode decomposition is performed on the minimum transmission potential energy sequence to extract representative eigencomponents. The representative eigencomponents are mapped to a zero-mean symmetric domain using a radial normalization kernel. Finally, the information entropy aggregation method is used to aggregate local entropy with adaptive weights to generate a coupling strength flag. When the coupling strength flag is less than 1, a qualified flag is written.

[0060] The size of the coupling strength indicator reflects the degree of coupling between the sampling link and the pulse link in a high-power fiber laser system and its impact on power monitoring. When the coupling strength indicator value is small, it indicates that the sampling link and the pulse link tend to be decoupled, the impact of alias distortion is small, and the system can more accurately monitor the power output, thereby maintaining the stability of the processing process. Conversely, when the coupling strength indicator value is large, the degree of coupling between the sampling link and the pulse link is enhanced, and the impact of alias distortion is aggravated, which may cause the feedback system to misjudge power changes, trigger power oscillations, and reduce output stability. Therefore, the size of the coupling strength indicator is an important indicator for evaluating the accuracy and stability of the system's power monitoring at high pulse repetition frequencies. The smaller the value, the better the system performance.

[0061] The abstract assessment quantifies the coupling relationship between the reentrant energy rate amplitude and the jitter dispersion rate using a joint probability density field and an optimal transmission algorithm. The Wasserstein distance effectively measures the difference between the joint probability density field and the reference decoupling distribution, accurately reflecting the coupling strength. Multi-scale empirical mode decomposition extracts representative eigencomponents, combined with information entropy aggregation methods, further improving the robustness and accuracy of the assessment. This multi-level analysis ensures the reliability of the coupling strength identification and provides a scientific basis for subsequent sampling frequency adjustments.

[0062] The feature evaluation module separates the reentrant energy component by inputting the synchronously aligned data stream into a sparse reconstruction model, accurately identifying and quantifying pseudo-power fluctuations caused by alias distortion. By extracting the reentrant energy amplitude and jitter dispersion, it quantifies the impact of alias distortion and the characteristics of random temporal jitter, respectively, comprehensively evaluating the coupling state between sampling and pulses. The abstract evaluation component utilizes a joint probability density field, optimal transmission algorithm, multi-scale empirical mode decomposition, and information entropy aggregation to generate a coupling strength signature, accurately characterizing the degree of decoupling between the sampling and pulse links. The feature evaluation module significantly improves the performance of high-power fiber laser systems under high-speed conditions. Separating the reentrant energy component eliminates the interference of pseudo-power fluctuations on monitoring. Coupling feature extraction provides a clear basis for sampling frequency optimization. The coupling strength signature generated by the abstract evaluation directly guides frequency adjustment, enabling the system to achieve stable power output and monitoring at high pulse repetition rates, ensuring high consistency and reliability of the processing process. In particular, the abstract evaluation's technical features, through multi-dimensional analysis and entropy aggregation, accurately capture coupling dynamics, significantly reducing the risk of power distortion caused by insufficient sampling, thereby optimizing overall system stability.

[0063] The feature evaluation module utilizes a sparse reconstruction model to separate the reentrant energy components, extract coupling features, and generate a qualified signature. However, in high-speed operating conditions with dynamically changing pulse repetition frequencies, relying solely on these measures is insufficient to completely eliminate alias distortion. Real-time optimization and adjustment of the sampling frequency becomes crucial for maintaining monitoring link stability and processing consistency. The frequency adjustment module aims to achieve continuous decoupling of the sampling link from the pulse link by dynamically adjusting the sampling frequency to a fractional ratio coprime with the pulse repetition frequency, thereby eliminating alias distortion at its root.

[0064] The frequency adjustment module includes the following: S3.1, read the qualified mark: Obtain a pass flag from the feature evaluation module to confirm whether the coupling strength between the sampling frequency and the pulse repetition frequency meets the adjustment conditions. If the pass flag is "passed," meaning the coupling strength flag is less than 1, the sampling frequency adjustment process begins. If the pass flag is not obtained, or the coupling strength flag is greater than or equal to 1, the current sampling frequency remains unchanged and no further adjustments are performed.

