All-solid-state laser thermal compensation method and device based on machine learning and laser

Abstract

The invention discloses an all-solid-state laser thermal compensation method and device based on machine learning and a laser, and the method comprises the steps: selecting the type of a lens model according to the parameters of a working substance of a to-be-compensated all-solid-state laser and the power of pump light, and the type of the lens model is a thin lens model or a thick lens model; and performing thermal compensation on the all-solid-state laser by using the thin lens model or the thick lens model and a negative lens. The built thick lens model gives consideration to the influence of the temperature gradient, the thermal stress and the end face deformation on the multi-section bonded laser crystal at the same time, the thermal lens effect of the multi-section doped crystal can be simulated more accurately, and therefore the performance of the all-solid-state laser is improved.

Description

Technical Field

[0001] The present invention relates to the field of laser thermal compensation, and in particular to a full-solid-state laser thermal compensation method, device and laser based on machine learning. Background Art

[0002] 2μm lasers are widely used in laser ranging, laser processing, laser medical treatment and other fields due to their strong penetrating ability, easy absorption by water molecules and safety for human eyes. However, due to the thermal lens effect, excessive accumulation of heat in the medium will lead to uneven temperature distribution inside the laser crystal, forming a thermal gradient, which will not only cause thermal expansion and thermal distortion of the crystal, but also lead to non-uniform changes in the refractive index, thus causing the laser beam quality to deteriorate. In addition, excessive heat may also cause stress concentration on the crystal surface, increase the risk of crystal breakage, and shorten the service life of the laser.

[0003] At present, researchers mainly alleviate the thermal effects of 2μm all-solid-state lasers from the aspects of pumping mode, crystal structure and cooling technology. For example, when performing thermal compensation on all-solid-state lasers using crystals with single doping concentration or multiple doping concentrations, a thermal lens model is used to achieve thermal compensation within the full pump power pumping range.

[0004] However, the same thermal lens model will produce quite different thermal compensation amounts for bonded crystals or single doped crystals in different pump power ranges. Therefore, a flexible and efficient all-solid-state laser thermal compensation method is needed to improve the output performance of 2μm lasers. Summary of the invention

[0005] The purpose of the present invention is to provide a method, device and laser for thermal compensation of an all-solid-state laser based on machine learning, which can solve at least one of the technical problems mentioned above. The specific scheme is as follows:

[0006] According to the specific embodiments disclosed in the present invention, the first aspect of the present invention provides a method for thermal compensation of an all-solid-state laser based on machine learning, comprising: selecting a lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, wherein the lens model type is: a thin lens model or a thick lens model;

[0007] Thermally compensating the all-solid-state laser using the thin lens model or the thick lens model and a negative lens;

[0008] The expression of the thick lens model is:

[0009]

[0010] The expression of the thin lens model is:

[0011]

[0012] Wherein, ωp represents the spot radius of the pump light; Pin represents the pump light power;

[0013] k c represents thermal conductivity; l c represents the crystal length;

[0014] n0 represents the refractive index of the crystal; b represents the radius of the crystal;

[0015] ζ represents the thermal load ratio of the crystal;

[0016] α T represents the absorption coefficient; C r,φ represents the photoelastic coefficient;

[0017] dn / dt represents the thermo-optical coefficient.

[0018] Preferably, the lens model type is selected according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, including:

[0019] Constructing a neural network screening model, inputting the parameters of the working substance and the pump light power into the neural network screening model to obtain the lens model type;

[0020] The working material includes: a crystal with a single doping concentration or a laser crystal bonded in multiple sections;

[0021] The parameters of the working material include: the doping concentration of the laser crystal and the corresponding length of the laser crystal segment.

[0022] Preferably, the neural network screening model is based on a back propagation neural network to create a multi-layer perceptron regression model;

[0023] The parameters of the input layer of the neural network screening model include: a simulation result of a thick lens model obtained using the thick lens model, a simulation result of a thin lens model obtained using the thin lens model, a measured thermal focal length value, parameters of a working substance, and a pump light power value.

[0024] Preferably, when the all-solid-state laser is thermally compensated using a thin lens model and a negative lens, the position distance of the negative lens in the resonant cavity of the all-solid-state laser satisfies:

[0025]

[0026] Wherein, L1 is the distance from the front end surface of the working material to the total reflection mirror;

[0027] L2 is the distance from the rear end face of the working material to the output mirror;

[0028] R2 is the radius of curvature of the output mirror.

