Method and system for rapidly predicting impulse in laser cleared fragments
By constructing an impulse prediction model based on machine learning and utilizing plasma spectral information and a multi-camera system, rapid impulse prediction for laser removal of space debris is achieved, solving the problems of complex and slow impulse calculation in existing technologies and improving the hit rate and efficiency of laser removal.
Patent Information
- Application Number
- CN202510503534.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-21
- Publication Date
- 2025-09-09
AI Technical Summary
When using lasers to remove space debris, existing technologies have complex and slow impulse calculations, making it impossible to achieve real-time adjustment and control, which affects the hit rate and resource utilization efficiency.
By acquiring the impulse data and spectral intensity data of laser ablation of space debris targets, an impulse prediction model based on machine learning is constructed. The plasma spectrum information is used for rapid prediction. By combining the support vector machine and BP neural network methods, the model parameters and spectral line combinations are optimized to achieve real-time prediction of impulse.
It realizes the rapid impulse prediction of laser removal of space debris, improves the hit rate and operation efficiency, simplifies the device structure, meets the requirements of real-time and accuracy, and is suitable for various target types.
Smart Images
Figure CN120611207A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to a method and system for predicting impulse in laser micro-propulsion, belonging to the field of laser micro-propulsion, and in particular to a method and system for rapidly predicting impulse in laser debris removal. Background Art
[0002] With humanity's continued exploration of space, the amount of space debris has increased dramatically. These debris, constantly orbiting Earth at high speeds, have created a unique debris environment around Earth, leading to a serious space debris problem. If a spacecraft strikes space debris during its operation, it can alter the surface properties of the orbiting spacecraft at best, or even cause it to explode. Therefore, these debris threatens the safe operation of existing spacecraft and even poses a significant risk to future space missions. Laser debris removal has attracted widespread attention due to its non-contact operation, high efficiency, zero pollution, high continuity, and great flexibility. Laser debris removal works by irradiating the debris with a high-energy laser beam, causing the surface material to evaporate instantly or generate a plasma jet, which in turn exerts a reaction force. This reaction force changes the debris's momentum, generating an impulse. This impulse reduces the debris's orbital velocity and lowers its perigee, thereby altering its orbit. This allows the irradiated debris to enter the atmosphere and burn up, or to be transferred to an orbit where it does not affect other spacecraft.
[0003] During the entire laser debris removal process, impulse calculation is crucial to the debris' trajectory change. Accurately calculating impulse is crucial for the success of laser debris removal technology. However, current technology for measuring impulse changes requires real-time data on the debris' position, velocity, and attitude to adjust laser parameters and calculate the required impulse. This entire process is time-consuming and slow. Summary of the Invention
[0004] According to one aspect of the present application, a method for rapidly predicting impulse in laser debris removal is provided. This method can achieve rapid prediction of impulse, which is of great significance for real-time adjustment and control of laser space debris removal, improving hit rate and effectiveness, and optimizing resource utilization.
[0005] The method for rapidly predicting the impulse in laser debris removal comprises the following steps:
[0006] (1) Simultaneously obtain impulse data and spectral intensity data when laser ablation is performed on a target material of space debris; obtain multiple sets of impulse data and spectral intensity data by continuously changing the laser parameters emitted by the laser light source;
[0007] (2) preprocessing the data obtained in step (1) to form a valid data set;
[0008] (3) constructing an impulse prediction model based on machine learning, the output of which is an impulse prediction value; and calculating the optimal model parameter combination and the optimal spectral line prediction combination of the model based on the valid data set;
[0009] (4) During the laser debris removal process, the plasma spectrum information of the target debris target material is obtained in real time and input into the trained impulse prediction model for rapid impulse prediction; the plasma spectrum information includes spectral line intensity information of two or more specific wavelengths for the target material type.
[0010] Optionally, the target material needs to be pre-treated before executing step (1) to remove and prevent the oxide film from forming again.
