Monocrystalline silicon wafer low-loss slicing device and method based on fluid pressure and thermal field regulation

Abstract

This invention provides a low-loss slicing device and method for monocrystalline silicon wafers based on fluid pressure and thermal field control. It includes an intelligent jet cooling and sensing unit with a high-precision nozzle array and integrated miniature pressure sensors at the nozzle outlet or along the path adjacent to the cutting point; a central controller employing an improved incremental control algorithm to process the fluid pressure signal at the cutting point acquired in real time by the intelligent jet cooling and sensing unit; dynamically calculating the tension compensation amount acting on the tension regulator and introducing a nonlinear gain factor to adapt to the strong nonlinearity of the cutting process; the tension regulator adjusting the tension according to the tension compensation amount calculated by the central controller to achieve instantaneous matching between cutting resistance and wire mesh support force; and an actuator adjusting the jet parameters of the intelligent jet cooling and sensing unit according to the instructions of the central controller. This invention can solve the problems of low control accuracy, severe thermal damage, and lag in dynamic response existing in the prior art.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic and semiconductor material processing technology, specifically to a low-loss slicing device and method for monocrystalline silicon wafers based on fluid pressure and thermal field control, which is applicable to high-precision, low-loss slicing processing of monocrystalline silicon rods. Background Technology

[0002] The basic working principle of a high-speed reciprocating diamond wire slicer is as follows: hundreds of kilometers of diamond wire are guided by a main roller and maintained at a constant tension under a tension control system, forming a dense wire mesh. A silicon rod is propelled into the wire mesh by a feeding mechanism, and the diamond wire grinds and cuts the silicon rod in high-speed reciprocating motion. During the cutting process, a large amount of cutting fluid is sprayed onto the cutting area for cooling, lubrication, and chip removal.

[0003] Currently, traditional high-speed reciprocating diamond wire slicers have the following disadvantages: Bowing and Edge Chipping: During the dicing process, especially towards the end of the silicon ingot, the diamond wire bends and deforms due to grinding resistance (i.e., bowing). Current technologies, employing constant overall tension and uniform coolant spraying, struggle to compensate for bowing in real-time and locally. This leads to microcracks and edge chipping on the silicon wafer, particularly at the start and end of the dicing process (ExitChips), affecting its mechanical strength and subsequent process yield.

[0004] Silicon powder removal and thermal management issues: The uniformly sprayed coolant has insufficient penetration into the center of the wire mesh, making it difficult to effectively and promptly remove the high-temperature silicon powder generated during cutting from deep within the slits. The accumulation of silicon powder can lead to a surge in localized frictional heat, potentially causing localized thermal stress damage to the silicon wafer and exacerbating diamond wire wear.

[0005] Cutting stability and wire breakage risk: The constant tension mode cannot adapt to the constantly changing load during the cutting process. When encountering hard points inside the crystal or irregular shapes of the silicon rod, localized impact loads may cause increased wire vibration, increasing the risk of wire breakage, resulting in production interruptions and material losses.

[0006] Control and Response Lag: The existing tension control system and coolant system are independent of each other. Tension adjustment is based on preset curves or visual feedback from the bow, while cooling is based on temperature feedback, resulting in separate decision-making between the two. This leads to a lag in response to changes in the micro-state of the cutting point (such as instantaneous frictional heat and silicon powder accumulation pressure), making it impossible to achieve instantaneous coordination at the most critical points.

[0007] Energy Dispersion Rather Than Guidance: The coolant is mainly sprayed over a large area, dispersing energy. Its main function is to remove the heat that has been generated, rather than actively using fluid energy to assist the cutting process (such as using fluid dynamic pressure to assist chip breaking or drag reduction).

[0008] Blind wire bow compensation: Existing dynamic tension adjustment only aims to "straighten the wire mesh", but does not consider the local dynamic pressure changes at the cutting point of the wire mesh caused by coolant impact and silicon powder ejection. This compensation is inaccurate at the microscale. Summary of the Invention

[0009] This invention provides a low-loss slicing device and method for monocrystalline silicon wafers based on fluid pressure and thermal field control, aiming to solve the problems of low wire bow control accuracy, severe thermal damage, and lag in dynamic response in the prior art. By constructing a multi-physical field collaborative control system of fluid pressure, tension, and temperature, the microscopic physical state of the slicing process can be precisely controlled, which can reduce the thickness of the subsurface damage layer of the silicon wafer, and reduce the wire breakage rate and diamond wire consumption, thereby achieving the goal of reducing costs.