[0065] The presence of a qualified identifier is a necessary condition for triggering the sampling frequency adjustment process. The reason is to ensure that the sampling frequency adjustment is only performed when the coupling degree between the sampling link and the pulse link reaches an acceptable range, thereby avoiding invalid or unnecessary frequency changes. In the programmable waveform modulated high-power fiber laser system, the feature evaluation module generates a coupling strength identifier by analyzing the synchronous alignment data stream. When the coupling strength identifier is less than 1, the qualified identifier is written, indicating that the impact of alias distortion has been significantly reduced and the degree of coupling between the sampling frequency and the pulse repetition frequency is insufficient to cause serious power foldback pseudo-fluctuations. At this point, the qualified identifier indicates that the system is qualified to enter the sampling frequency adjustment process. By adjusting the sampling frequency to a fractional ratio that is coprime with the pulse repetition frequency, the potential integer multiple relationship is further eliminated, and the occurrence of alias distortion is eliminated from the root. This mechanism not only improves the targetedness and efficiency of the adjustment, but also ensures the power monitoring accuracy and processing stability of the system at high pulse repetition frequencies.

[0066] S3.2, calculate the coprime fraction ratio sampling frequency: In order to avoid the sampling frequency and the pulse repetition frequency forming an integer multiple relationship, the sampling frequency that is coprime with the current pulse repetition frequency is calculated. The specific calculation process is as follows: First, obtain the current pulse repetition frequency and the current sampling frequency, and then determine the greatest common divisor of the two. If the greatest common divisor is 1, it means that the current sampling frequency is coprime with the pulse repetition frequency and no further adjustment is required. If the greatest common divisor is greater than 1, the sampling frequency needs to be adjusted. To do this, set an adjustment step value, such as 10 kHz, and initially set the target sampling frequency to the current sampling frequency plus the adjustment step value. Next, iteratively adjust the target sampling frequency, each time increasing or decreasing the target sampling frequency by the adjustment step value, and recalculating the greatest common divisor of the target sampling frequency and the pulse repetition frequency until the greatest common divisor is 1. At the same time, it is necessary to ensure that the target sampling frequency is within the hardware sampling rate range to ensure that the adjusted sampling frequency is achievable within the system hardware capabilities.

[0067] By calculating a sampling frequency that is coprime with the pulse repetition frequency (PRF), ensuring that the two do not share a common divisor greater than 1, we fundamentally eliminate aliasing distortion caused by integer frequency multiples. We employ an iterative adjustment method to calculate the target sampling frequency. This method is simple and computationally efficient, enabling rapid determination of the appropriate sampling frequency value. Furthermore, by accounting for the hardware sampling rate range limitations, we ensure the feasibility and practicality of the adjustment results, avoiding ineffective adjustments that exceed hardware capabilities.

[0068] S3.3, adjust the real-time sampling frequency: Based on the calculated target sampling frequency, a frequency adjustment instruction is generated. This instruction contains the specific value of the target sampling frequency and is sent to the sampling clock generator. Upon receiving the instruction, the sampling clock generator adjusts its internal clock frequency so that the real-time sampling frequency precisely matches the target sampling frequency, thus completing the real-time update of the sampling frequency.

[0069] Adjusting the real-time sampling frequency to a target value coprime with the pulse repetition frequency ensures that the sampling sequence does not produce foldback artifacts due to integer multiples of the frequency. This adjustment significantly improves the accuracy of power monitoring, preventing high-frequency power pulsations from being mistakenly folded back into the low-frequency region, thereby enhancing the measurement accuracy and reliability of the system.

[0070] S3.4, synchronously update the sampling frequency label: The adjusted sampling frequency is written digitally to the sampling frequency tag and stored in a register. The stored sampling frequency tag can be directly read and used in subsequent steps, ensuring that all parts of the system can operate based on the latest sampling frequency data.

[0071] The frequency adjustment module reads the qualified identifier, calculates, and adjusts the real-time sampling frequency to a fractional ratio coprime with the pulse repetition frequency, completely eliminating aliasing distortion. This adjusted sampling frequency ensures that the sampling sequence does not fold back high-frequency power pulsations to the low-frequency region due to integer multiples of the frequency, significantly improving power monitoring accuracy. Furthermore, the synchronously updated sampling frequency tag provides a precise frequency reference for subsequent closed-loop power feedback, enabling the system to optimize the pump drive signal based on the latest sampling conditions, ultimately achieving stable output power maintenance.