[0029] Preferably, when the all-solid-state laser is thermally compensated using a thick lens model and a negative lens, the position distance of the negative lens in the resonant cavity of the all-solid-state laser satisfies:

[0030]

[0031]

[0032] Wherein, L1 is the distance from the front end surface of the working material to the total reflection mirror;

[0033] L2 is the distance from the rear end face of the working material to the output mirror;

[0034] R2 is the radius of curvature of the output mirror.

[0035] Preferably, the focal length of the negative lens ranges from -5 mm to -150 mm.

[0036] Preferably, the working material is: a multi-segment bonded Tm:YAG laser crystal; the first segment is undoped Tm 3+ , length is 3mm; the second segment Tm 3+ The doping concentration is 2at.%, and the length is 3mm; the third section Tm 3+ The doping concentration is 3.5at.%, and the length is 5mm.

[0037] Preferably, the distance between the front end surface of the working substance and the total reflection mirror ranges from 0 to 55 mm;

[0038] The distance between the rear end face of the working substance and the negative lens is in the range of 0 to 150 mm;

[0039] The distance between the negative lens and the output mirror ranges from 0 to 210 mm.

[0040] According to the specific embodiments disclosed in the present invention, the second aspect of the present invention discloses a full-solid-state laser thermal compensation device based on machine learning, which is used to implement the above method, including:

[0041] A lens model selection unit, used for selecting a lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, wherein the lens model type is a thin lens model or a thick lens model;

[0042] A thermal compensation calculation unit is used to perform thermal compensation on the all-solid-state laser by utilizing the thin lens model or the thick lens model and a negative lens.

[0043] According to the specific embodiments disclosed in the present invention, the third aspect of the present invention discloses an all-solid-state laser based on machine learning, including the above-mentioned all-solid-state laser thermal compensation device based on machine learning.

[0044] Compared with the prior art, the above solution disclosed in the present invention has at least the following beneficial effects:

[0045] The present invention simulates the thermal lens effect of laser crystals under different doping conditions by constructing a thin lens model and a thick lens model, and obtains a lens model type that can generate a matching doped crystal to be compensated according to the input pump power value by constructing a neural network screening model, that is, a thin lens model or a thick lens model is selected for thermal compensation, and the negative lens is further used to thermally compensate the laser. The thick lens model constructed by the present invention takes into account the influence of temperature gradient, thermal stress and end face deformation on multi-segment bonded laser crystals at the same time, and can more accurately simulate the thermal lens effect of multi-segment doped crystals, thereby improving the performance of all-solid-state lasers. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the disclosure of the present invention, and together with the specification, are used to explain the principles disclosed in the present invention. Obviously, the drawings described below are only some embodiments disclosed in the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0047] Figure 1 It is a flow chart of a method for thermal compensation of an all-solid-state laser based on machine learning of the present invention;

[0048] Figure 2a It is the temperature distribution diagram of the Tm:YAG crystal with a single doping concentration of the present invention;

[0049] Figure 2b It is a thermal stress distribution diagram of a Tm:YAG crystal with a single doping concentration of the present invention;

[0050] Figure 3a It is the temperature distribution diagram of the multi-stage bonded Tm:YAG crystal of the present invention;

[0051] Figure 3b It is a one-dimensional temperature curve distribution diagram of the multi-segment bonded Tm:YAG crystal of the present invention;

[0052] Figure 3c It is the thermal stress distribution diagram of the multi-stage bonded Tm:YAG crystal of the present invention;

[0053] Figure 4 A schematic diagram of a neural network screening model of the present invention;

[0054] Figure 5 It is a schematic diagram of the present invention for collecting a neural network screening model data set;

[0055] Figure 6a This is a simulation diagram of the stable region of the Tm:YAG laser resonator cavity using a thin lens model in the present invention;

[0056] Figure 6b This is a simulation diagram of the stable region of the Tm:YAG laser resonator cavity using a thick lens model in the present invention;

[0057] Figure 7a , Figure 7b , Figure 7c Schematic diagram of thermal compensation effect under different lens models of the present invention;

[0058] Figure 8 It is a schematic diagram of the thermal compensation effect of different focal lengths and different positions of the negative lens inserted into the cavity of the present invention;

[0059] Fig. 9 It is a structural schematic diagram of an all-solid-state laser thermal compensation device based on machine learning of the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical scheme and advantages disclosed in the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments disclosed in the present invention, rather than all the embodiments. Based on the embodiments disclosed in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection disclosed in the present invention.