[0011] Preferably, the preprocessing specifically includes:
[0012] Use dilute acid to dissolve the oxide film on the target, then rinse thoroughly with deionized water; then polish the target surface; dry it as soon as possible after cleaning.
[0013] Optionally, the laser parameters in step (1) include laser energy parameters.
[0014] Optionally, the preprocessing includes but is not limited to at least one of the following: processing of missing values, processing of outliers, normalization of data, and alignment of spectral line intensity graphs with characteristic spectrum wavelengths in the NIST database.
[0015] Optionally, step (3) includes:
[0016] Support vector machine (SVM) and BP neural network methods were selected; the optimal model parameter combination of the impulse prediction model and the optimal spectral line prediction combination were obtained through ergodic search.
[0017] Optionally, in step (4), the specific wavelength of the plasma spectrum information is selected with reference to the optimal spectral line prediction combination.
[0018] According to another aspect of the present application, a system for rapidly predicting impulse in laser debris removal is provided, comprising:
[0019] The sample data acquisition unit is used to obtain impulse data and spectral intensity data when laser ablation of space debris targets; by continuously changing the laser parameters emitted by the laser light source, multiple sets of impulse data and spectral intensity data are obtained;
[0020] A target data acquisition unit, configured to obtain plasma spectrum information of the target fragment target material, wherein the plasma spectrum information includes spectral line intensity information of two or more specific wavelengths for the target material type;
[0021] A data processing unit is configured to preprocess the multiple sets of impulse data and spectral intensity data to form a valid data set; construct an impulse prediction model based on machine learning, the output of which is an impulse prediction value; calculate the optimal model parameter combination and the optimal spectral line prediction combination of the model based on the valid data set; obtain plasma spectral information and input it into the trained impulse prediction model to quickly predict the impulse.
[0022] Optionally, the sample data acquisition unit uses a torsion pendulum impulse measurement device, a plasma spectrum measurement device and a laser ablation device; the lasers in the torsion pendulum impulse measurement device, the plasma spectrum measurement device and the laser ablation device are synchronized;
[0023] The laser ablation device is used to provide lasers with variable energy, pulse width, and frequency, and to provide different ablation conditions;
[0024] The torsion pendulum impulse measuring device is used to measure the average displacement of the torsion pendulum on which the target material is mounted during laser irradiation, and calculate the impulse of the laser ablation target material;
[0025] The plasma spectrum measuring device is used to obtain spectrum intensity data under different ablation conditions.
[0026] Optionally, the variable energy is adjustable within the range of 0-100 mJ.
[0027] Optionally, the pulse width is in the order of nanoseconds.
[0028] Optionally, the frequency is adjustable within the range of 1-10 Hz.
[0029] Considering that in actual engineering applications, when a microthruster is operating in space, it is impossible to use a spectrometer to collect spectra, as the spectrometer is a precision scientific research instrument. Therefore, the target data acquisition unit optionally uses an impulse measurement device based on a multi-camera system to collect two or more plasma emission spectra and then predict the impulse of the microthruster based on a pre-trained impulse prediction model.
[0030] Optionally, the impulse measurement device of the multi-camera system adopts a dual-camera spectrum acquisition device, including two CCD detectors and corresponding wave plate wheels and bandpass filter groups; wherein,
[0031] The two CCD detectors are used to synchronously collect plasma signals of two specific wavelengths and send spectral intensity data to a data processing unit;
[0032] The wave plate wheel and bandpass filter assembly are used to switch appropriate filters according to the selected spectrum lines of specific wavelengths according to the target material type of space debris.
[0033] Optionally, the specific wavelength is selected with reference to the best spectral line prediction combination calculated by the data processing unit.
[0034] Optionally, the range of the bandpass filter is dynamically adjusted according to the effective characteristic spectrum of the sample.