[0010] The present invention achieves the above objectives through the following technical solutions: A low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control, comprising: The intelligent jet cooling and sensing unit is equipped with a high-precision nozzle array and a miniature pressure sensor integrated at the nozzle outlet or on the path close to the cutting point. This sensor is used to detect the dynamic fluid pressure formed near the cutting point after the coolant is injected into the cutting slit in real time, so as to reflect the silicon powder discharge, the cutting friction state, and the coolant penetration depth. The central controller employs an improved incremental control algorithm, using the cutting point fluid pressure signal P(t)(t) collected in real time by the intelligent jet cooling and sensing unit as the core feedback quantity. Based on the current cutting depth h(t) and feed rate vf(t), it determines the target pressure window [Pmin(h,vf), Pmax(h,vf)] through a built-in process database or empirical formulas, and defines the real-time pressure deviation e(t) = P(t). Pt (h,vf), where Pt can be the center value of the window or dynamically adjusted according to the process stage; the tension compensation amount ΔT(t) acting on the tension regulator at the outlet end or the tension regulator at the inlet end is dynamically calculated. The tension compensation amount ΔT(t) consists of a proportional term, an integral term and a differential term, and a nonlinear gain factor is introduced to adapt to the strong nonlinearity of the cutting process. The tension regulator, including the inlet tension regulator and the outlet tension regulator, adjusts the tension according to the tension compensation amount ΔT(t) calculated by the central controller to achieve instantaneous matching between the cutting resistance and the wire mesh support force; The actuator, according to the instructions of the central controller, adjusts the jet parameters of the intelligent jet cooling and sensing unit. The jet parameters include at least jet pressure, angle, temperature or flow rate.

[0011] According to the present invention, a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control is provided. The tension compensation amount ΔT(t) consists of a proportional term, an integral term, and a differential term, and is expressed by the following formula:

[0012] Among them, the proportional term adopts In form, To determine the real-time pressure deviation, a piecewise proportional gain method is used. Value: When the absolute value of the pressure deviation | When | is less than the preset first threshold, take = When the absolute value of the pressure deviation is | When | is greater than the preset second threshold, take = When the absolute value of the pressure deviation is | | When the value is in the intermediate region between the first and second thresholds, linear interpolation is used to determine the value. The value of makes the proportional gain at and Smooth transition between them; The integral term is used to eliminate steady-state error by adjusting the pressure deviation. The integral is obtained by performing an integral operation; the integral limit is set, that is, the output value of the integral term is restricted so that it does not exceed the preset maximum and minimum values, and the integral coefficient of the integral term is reduced or the integral operation is paused at the beginning and end of the cutting stage, depending on the actual cutting situation. The differential term is used to reflect the pressure deviation. The pressure change trend, through pressure deviation The result is obtained by performing differentiation; a first-order low-pass filter is connected in series before the differentiation operation to filter the pressure signal. The time constant of the first-order low-pass filter is... T f Selected based on sampling frequency and noise level; The calculated proportional, integral, and differential terms are added together to obtain the final tension compensation amount. .

[0013] According to the present invention, a low-loss slicing device for monocrystalline silicon wafers based on fluid pressure and thermal field regulation includes a central controller with a strategy for allocating the total tension compensation amount ΔT(t) in a dual-regulator independent control scenario, comprising: The total compensation amount ΔT(t) is proportionally allocated to the inlet tension regulator and the outlet tension regulator, expressed as: Inlet tension compensation amount:

[0014] Outlet tension compensation: .

[0015] According to the present invention, a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control is provided. The allocation coefficient α is dynamically adjusted according to the slicing stage, and the adjustment rule is as follows: In the early stage of cutting, α>1, the focus is on tension adjustment at the inlet end to suppress the initial wire bow and ensure stability in the initial stage of cutting. In the middle stage of cutting, α=1, and the inlet and outlet ends are balanced to maintain a stable tension state of the wire mesh during cutting. In the later stage of cutting, α<1, the focus is on tension adjustment at the outlet end to compensate for the tension changes caused by the weakening of the silicon rod material support.