[0061] The terms used in the embodiments disclosed in the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the disclosure of the present invention. The singular forms "a", "said" and "the" used in the embodiments disclosed in the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings, and "multiple" generally includes at least two.

[0062] It should be understood that although the terms first, second, third, etc. may be used to describe in the disclosed embodiments of the present invention, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the disclosed embodiments of the present invention, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0063] It should also be noted that the term "includes", "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, so that a commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprising a ..." do not exclude the existence of other identical elements in the commodity or device including the elements.

[0064] The optional embodiments disclosed in the present invention are described in detail below with reference to the accompanying drawings.

[0065] In an end-pumped solid-state laser system, the release of thermal energy will accumulate in the laser medium, causing phenomena such as temperature gradient, thermal stress and end-face deformation, which together affect the thermal focal length characteristics of the gain medium. Usually, in order to simplify complex calculations, the laser crystal is simplified into a thin lens model, focusing on the thermal effect of the temperature gradient. However, the influence of factors such as the substrate material, doping concentration, structural characteristics and pumping method of the gain medium on the thermal effect under different conditions cannot be ignored. Especially for commonly used bonded crystals, using a thin lens model for analysis may lead to large simulation errors.

[0066] Therefore, the first embodiment of the present invention provides a method for thermal compensation of an all-solid-state laser based on machine learning, comprising the following steps:

[0067] Step S1, selecting a lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, wherein the lens model type is a thin lens model or a thick lens model.

[0068] Step S2: thermally compensate the all-solid-state laser by using the thin lens model or the thick lens model and a negative lens.

[0069] This embodiment can take into account the influence of temperature gradient, thermal stress and end face deformation on multi-segment bonded laser crystals by constructing a thick lens model, thereby more accurately simulating the thermal lens effect of multi-segment doped crystals.

[0070] Specifically, in step S1, taking the working material of the all-solid-state laser as a single-doped Tm:YAG crystal and a multi-segment bonded Tm:YAG crystal as an example, the thin lens model and the thick lens model constructed by the present invention are further described.

[0071] like Figure 2a and Figure 2bAs shown in the figure, the temperature distribution and thermal stress distribution of the Tm:YAG crystal with a single doping concentration are shown. It can be seen that the temperature and deformation position inside the crystal are mainly concentrated in a small part of the front end of the laser crystal, resulting in a large temperature difference in the entire laser crystal, which is suitable for the thin lens model.

[0072] The expression for constructing the thin lens model is:

[0073]

[0074] Wherein, ωp represents the spot radius of the pump light; Pin represents the pump light power;

[0075] k c represents thermal conductivity; l c represents the crystal length;

[0076] n0 represents the refractive index of the crystal; b represents the radius of the crystal;

[0077] ζ represents the thermal load ratio of the crystal;

[0078] α T Indicates: Absorption coefficient; C r,φ represents the photoelastic coefficient; dn / dt represents the thermo-optical coefficient.

[0079] Figure 3a , Figure 3b and Figure 3c The temperature distribution, one-dimensional temperature curve distribution and thermal stress distribution of the multi-section bonded Tm:YAG crystal 3 (0at.%) + 3 (2at.%) + 5 (3.5at.%) are shown respectively. It can be seen that the temperature has reached 310.11K at the beginning of the laser crystal bonding surface. After a section that is more severely heated, the temperature can still reach 313.17K at the second bonding surface. Figure 3c It can also be seen that the portion of the laser crystal affected by thermal stress accounts for a large portion of the total crystal length. However, compared with the single doping concentration Tm:YAG crystal, the overall temperature of the bonded crystal drops more due to the homogenization absorption, which solves some of the thermal effects to a certain extent. Therefore, in order to better reflect the thermal lens effect, based on the thin lens model, the present invention constructs a more complex thick lens model, which comprehensively considers the temperature gradient f T 、 Thermal stress S and end deformation f D The impact on laser crystals enables more accurate simulation and analysis of the behavior of thermal effects.

[0080] Specifically, the laser crystal absorbs energy during the pumping process, resulting in uneven temperature distribution inside the crystal, forming a radial temperature gradient f T , causing the beam to converge or diverge.

[0081] The non-uniform thermal expansion of the crystal will generate thermal stress inside S Thermal stress causes stress birefringence effect inside the crystal, which changes the propagation path of light and produces thermal focus.