[0072] In the frequency adjustment module, the programmable waveform modulated high-power fiber laser system reads the qualified identifier, adjusts the sampling frequency to a fractional ratio that is coprime with the pulse repetition frequency, and simultaneously updates the sampling frequency label, thereby eliminating the root cause of alias distortion. However, simply adjusting the sampling frequency cannot completely eliminate the impact of fluctuations in power monitoring. To further ensure the stability of the output power, the power stabilization module introduces a real-time correction mechanism for the pump drive signal. Using the sampling frequency label and qualified identifier, the pump drive signal is precisely adjusted by setting the feedforward gain matrix, ultimately achieving the goal of maintaining the output power aftershock within the set upper limit.

[0073] The power stabilization module includes the following: S4.1, read the sampling frequency label and qualified mark: The sampling frequency tag is read from the register to obtain the specific value of the sampling frequency adjusted by the frequency adjustment module. At the same time, the qualified flag written by the characteristic evaluation module is read to confirm whether the degree of coupling between the sampling frequency and the pulse repetition frequency meets the preset requirements, that is, to determine whether the coupling characteristic index is less than 1.

[0074] S4.2, set the feedforward gain matrix: The feedforward gain matrix is ​​a matrix of gain coefficients used to adjust the pump drive signal. Its dimensions are determined by the number of control channels of the pump source. The process of setting the feedforward gain matrix includes the following steps: First, determine whether the feedforward gain matrix needs to be adjusted based on the qualified mark; if the qualified mark indicates that the degree of coupling between the sampling frequency and the pulse repetition frequency meets the requirements, that is, the coupling characteristic index is less than 1, then calculate the ratio of the current sampling frequency to the pulse repetition frequency; next, compare the calculated ratio with a preset reference ratio to obtain the difference between the two; then, combine the preset adjustment coefficient and adjust the reference feedforward gain matrix according to the above difference to generate a new feedforward gain matrix to compensate for the dynamic impact of the sampling frequency change on power monitoring.

[0075] The purpose of setting the feedforward gain matrix is ​​to compensate for the interference of sampling frequency changes on power monitoring by dynamically adjusting the gain of the pump drive signal. This method ensures that the output power of the pump source remains consistent with the target power level. The feedforward gain matrix is ​​adjusted based on the current system state, enabling rapid response to sampling frequency changes, improving system control stability and response speed, and reducing power fluctuations that may be caused by feedback delays.

[0076] S4.3, apply correction to the pump drive signal: The original pump drive signal is modified based on the specified feedforward gain matrix to generate a modified pump drive signal. Specifically, the feedforward gain matrix is ​​applied to the original pump drive signal, adjusting the signal's amplitude or phase to generate a modified pump drive signal that accounts for the effects of sampling frequency variations. The modified pump drive signal is then sent to the pump source controller, which adjusts the pump source's output power to match the target power level.

[0077] The purpose of implementing pump drive signal correction is to quickly compensate for power fluctuations caused by sampling frequency changes by adjusting the pump source drive signal in real time. This correction mechanism effectively suppresses power instabilities caused by sampling frequency adjustments, improves the system's power control accuracy at high pulse repetition rates, and ensures stable laser output.

[0078] The pump source is the energy supply unit of the laser, typically composed of a high-power semiconductor laser (LD) or fiber laser. Its primary function is to provide excitation energy to the gain medium (usually a doped fiber) to achieve laser amplification and generation. Specifically, the pump source converts electrical energy into light energy and injects it into the doped fiber, exciting rare earth ions (such as erbium and ytterbium) to high energy levels. Subsequently, laser output is generated through a process called stimulated emission. In high-power fiber lasers, the performance of the pump source directly affects the efficiency and output power of the laser system.

[0079] The pump source controller is a device that precisely manages the pump source, responsible for regulating key parameters such as its output power, wavelength, and temperature to ensure the stability and reliability of the laser system. Its main functions include the following aspects: Power control: By adjusting the current or voltage of the pump source, its output power can be precisely controlled, thereby indirectly adjusting the output power of the laser.