[0035] The beneficial effects of this application include:
[0036] 1) The method and system for rapidly predicting impulse during laser debris removal, provided in this application, can directly provide specific spectral intensity and impulse information generated by laser ablation of the target material, serving as a basis for accurate laser debris removal. This effectively addresses shortcomings in laser space debris removal technology, such as complex impulse measurement systems, slow and time-consuming impulse data acquisition, and the lack of a real-time feedback mechanism. By directly using the spectral information generated by the laser ablation target material in combination with a machine learning algorithm to predict impulse size, this application offers the advantages of clear principles, simple apparatus, and strong operability compared to traditional impulse acquisition methods.
[0037] 2) This application uses multi-dimensional plasma spectral information, and then can fully utilize the impulse prediction model based on machine learning, without relying on physical effect derivation and linear calibration, to achieve data-driven based on nonlinear models.
[0038] 3) The dual-camera system of this application is also equipped with a bandpass filter set to accurately screen the target material's unique characteristic spectral lines (such as 394.40nm / 463.15nm for aluminum targets), and can directly collect the target material's characteristic spectral line intensity for impulse prediction. The wave plate wheel automatically switches between multiple sets of filters without manual switching, and can adapt to different target materials within 1ms, ensuring real-time response, thereby achieving rapid impulse prediction. Rapid impulse prediction allows the laser system to dynamically adjust the power, irradiation angle, and duration according to actual conditions, thereby improving the accuracy and efficiency of the operation. This is crucial for ensuring that the laser beam can continuously and effectively apply the required force to change the debris trajectory and improve debris removal efficiency.
[0039] 4) This application can achieve end-to-end rapid prediction: the target data acquisition unit can use an impulse measurement device based on a multi-camera system. The multi-camera system is triggered synchronously with the laser pulse through a synchronization device to achieve microsecond synchronization, ensuring that the measured multiple spectral line intensity data strictly correspond to the same ablation event, avoiding the prediction deviation introduced by asynchronous acquisition. Although this application can support multi-camera to achieve multi-dimensional input, experiments have shown that two-dimensional input is accurate enough, and considering actual engineering applications, two or three spectral lines are the most likely to achieve engineering applications. Therefore, this application can maintain low system complexity while ensuring accuracy, and can adapt to the control period of nanosecond pulse lasers to meet the real-time requirements of laser pulse frequency (1-10Hz). BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 This is a schematic flow chart of a method and system for rapidly predicting impulse in laser debris removal in one embodiment of the present application.
[0041] Figure 2 Schematic diagram of a torsion pendulum measurement device and a plasma spectrum measurement device in one embodiment of the present application.
[0042] Figure 3 This is a schematic diagram of the operation of the target data acquisition unit in one embodiment of the present application.
[0043] Figure 4 This is a schematic diagram of the connection between the target data acquisition unit and the data processing unit in one embodiment of the present application.
[0044] Figure 5 This is the plasma spectrum under the conditions of laser irradiation of aluminum target in the preferred example of this application.
[0045] Figure 6 This is a graph showing the changes in laser energy, spectral intensity and impulse in a preferred example of this application.
[0046] Figure 7 This is a comparison of the prediction accuracy of the SVM model and the BP neural network model in the preferred example of this application.
[0047] Figure 8 This is a comparison between the SVM model predicted data and the actual data in the preferred example of this application.
[0048] List of parts and reference numerals:
[0049] 1. Torsional pendulum impulse measurement device; 2. Plasma spectrum measurement device; 3. Laser ablation device; 4. Spectral acquisition device based on dual cameras; 5. Back-end data processing system;
[0050] 101. Pendulum beam; 102. Pendulum bracket; 103. Pendulum pivot; 104. Counterweight; 105. Displacement sensor;
[0051] 201. Spectrometer; 202. ICCD; 203. Lens; 204. Optical fiber;
[0052] 301. Laser; 302. Half-wave plate; 303. Glan prism; 304. Focusing lens; 305. Translation stage; 306. Synchronization device;
[0053] 401. Two CCD detectors; 402. Two wave plate wheels; 403. Bandpass filter set; 501. Computer; 502. Data processing software. DETAILED DESCRIPTION
[0054] The present application is described in detail below with reference to embodiments, but the present application is not limited to these embodiments.