[0082] The thermal load causes mechanical deformation of the laser crystal end face, namely, the end face deformation f D The deformed end surface acts as a geometric lens for the laser beam, changing the convergence or divergence characteristics of the beam.

[0083] Therefore, the expression for constructing the thick lens model is:

[0084]

[0085] Wherein, ωp represents the spot radius of the pump light; Pin represents the pump light power;

[0086] k c represents thermal conductivity; l c represents the crystal length;

[0087] n0 represents the refractive index of the crystal; b represents the radius of the crystal;

[0088] ζ represents the thermal load ratio of the crystal;

[0089] α T Indicates: Absorption coefficient; C r,φ represents the photoelastic coefficient;

[0090] dn / dt represents the thermo-optical coefficient.

[0091] Furthermore, when the thermal focal length of the laser crystal is simulated using the constructed thin lens model and thick lens model, there is a difference in the accuracy of the thermal focal length simulated using the thin lens model and the thick lens model within the same pump light power range. Since the laser crystal changes in characteristics when it is irradiated by pump light, a matching thin lens model or thick lens model should be selected within different pump light power ranges.

[0092] In the initial stage, when the pump light irradiates the laser crystal, the heat conduction is not significant, and part of the heat energy is absorbed by the matrix without doped ions, resulting in little difference in the overall temperature of the laser crystal. The end face deformation is relatively small compared to the overall shape change of the laser crystal, so it is suitable to use a thick lens model.

[0093] As the pump power gradually increases, the rate at which the doped ions absorb photons increases, causing the temperature inside the laser crystal to rise rapidly. The heated lens effect on the crystal bonding surface is significant, and there is a significant difference in temperature with the back end of the laser crystal. The deformation also begins to increase, and the thin lens model is more suitable at this time.

[0094] As the pump power increases further, the temperature of the bonding surface of the laser crystal reaches a certain level, and the heat begins to transfer backward. At the same time, the deformation caused by thermal stress gradually increases. At this time, it is suitable to use the thick lens model for simulation.

[0095] Therefore, in a preferred embodiment of the present invention, a neural network screening model is established, the parameters of the laser crystal and the pump light power are input into the neural network screening model, and the matching lens model type is obtained using the neural network screening model.

[0096] Specifically, the neural network screening model is based on the back propagation neural network to create a multi-layer perceptron regression model. Figure 4 As shown, the parameters of the input layer of the neural network screening model include: the simulation results of the thick lens model obtained using the thick lens model, the simulation results of the thin lens model obtained using the thin lens model, the measured thermal focal length value, the parameters of the working material and the pump light power value.

[0097] The number of neurons in the hidden layer of the neural network screening model is set to (20, 20, 20), the maximum number of iterations is set to 1000, and the learning rate is 0.01.

[0098] The output layer of the neural network screening model includes: thin lens model, thick lens model and actual thermal focal length.

[0099] During the model training process, the model calculates the predicted value through forward propagation, and then uses the back-propagation algorithm to adjust the model weight according to the difference between the predicted value and the actual value, so as to gradually optimize the model to approach the real value. In the output stage, it is possible to make predictions by inputting any pump light power Pin, and compare it with the measured data and the predicted results of the thin lens model and the thick lens model. Finally, for a certain laser crystal, at a specific pump power point, the predicted actual thermal focal length and the results of selecting the thin lens model or the thick lens for simulation are output.

[0100] In other embodiments of the present invention, the thermal focal length of the laser crystal is simulated by using formulas (1) and (2), and the measured data is analyzed and compared with the results using the thin lens model and the thick lens model to form a data set for training the neural network screening model.

[0101] Specifically, build Figure 5As shown in the schematic diagram, a 785nm LD is first used as a pump source 1 to irradiate the crystal, and its light spot is focused on the first bonding surface of the Tm:YAG crystal 2, and a 45° folding mirror 4 is placed between the Tm:YAG crystal 2 and the coupling lens tube 3. The Tm:YAG crystal 2 produces a thermal lens effect after being heated, causing the red light emitted from the He-Ne laser 6 to pass through the Tm:YAG crystal 2, which has already shown a convex lens effect, and gradually lose the parallel transmission property and begin to focus.