[0080] Temperature management: Real-time monitoring and adjustment of the operating temperature of the pump source to avoid damage to the life and performance of the equipment due to overheating.

[0081] Wavelength stabilization: By fine-tuning the temperature or current, the wavelength of the pump source is matched to the absorption peak of the gain medium, improving energy conversion efficiency.

[0082] Safety protection: Integrated over-current, over-voltage, over-temperature and other protection mechanisms ensure the safe operation of the pump source and the entire system.

[0083] In high-power fiber lasers, the pump source provides the energy foundation, while the pump source controller ensures system stability and output consistency through real-time regulation. Especially in programmable waveform modulation scenarios, the pump source controller dynamically adjusts the pump source output based on feedback signals to maintain laser power stability and beam quality, meeting the requirements of high-precision applications. Both are indispensable and together determine the performance of the laser system.

[0084] S4.4, maintain the output power aftershock within the set upper limit: The laser's real-time output power is continuously monitored and compared with the preset target power value. The difference between the two is calculated, representing the output power aftershock. If the absolute value of the output power aftershock exceeds a preset upper limit, a proportional-integral-differential control algorithm is used to gradually adjust the feedforward gain matrix to reduce the absolute value of the output power aftershock until it is less than or equal to the preset upper limit. The specific adjustment process involves determining the adjustment direction and amplitude based on the magnitude and trend of the output power aftershock, and iteratively updating the feedforward gain matrix until the aftershock control requirements are met.

[0085] Maintaining output power aftershock within a set upper limit involves real-time monitoring and adjustment of the output power through a closed-loop feedback mechanism, preventing excessive aftershock from causing amplification of stimulated scattering or thermal lensing. This approach effectively maintains beam quality stability and uniform energy distribution, ensuring reliable optical performance for high-precision machining.

[0086] The power stabilization module reads the sampling frequency label and qualified identification, sets the feedforward gain matrix, implements real-time corrections to the pump drive signal, and utilizes a closed-loop feedback mechanism to ensure that the output power aftershock remains within the set upper limit. This technical logic significantly improves the power control accuracy of programmable waveform-modulated high-power fiber lasers at high pulse repetition rates, effectively suppresses the effects of aliasing distortion and power fluctuations, and provides stable beam quality and energy distribution for high-precision machining.

[0087] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0088] It should be noted that the system of the present invention can be deployed on the device itself to realize embedded applications, and can also be run on a PC or other terminal with a user interface, thereby meeting a variety of hardware environments and usage requirements.

[0089] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

[0090] It should be noted that, in this document, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or device. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0091] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A programmable waveform modulated high power fiber laser system, characterized in that: include: Jitter injection module, signature evaluation module, frequency adjustment module and power stabilization module; Jitter injection module: injects slight random time jitter into the power monitoring branch and records the timestamp sequence. It simultaneously captures the power sampling sequence and outputs a synchronously aligned data stream. Feature evaluation module: This module inputs the synchronously aligned data stream into the sparse reconstruction model to separate the return energy component, extracts the sampling and pulse coupling features, performs abstract evaluation, determines the degree of decoupling between the sampling link and the pulse link, and obtains a qualified mark. Frequency adjustment module: reads the qualified mark, adjusts the real-time sampling frequency to a fractional ratio that is coprime with the pulse repetition frequency, and simultaneously updates the sampling frequency label; Power stabilization module: This module sets the feedforward gain matrix based on the sampling frequency label and the pass indicator, and corrects the pump drive signal to keep the output power aftershock within the set upper limit.

2. A programmable waveform modulated high power fiber laser system according to claim 1, characterized in that: The jitter injection module includes the following: In the power monitoring branch, an independent slight random time offset is introduced for each sampling period to disrupt the fixed correspondence between the sampling moment and the pulse moment. The slight random time offset is generated through uniform distribution. At the same time, the actual sampling moment of each sampling point is recorded. The actual sampling moment is the sum of the nominal sampling moment and the slight random time offset. The nominal sampling moment is the product of the sampling point number and the sampling period. Furthermore, at each actual sampling moment, a high-speed analog-to-digital converter is used to collect the instantaneous value of the laser output power to form a power sampling sequence. Finally, the actual sampling moment is paired with the corresponding instantaneous power value to generate a synchronously aligned data stream.