[0016] According to the present invention, a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control is provided. The actuator includes a jet parameter adjustment component group, and the specific implementation includes: The jet pressure regulating component adopts a combination structure of a high-pressure proportional pump and an electro-hydraulic proportional pressure valve. The input end of the high-pressure proportional pump is connected to the coolant storage tank, and the output end of the high-pressure proportional pump is connected to the high-precision nozzle array through a pipeline. The electro-hydraulic proportional pressure valve is installed on the pipeline between the pump and the nozzle. The jet angle adjustment component includes an electric rotary nozzle holder and an angle sensor. A high-precision nozzle array is fixedly installed on the electric rotary nozzle holder, and the angle sensor provides real-time feedback on the angle between the nozzle outlet and the cutting line feed direction. The jet temperature regulating component consists of a coolant precooling heat exchanger, a cryogenic medium storage tank, and a three-way proportional mixing valve. The precooling heat exchanger precools the coolant to 0~5℃ through circulating cooling water. The cryogenic medium storage tank stores liquid nitrogen or ethylene glycol cryogenic medium. The three ports of the three-way proportional mixing valve are respectively connected to the precooling coolant pipeline, the ambient temperature coolant pipeline, and the cryogenic medium pipeline. The jet flow rate regulating component uses a frequency-controlled magnetic circulation pump, which is installed on the coolant return line to regulate the coolant circulation flow rate.

[0017] According to the present invention, a low-loss slicing device for monocrystalline silicon wafers based on fluid pressure and thermal field control is provided. The jet pressure regulating component is used to receive the pressure regulating command of the central controller. By changing the output pressure of the proportional pump or the opening of the proportional valve, the pressure of the coolant injected into the cutting slit is adjusted in real time so that the coolant pressure matches the fluid pressure target window [Pmin, Pmax] at the cutting point. When the pressure deviation e(t) is large, the controller commands the proportional pump to increase the pressure or the proportional valve to increase the opening to enhance the penetration ability of the coolant into the cutting slit and discharge silicon powder. The jet angle adjustment component is used to receive the angle adjustment command from the central controller. It drives the nozzle seat to rotate around the vertical axis of the cutting line through the electric drive mechanism, and adjusts the relative angle between the nozzle outlet and the cutting point. When the cutting depth h(t) increases, the controller commands the nozzle seat to tilt deeper into the cutting slit. The three-way proportional mixing valve is used to receive temperature regulation commands from the central controller and mix precooling coolant and low-temperature medium in a set ratio, so that the temperature of the coolant injected into the cutting point drops to the target low-temperature range, forming a local supercooled zone in front of the cutting point, and reducing cutting resistance by utilizing the low-temperature brittleness of the material. The magnetic circulation pump is used to receive flow regulation commands from the central controller. By changing the motor speed, it adjusts the flow rate of coolant injected into the cutting slit per unit time. When the pressure deviation e(t) is positive, the controller commands the pump to speed up and increase the flow rate, thereby enhancing the flushing effect of the coolant on the silicon powder.

[0018] According to the present invention, a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control optimizes the calculation of tension compensation amount ΔT(t) by introducing a nonlinear gain factor to address the strong nonlinear characteristics of the slicing process. The optimization includes: Define nonlinear gain factor function This function uses real-time pressure deviation. e ( t ) and its rate of change As input variables, output a gain coefficient used to adjust the tension compensation amount; The nonlinear gain factor function adopts a piecewise nonlinear mapping relationship and dynamically adjusts the gain coefficient according to the different ranges of pressure deviation and its rate of change during the cutting process. In tension compensation amount Δ T ( t Based on the calculation, a nonlinear gain factor is introduced. G After weighting and adjusting the proportional, integral, and differential terms, the optimized formula for calculating the tension compensation is as follows:

[0019] in, For segmented proportional gain, For integral gain, This is the differential gain; The central controller uses real-time collected pressure deviation data. e ( t ) and its rate of change According to the preset nonlinear gain factor function Dynamically update nonlinear gain factor G In each control cycle Ts Inside, the central controller recalculates the nonlinear gain factor. GAnd apply it to the calculation of tension compensation for the current cycle.