[0102] By manually moving the ceramic white plate 5 to observe the evolution of the light spot, it is found that the light spot gradually focuses from a divergent ring to a focus, and the brightest point reflects the thermal focal length formed by the Tm:YAG crystal 2. In order to deal with the equivalence of the thermal focal length of the image side and the thermal focal length of the object side, and calculate the true thermal focal length value. Considering that there may be errors in manually adjusting the ceramic white plate and observing the light spot size with the human eye, multiple measurements were performed to improve accuracy. The thermal focal length of the crystal was simulated by using formulas (1) and (2), and the measured data was analyzed and compared with the results using the thin lens and thick lens models to form a data set.

[0103] As a verification of the accuracy of the model, Figure 6 shows the comparison results at different power points when the neural network screening model of the present invention is used for prediction. It can be seen that when the pump power is 5-9.5W and 18-30W, the simulation accuracy of the thick lens model is higher; while in the range of 9.5-18W, the thin lens model is more accurate. The maximum error of the model is only 1.8mm, and the minimum error accuracy can reach 1.9%.

[0104] In step S2, the length of the thermal focal length is predicted by the thin lens model or thick lens model selected at different power points in step S1, and negative lenses with different focal lengths are selected for thermal compensation. When the thin lens model or thick lens model is used respectively, the optical transfer matrix (ABCD matrix) is used, and according to the laser resonant cavity stability condition -1<(A+D) / 2<1, the following can be obtained respectively:

[0105] When using the thin lens model and the negative lens to perform thermal compensation on the all-solid-state laser, the position distance of the negative lens in the resonant cavity of the all-solid-state laser satisfies:

[0106]

[0107] Wherein, L1 is the distance from the front end surface of the working material to the total reflection mirror;

[0108] L2 is the distance from the rear end face of the working material to the output mirror;

[0109] R2 is the radius of curvature of the output mirror.

[0110] When using the thick lens model and the negative lens to perform thermal compensation on the all-solid-state laser, the position distance of the negative lens in the resonant cavity of the all-solid-state laser satisfies:

[0111]

[0112] Wherein, L1 is the distance from the front end surface of the working material to the total reflection mirror;

[0113] L2 is the distance from the rear end face of the working material to the output mirror;

[0114] R2 is the radius of curvature of the output mirror.

[0115] In a preferred embodiment of the present invention, the focal length of the thin lens model is greater than that of the thick lens model, and when the following positional relationships are met, thermal compensation can be effectively performed on the laser:

[0116] When the focal length range of the negative lens is -5mm to -150mm;

[0117] The distance between the front face of the Tm:YAG crystal in the resonant cavity and the total reflection mirror ranges from 0 to 55 mm;

[0118] The distance between the rear end face of the Tm:YAG crystal and the negative lens ranges from 0 to 150 mm;

[0119] The distance between the negative lens and the output mirror ranges from 0 to 210 mm.

[0120] Taking the working material as a multi-segment bonded crystal Tm:YAG 3(0at.%)+3(2at.%)+5(3.5at.%) as an example, negative lenses with focal lengths of -25mm, -50mm, -75mm, -100mm and -125mm were introduced at 50mm on the rear end face of the multi-segment bonded crystal, respectively, and a simulation comparison diagram of the stable zone of the Tm:YAG laser resonant cavity was obtained.

[0121] like Figure 6a As shown, when thermal compensation is performed for pump powers in the range of 8-18W, the compensation effects of different negative focal lengths are different.

[0122] Specifically, when the focal length is F = -25mm, almost all power points get the maximum cavity length compensation. When the power is 17W, the compensation range with a focal length of F = -125mm reaches the maximum value, and the length between the negative lens and the output mirror is 194.7mm. When the focal length is F = -25mm, the minimum value of the length L3 between the negative lens and the output mirror is only 16.7mm. This shows that within this pump power range, the longer the focal length, the better the compensation effect, and the thin lens model should be used.

[0123] like Figure 6bAs shown in the figure, for thermal compensation in the pump power range of 18-30W, all power points can be effectively compensated only when the focal length is F = -25mm and L3 = 200mm, while the compensation effect of the negative lens at other focal lengths is almost halved. When the thick lens model is used, the maximum compensation range achieved by the negative lens with F = -25mm can reach 173.4mm when the pump power is 30W. However, when the pump power is 18W, it can be observed that the compensation effect of the focal length of F = -125mm is significantly better than that of the focal length of F = -25mm.