[0055] This application provides a method for rapidly predicting impulse in laser debris removal, comprising the following steps:
[0056] (1) Simultaneously obtain impulse data and spectral intensity data when laser ablation is performed on a target material of space debris; obtain multiple sets of impulse data and spectral intensity data by continuously changing the laser parameters emitted by the laser light source;
[0057] (2) preprocessing the data obtained in step (1) to form a valid data set;
[0058] (3) constructing an impulse prediction model based on machine learning, the output of which is an impulse prediction value; and calculating the optimal model parameter combination and the optimal spectral line prediction combination of the model based on the valid data set;
[0059] (4) During the laser debris removal process, the plasma spectrum information of the target debris target material is obtained in real time and input into the trained impulse prediction model for rapid impulse prediction; the plasma spectrum information includes spectral line intensity information of two or more specific wavelengths for the target material type.
[0060] See Figure 1 , which shows the process of a method for rapidly predicting the impulse in laser debris removal in one embodiment of the present application. Wherein, the step (1) may specifically include the following steps:
[0061] (11) Construct a pendulum impulse measurement device, a plasma spectrum measurement device, and a laser ablation device;
[0062] (12) A large amount of impulse data is obtained by using a torsion pendulum impulse measurement device, and simultaneously, a large amount of spectral intensity data is obtained by using a plasma spectrum measurement device.
[0063] In one embodiment, the target material needs to be pre-treated before executing step (1) to remove and prevent the oxide film from forming again to prevent it from affecting the prediction results.
[0064] Furthermore, the target material pretreatment includes: using dilute acid to dissolve the oxide film on the target material, and then thoroughly rinsing it with deionized water; then polishing the target material surface; and drying it as soon as possible after cleaning.
[0065] In one embodiment, the laser parameters in step (1) include laser energy parameters, including but not limited to variable energy, pulse width, and frequency.
[0066] In one embodiment, the preprocessing means in step (2) include but are not limited to: processing of missing values, processing of outliers, normalization of data, and alignment of spectral line intensity graphs with characteristic spectrum wavelengths in the NIST database.
[0067] In one embodiment, step (3) comprises:
[0068] Support vector machine (SVM) and BP neural network methods were selected; the optimal model parameter combination of the impulse prediction model and the optimal spectral line prediction combination were obtained through ergodic search.
[0069] In a specific embodiment, the optimal model parameters can be obtained using a grid search method.
[0070] In one embodiment, in step (4), the selection of the specific wavelength of the plasma spectrum information can refer to the optimal spectral line prediction combination.
[0071] The present application also provides a system for rapidly predicting impulse in laser debris removal, comprising:
[0072] The sample data acquisition unit is used to obtain impulse data and spectral intensity data when laser ablation of space debris targets; by continuously changing the laser parameters emitted by the laser light source, multiple sets of impulse data and spectral intensity data are obtained;
[0073] A target data acquisition unit, configured to obtain plasma spectrum information of the target fragment target material, wherein the plasma spectrum information includes spectral line intensity information of two or more specific wavelengths for the target material type;
[0074] A data processing unit is configured to preprocess the multiple sets of impulse data and spectral intensity data to form a valid data set; construct an impulse prediction model based on machine learning, the output of which is an impulse prediction value; calculate the optimal model parameter combination and the optimal spectral line prediction combination of the model based on the valid data set; obtain plasma spectral information and input it into the trained impulse prediction model to quickly predict the impulse.
[0075] In one embodiment, Figure 1 As shown, the sample data acquisition unit uses a torsion pendulum impulse measurement device 1, a plasma spectrum measurement device 2 and a laser ablation device 3. The lasers 301 in the torsion pendulum impulse measurement device 1, the plasma spectrum measurement device 2 and the laser ablation device 3 are synchronized.