3. The programmable waveform modulated high power fiber laser system according to claim 2, characterized in that: The feature assessment module includes the following: The power sampling sequence in the synchronous aligned data stream is sparsely decomposed using a dictionary learning method to separate the true power component and the retrieval energy component, from which the sampling and pulse coupling characteristics are extracted, including the retrieval energy amplitude and jitter dispersion rate.

4. The programmable waveform modulated high power fiber laser system according to claim 3, characterized in that: The feature evaluation module also includes the following: By collecting the reentry energy rate amplitude sequence and the jitter dispersion rate sequence within a sliding time window, a joint probability density field is constructed. The Wasserstein distance with the reference decoupling distribution is calculated using the optimal transmission algorithm to obtain the minimum transmission potential energy sequence. Multi-scale empirical mode decomposition is then performed on it to extract representative eigencomponents. After radial normalized kernel mapping, adaptive weighted aggregation of local entropy is used to generate a coupling strength label. The degree of decoupling between the sampling link and the pulse link is judged to obtain a qualified label.

5. The programmable waveform modulated high power fiber laser system according to claim 3, characterized in that: The feature evaluation module also includes the following: The impact of alias distortion is quantified by calculating the amplitude of the folded energy rate by performing fast Fourier transform on the folded energy component and combining it with an adaptive frequency band selection method based on spectral entropy.

6. The programmable waveform modulated high power fiber laser system according to claim 3, characterized in that: The feature evaluation module also includes the following: The deviation between the actual sampling time and the nominal sampling time is calculated, and the mutual information entropy method is used to calculate the jitter dispersion rate and quantify the randomness of random time jitter.

7. The programmable waveform modulated high power fiber laser system according to claim 4, characterized in that: The frequency adjustment module includes the following: When the qualified mark is qualified, the sampling frequency adjustment process is started; the current pulse repetition frequency and the current sampling frequency are obtained, and the greatest common divisor of the current pulse repetition frequency and the current sampling frequency is calculated. If the greatest common divisor is 1, the current sampling frequency is kept unchanged; If the greatest common divisor is greater than 1, set the adjustment step value and iteratively adjust the target sampling frequency until the greatest common divisor of the target sampling frequency and the pulse repetition frequency is 1, while ensuring that the target sampling frequency is within the hardware sampling rate range.

8. The programmable waveform modulated high power fiber laser system according to claim 7, characterized in that: The frequency adjustment module also includes the following: A frequency adjustment instruction including a target sampling frequency is generated and sent to a sampling clock generator; the sampling clock generator adjusts the internal clock frequency according to the frequency adjustment instruction so that the real-time sampling frequency matches the target sampling frequency; the adjusted sampling frequency is written into a sampling frequency tag and stored in a register.

9. The programmable waveform modulated high power fiber laser system according to claim 8, characterized in that: The power stabilization module includes the following: Read the sampling frequency label and the qualified flag from the register to obtain the adjusted sampling frequency and the coupling status of the sampling frequency and the pulse repetition frequency; When the qualified mark indicates that the coupling characteristic index is less than 1, the reference feedforward gain matrix is ​​adjusted according to the ratio of the sampling frequency to the pulse repetition frequency, combined with the preset reference ratio and the adjustment coefficient, to generate a new feedforward gain matrix.

10. The programmable waveform modulated high power fiber laser system according to claim 9, characterized in that: The power stabilization module also includes the following: The original pump driving signal is corrected using the new feedforward gain matrix to generate a corrected pump driving signal, which is sent to the pump source controller to adjust the output power of the pump source; The real-time value of the laser output power is continuously monitored, and the difference between it and the target power value is calculated as the output power aftershock. When the absolute value of the output power aftershock exceeds the preset upper limit, the proportional-integral-differential control algorithm is used to dynamically adjust the feedforward gain matrix until the absolute value of the output power aftershock is less than or equal to the preset upper limit.

Images