[0124] It can be further seen that the combination of the negative lens and the lens model of the present invention can effectively perform thermal compensation, and no matter which model type is used, after the negative lens is inserted, the length L3 between the negative lens and the output mirror shows a significant extension trend relative to the case without the negative lens, which not only expands the total cavity length of the resonant cavity, but also provides a potential opportunity for introducing more optical elements into the resonant cavity.

[0125] 1. Comparison of experimental compensation effects of thin lens and thick lens models

[0126] In order to ensure the space for the negative lens to be inserted, the maximum total cavity length of the resonant cavity is set to 250mm, and a negative lens with F=-100mm is selected for compensation. The distance between the front end face of the crystal and the total reflection mirror is set to 6mm, the distance between the rear end face of the crystal and the negative lens is set to 4mm, and the distance between the negative lens and the output mirror is 190mm and 195mm respectively for the following comparative experiment.

[0127] Figure 7a This is the trend diagram of the continuous laser output power without negative lens under the thick lens model. It can be seen that when the pump power is 30.31W, the maximum output powers obtained are 7.01W and 8.35W respectively;

[0128] Figure 7b This is the trend diagram of the laser output power after using a negative lens with a focal length of -100mm under the thick lens model. It can be seen that the addition of the negative lens plays a role in thermal compensation when the pump power is in the range of 14.41W to 30.31W. Among them, when the pump power is 18.23W and 23.68W, the output power increases by 0.581W and 0.51W respectively.

[0129] Figure 7c The trend chart of laser output power using a -100mm negative lens under the thin lens model shows that when the pump power is 18.23W and 23.68W respectively, the output power only increases by 0.131W and 0.38W;

[0130] At this time, when compensating the laser with a cavity length of 210 mm, the laser output power drops by 1.178 W.

[0131] Therefore, it is necessary to select the thin-thick lens model or the thick lens model according to the actual pump power value.

[0132] 2. Comparison of thermal compensation effects of negative focal lengths with different focal lengths and different positions in the insertion cavity based on the thick lens model, such as Figure 8 shown.

[0133] Select a negative lens with F=-100mm for compensation, and the distance L1 between the front end of the crystal and the total reflector is 6mm. Figure 8 a and Figure 8 The total cavity length of b is 120mm, and the distance L2 between the rear end face of the laser crystal and the negative lens is 13mm and 18mm respectively. It can be seen that under the same cavity length conditions, the closer the distance between the negative lens and the crystal end face, the larger the adjustment range of the output power (the greater the pump power that can be tolerated).

[0134] In addition, when the cavity length is 90 mm, a negative lens with F=-75 mm is selected and the position of the negative lens is adjusted to compensate. Figure 8 d and 8e show that when the distance between the rear end face of the laser crystal and the negative lens is L2 = 3 mm, the adjustment range of the output power is significantly greater than that of the case of L2 = 8 mm.

[0135] Based on the above phenomenon, we used Rezonator to simulate the spot size in the cavity, such as Figure 8 c and 8f. The results show that in the distance between the rear end face of the crystal and the negative lens, the spot radius of the combination of L2=18mm / L2=8mm is significantly larger than that of the combination of L2=13mm / L2=3mm. As the intracavity laser is compensated by the negative lens, the spot radius of the combination of L2=13 / L2=3mmmm gradually increases, and eventually exceeds the combination of L2=18mm / L2=8mm. At the output mirror, the difference in spot radius between the two reaches 14mm / 14.6mm. This is because the increase in spot size when L2=13mm / L2=3mm leads to a corresponding expansion of the intracavity mode volume, which in turn increases the photon density allowed for oscillation, thereby helping to increase the output power of the laser. In addition, the increase in the intracavity mode volume will weaken the interaction between different modes, enhance the stability of the mode, and thus improve the overall stability of the laser.

[0136] In summary, the change of the focal length and position of the negative lens will have different effects on the laser compensation, when: the focal length range of the negative lens is -5mm to -150mm;

[0137] The distance between the front face of the Tm:YAG crystal in the resonant cavity and the total reflection mirror ranges from 0 to 55 mm;

[0138] The distance between the rear end face of the Tm:YAG crystal and the negative lens ranges from 0 to 150 mm;

[0139] When the distance between the negative lens and the output mirror is in the range of 0 to 210 mm, the laser can be effectively thermally compensated.

[0140] The second embodiment of the present invention discloses a device embodiment that is consistent with the above embodiment, which is used to implement the method steps described in the above embodiment. The explanation based on the same name meaning is the same as the above embodiment, and has the same technical effect as the above embodiment, which will not be repeated here.