[0076] The torsion pendulum impulse measuring device 1 is used to measure the average displacement of the torsion pendulum on which the target material is mounted during laser irradiation, and calculate the impulse of the laser ablation target material.
[0077] The plasma spectrum measuring device 2 is used to obtain spectrum intensity data under different ablation conditions.
[0078] The laser ablation device 3 is used to provide laser light with variable energy, pulse width and frequency, and to provide different ablation conditions.
[0079] In one embodiment, the variable energy is adjustable within the range of 0-100 mJ.
[0080] In one embodiment, the pulse width is on the order of nanoseconds.
[0081] In one embodiment, the frequency is adjustable within the range of 1-10 Hz.
[0082] In one embodiment, Figure 2 As shown, the torsion pendulum impulse measuring device 1 includes: a torsion pendulum beam 101 supported by a flexible pivot, a torsion pendulum bracket 102, a torsion pendulum pivot 103, a counterweight 104, a displacement sensor 105, and a damper.
[0083] When the laser abslates the target, the thrust generated causes the oscillating beam 101 to rotate about the oscillating pivot 103, with the flexible pivot providing a restoring torque. Therefore, by measuring the swing displacement signal of the beam, the average thrust of the laser ablation target can be indirectly measured. The oscillating impulse measurement device 1 utilizes a displacement sensor 105 to record the linear displacement changes of the oscillation of the oscillating beam 101, thereby calculating the impulse of the laser ablation target based on the average displacement. When measuring the impulse, it is necessary to ensure that the entire device is not connected to any wires or pipes to avoid the influence of external connection torque on the impulse measurement results. The damper is used to quickly stabilize the oscillating beam 101, facilitating impulse measurement.
[0084] In a preferred embodiment, the measurement range of the torsional impulse is 10 -7 ~10 -5 μN·s.
[0085] In a preferred embodiment, the laser for ablating the target material is a pulsed laser.
[0086] Plasma spectrum measurement device 2 includes a spectrometer 201, an ICCD 202, a lens 203, and an optical fiber 204. Laser ablation device 3 includes a laser 301, a half-wave plate 302, a Glan prism 303, a focusing lens 304, a translation stage 305, and a synchronization device 306. Lens 203 collects plasma light, which is then connected to spectrometer 201 via optical fiber 204. ICCD 202 receives the optical signal from spectrometer 201 and the trigger signal from synchronization device 306 as input. Its output is an electrical signal of spectral intensity, which is ultimately transmitted to back-end data processing system 5.
[0087] In a preferred embodiment, the wavelength of the laser 301 is in the near-infrared band.
[0088] In a preferred embodiment, the energy of the laser 301 is adjustable within the range of 0-100 mJ.
[0089] In a preferred embodiment, the pulse width of the laser 301 is in the nanosecond order.
[0090] In a preferred embodiment, the frequency of the laser 301 is adjustable within the range of 1-10 Hz.
[0091] In one embodiment, the target data acquisition unit adopts an impulse measurement device based on a multi-camera system.
[0092] In a preferred embodiment, Figure 4 As shown, the impulse measurement device based on the multi-camera system adopts a dual-camera spectrum acquisition device 4, including two CCD detectors 401, two wave plate wheels 402, and corresponding bandpass filter groups 403.
[0093] In a preferred embodiment, the data processing unit adopts a back-end data processing system 5 including a computer 501 and data processing software 502 .
[0094] like Figure 3 As shown, two CCD detectors 401 synchronously detect plasma spectrum information of space debris, including:
[0095] Step 1: Confirm the target type of space debris;
[0096] Step 2: Select the appropriate filter based on the target material type and the specific wavelength spectrum selected in the above step;
[0097] Step 3: Trigger two CCD detectors when the laser ablates the target;
[0098] Step 4: Two CCD detectors collect plasma signals and transmit the signals simultaneously to the back-end data processing system 5;
[0099] Step 5: The back-end data processing system 5 processes the captured image, inputs the preferred characteristic parameters Al(II)358.66nm+Al(II)463.15nm into the model, and calculates the impulse data.
[0100] In one embodiment, the bandpass filter set 403 replaces filters through a wave plate wheel according to the sample and selects the corresponding spectral line wavelength range according to the intensity of the spectrum display.
[0101] In a preferred embodiment, the filtering range of the filter is determined according to the specific material of the space debris irradiated by the laser.
[0102] In a preferred embodiment, the specific wavelength is selected with reference to the best spectral line prediction combination calculated by the data processing unit.
[0103] Example 1
[0104] Consider the case of an aluminum target as the target material for space debris. Aluminum is a lightweight and durable metal widely used in the structural components of spacecraft and satellites. When these spacecraft or satellites collide, explode, or decommission in orbit, a significant portion of the debris produced is made of aluminum. Other common materials include stainless steel, titanium alloys, and composite materials, which also appear in space debris, but aluminum accounts for a significant proportion. In practice, the bandpass filter range can be adjusted based on the effective characteristic spectral lines of the sample.
[0105] First, the aluminum target is pretreated accordingly: Before the experiment, since there is usually an oxide film on the surface of aluminum, which may affect the experimental results, it is necessary to use a dilute acid (such as dilute hydrochloric acid or nitric acid) to dissolve this oxide film, and then rinse it thoroughly with deionized water to clean the surface of the aluminum. Because the experiment requires a smooth surface, the aluminum needs to be mechanically polished. Then wipe it dry with a lint-free cloth, or heat it slightly in a low-temperature oven to remove moisture. The aluminum target is then placed in the entire laser ablation device 3.
[0106] like Figure 2 As shown, a laser light source is used to emit a high-energy laser beam to irradiate a target material, and a torsional pendulum impulse measurement device 1 and a plasma spectrum measurement device 2 are used to simultaneously obtain impulse value data and spectral intensity information when the laser irradiates the target material; and different laser energies are changed to obtain a data set for realizing impulse prediction.
[0107] The acquired data is then preprocessed, including processing of missing values, processing of outliers, data normalization, and alignment of the spectral line intensity graph with the characteristic spectrum wavelength in the NIST database to ensure the accuracy of the spectral data.
[0108] Based on the fact that the sample size is not rich enough, the support vector machine (SVM) and BP-neural network methods are selected. Through ergodic search, the best model parameter combination and the best spectral line prediction combination are obtained.
[0109] like Figure 7 As shown, the present application compares the SVM model and the BP neural network model, and it can be found that the prediction accuracy of the two is very close, indicating that the selected model has a strong adaptability to the measured data set.
[0110] See also Figure 8 From the prediction accuracy of various spectral lines, we can see that the two spectral lines can achieve good impulse prediction results.
[0111] The above descriptions are merely a few embodiments of the present application and do not constitute any form of limitation to the present application. Although the present application discloses the preferred embodiments as above, they are not intended to limit the present application. Any technical personnel familiar with the present profession, without departing from the scope of the technical solution of the present application, using the technical content disclosed above to make slight changes or modifications are equivalent to equivalent implementation cases and fall within the scope of the technical solution.
Claims
1. A method for rapidly predicting impulse in laser debris removal, characterized in that: The method comprises the following steps: (1) Simultaneously obtain impulse data and spectral intensity data during laser ablation of space debris targets; obtain multiple sets of impulse data and spectral intensity data by continuously changing the laser parameters emitted by the laser light source; (2) preprocessing the data obtained in step (1) to form a valid data set; (3) constructing an impulse prediction model based on machine learning, the output of which is an impulse prediction value; and calculating the optimal model parameter combination and the optimal spectral line prediction combination of the model based on the valid data set; (4) During the laser debris removal process, the plasma spectrum information of the target debris target material is obtained in real time and input into the trained impulse prediction model for rapid impulse prediction; the plasma spectrum information includes spectral line intensity information of two or more specific wavelengths for the target material type.