[0141] As shown in FIG2 , the present invention discloses a full-solid-state laser thermal compensation device based on machine learning, comprising:

[0142] A lens model selection unit 201 is used to select a lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, wherein the lens model type is a thin lens model or a thick lens model;

[0143] The thermal compensation calculation unit 202 performs thermal compensation on the all-solid-state laser by using the thin lens model or the thick lens model and a negative lens.

[0144] The third embodiment of the present invention discloses an all-solid-state laser based on machine learning, including the all-solid-state laser thermal compensation device based on machine learning in the above-mentioned embodiment.

[0145] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to the various embodiments disclosed in the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0146] The units involved in the embodiments disclosed in the present invention may be implemented by software or hardware, wherein the name of a unit does not limit the unit itself in some cases.

Claims

1. A method for thermal compensation of all-solid-state lasers based on machine learning, characterized in that: include: Selecting a lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, wherein the lens model type is a thin lens model or a thick lens model; Thermally compensating the all-solid-state laser using the thin lens model or the thick lens model and a negative lens; The expression of the thick lens model is: The expression of the thin lens model is: Wherein, ωp represents the spot radius of the pump light; Pin represents the pump light power; k c represents thermal conductivity; l c represents the crystal length; n0 represents the refractive index of the crystal; b represents the radius of the crystal; ζ represents the thermal load ratio of the crystal; α T represents the absorption coefficient; C r,φ represents the photoelastic coefficient; dn / dt represents the thermo-optical coefficient.

2. The method according to claim 1, characterized in that The method of selecting the lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power comprises: Constructing a neural network screening model, inputting the parameters of the working substance and the pump light power into the neural network screening model to obtain the lens model type; The working material includes: a crystal with a single doping concentration or a laser crystal bonded in multiple sections; The parameters of the working material include: the doping concentration of the laser crystal and the corresponding length of the laser crystal segment.

3. The method according to claim 2, characterized in that The neural network screening model is based on a back propagation neural network to create a multi-layer perceptron regression model; The parameters of the input layer of the neural network screening model include: a simulation result of a thick lens model obtained using the thick lens model, a simulation result of a thin lens model obtained using the thin lens model, a measured thermal focal length value, parameters of a working substance, and a pump light power value.

4. The method according to claim 1, characterized in that: When the thin lens model and the negative lens are used to perform thermal compensation on the all-solid-state laser, the position distance of the negative lens in the resonant cavity of the all-solid-state laser satisfies: Wherein, L1 is the distance from the front end surface of the working material to the total reflection mirror; L2 is the distance from the rear end face of the working material to the output mirror; R2 is the radius of curvature of the output mirror.

5. The method according to claim 1, characterized in that When the thick lens model and the negative lens are used to perform thermal compensation on the all-solid-state laser, the position distance of the negative lens in the resonant cavity of the all-solid-state laser satisfies: Wherein, L1 is the distance from the front end surface of the working material to the total reflection mirror; L2 is the distance from the rear end face of the working material to the output mirror; R2 is the radius of curvature of the output mirror.

6. The method according to claim 1, characterized in that The focal length of the negative lens ranges from -5 mm to -150 mm.

7. The method according to claim 5, characterized in that The working material is: a multi-segment bonded Tm:YAG laser crystal; the first segment is undoped Tm 3+ , length is 3mm; the second segment Tm 3+ The doping concentration is 2at.%, and the length is 3mm; the third section Tm 3+ The doping concentration is 3.5at.%, and the length is 5mm.

8. The method according to claim 7, characterized in that The distance between the front end surface of the working material and the total reflection mirror ranges from 0 to 55 mm; The distance between the rear end face of the working substance and the negative lens is in the range of 0 to 150 mm; The distance between the negative lens and the output mirror ranges from 0 to 210 mm.

9. A full-solid-state laser thermal compensation device based on machine learning, characterized in that: The method for implementing any one of claims 1 to 8 comprises: A lens model selection unit, used for selecting a lens model type according to the parameters of the working material of the all-solid-state laser to be compensated and the pump light power, wherein the lens model type is a thin lens model or a thick lens model; A thermal compensation calculation unit is used to perform thermal compensation on the all-solid-state laser by utilizing the thin lens model or the thick lens model and a negative lens.

10. An all-solid-state laser based on machine learning, characterized in that: include: The device as claimed in claim 9.

Images