2. The method for rapidly predicting impulse in laser debris removal according to claim 1, characterized in that: Before executing step (1), the target material needs to be pre-treated to remove and prevent the oxide film from forming again; Preferably, the preprocessing specifically includes: Use dilute acid to dissolve the oxide film on the target, then rinse thoroughly with deionized water; then polish the target surface; dry as soon as possible after cleaning.
3. The method for rapidly predicting impulse in laser debris removal according to claim 1, characterized in that: The laser parameters in step (1) include laser energy parameters.
4. The method for rapidly predicting impulse in laser debris removal according to claim 1, characterized in that: The preprocessing includes at least one of the following: processing of missing values, processing of outliers, normalization of data, and alignment of a spectral line intensity graph with a characteristic spectrum wavelength in a NIST database.
5. The method for rapidly predicting impulse in laser debris removal according to claim 1, characterized in that: The step (3) comprises: Support vector machine (SVM) and BP neural network methods were selected; the optimal model parameter combination of the impulse prediction model and the optimal spectral line prediction combination were obtained through ergodic search.
6. The method for rapidly predicting impulse in laser debris removal according to claim 1, characterized in that: In the step (4), the selection of the specific wavelength of the plasma spectrum information refers to the optimal spectrum line prediction combination.
7. A system for rapidly predicting impulse in laser debris removal, characterized in that: The system includes: The sample data acquisition unit is used to obtain impulse data and spectral intensity data when laser ablation of space debris targets; by continuously changing the laser parameters emitted by the laser light source, multiple sets of impulse data and spectral intensity data are obtained; A target data acquisition unit, configured to obtain plasma spectrum information of the target fragment target material, wherein the plasma spectrum information includes spectral line intensity information of two or more specific wavelengths for the target material type; A data processing unit is configured to preprocess the multiple sets of impulse data and spectral intensity data to form a valid data set; construct an impulse prediction model based on machine learning, the output of which is an impulse prediction value; calculate the optimal model parameter combination and the optimal spectral line prediction combination of the model based on the valid data set; obtain plasma spectral information and input it into the trained impulse prediction model to quickly predict the impulse.
8. The method for rapidly predicting impulse in laser debris removal according to claim 7, characterized in that: The sample data acquisition unit adopts a torsion pendulum impulse measurement device, a plasma spectrum measurement device and a laser ablation device; the lasers in the torsion pendulum impulse measurement device, the plasma spectrum measurement device and the laser ablation device are synchronized; The laser ablation device is used to provide lasers with variable energy, pulse width, and frequency, and to provide different ablation conditions; The torsion pendulum impulse measuring device is used to measure the average displacement of the torsion pendulum on which the target material is mounted during laser irradiation, and calculate the impulse of the laser ablation target material; The plasma spectrum measuring device is used to obtain spectrum intensity data under different ablation conditions; Preferably, the variable energy is adjustable within the range of 0-100 mJ; Preferably, the pulse width is in the order of nanoseconds; Preferably, the frequency is adjustable within the range of 1-10 Hz.
9. The method for rapidly predicting impulse in laser debris removal according to claim 7, characterized in that: The target data acquisition unit adopts an impulse measurement device based on a multi-camera system; Preferably, the impulse measurement device of the multi-camera system adopts a dual-camera spectrum acquisition device, including two CCD detectors and corresponding wave plate wheels and bandpass filter groups; wherein, The two CCD detectors are used to synchronously collect plasma signals of two specific wavelengths and send spectral intensity data to a data processing unit; The wave plate wheel and bandpass filter assembly are used to switch appropriate filters according to the selected spectrum lines of specific wavelengths according to the target material type of space debris.
10. The method for rapidly predicting impulse in laser debris removal according to claim 9, characterized in that: The specific wavelength is selected with reference to the best spectral line prediction combination calculated by the data processing unit; Preferably, the range of the bandpass filter is dynamically adjusted according to the effective characteristic spectrum of the sample.