[0020] A low-loss slicing method for single-crystal silicon wafers based on fluid pressure and thermal field control includes the following steps: Real-time monitoring of fluid pressure signals near the cutting point; using a miniature pressure sensor integrated in the intelligent jet cooling and sensing unit to sense the dynamic fluid pressure formed near the cutting point after the coolant is injected into the cutting slit. To determine whether the fluid pressure signal deviates from the preset target range, the target pressure window [Pmin(h,vf), Pmax(h,vf)] is determined based on the current cutting depth h(t) and feed rate vf(t) using the built-in process database or empirical formulas. This window serves as the preset target range, and the real-time pressure deviation e(t) = P(t) is defined. Pt(h,vf) is used to determine whether e(t) is within a reasonable range. When the fluid pressure signal deviates from the preset target range, perform the following operations simultaneously: Adjusting the jet parameters of the intelligent jet cooling unit: If the P value is too high, it indicates that the silicon powder discharge is obstructed and friction is aggravated. According to the instructions of the central controller, the jet pressure is increased instantaneously and / or the angle is adjusted to open the chip removal channel and reduce the P value. If the P value is too low, it indicates that the jet has failed to effectively enter the cutting point or the cutting is in an abnormally easy state. The central controller will work together to reduce the tension and adjust the jet to ensure effective energy input. The central controller starts the high-pressure chip removal and tension following mode, increases the jet pressure, and directionally impacts the cutting point to flush away the silicon powder. Based on the deviation in fluid pressure, the compensation amount for the inlet and / or outlet tension regulators is calculated and executed in real time: The central controller treats the increase or decrease of the P value as a signal of change in cutting resistance and immediately instructs the outlet tension regulator to perform a micro-incremental tension compensation proportional to the change in the P value; according to the dynamic adjustment of the tension distribution coefficient α in the cutting stage, the total compensation amount ΔT(t) is distributed to the inlet and outlet ends.

[0021] According to the present invention, a low-loss slicing method for monocrystalline silicon wafers based on fluid pressure and thermal field control is provided. When the fluid pressure signal P(t) at the cutting point is within a preset target range [Pmin(h,vf), Pmax(h,vf)], but the temperature detection value is higher than the ideal temperature threshold of the corresponding process stage, the controller activates the low-temperature cooling and tension fine-tuning mode. The actuator receives instructions from the central controller and adjusts the jet parameters through one or a combination of the following two methods: Pre-cooling regulation: A plate heat exchanger is added to the coolant supply pipeline, with its hot end connected to an independent refrigeration cycle system and its cold end exchanging heat with the coolant pipeline; the central controller adjusts the temperature deviation ΔT according to the temperature. tempThe compressor frequency of the refrigeration cycle system is dynamically adjusted so that the coolant temperature is pre-cooled to 5~15°C below the target temperature before entering the nozzle array, forming a low-temperature jet. Cryogenic Medium Mixing: A dual-medium mixing chamber is installed in the coolant storage tank, where the main coolant and the cryogenic auxiliary medium are mixed proportionally via an electromagnetic proportional valve; the central controller adjusts the mixing ratio based on ΔT. temp Calculate the required proportion of cryogenic medium to be mixed in, and instruct the electromagnetic proportional valve to adjust the opening so that the temperature of the mixed coolant meets the local subcooling requirements. The central controller uses a temperature change trend prediction module to predict temperature deviation ΔT. temp By performing differentiation, the rate of temperature change is obtained. ; according to The sign and size are determined by the central controller, which calculates the tension pre-adjustment amount ΔT using an improved incremental control algorithm. pre : when When ΔT > 0, pre Take a positive value to increase the wire mesh tension in advance to resist the increasing cutting resistance; when When <0, ΔT pre Taking a negative value reduces the wire tension in advance to avoid excessive constraint leading to wire breakage. Tension pre-adjustment amount ΔT pre According to the dynamic allocation coefficient α temp The tension is distributed to the inlet and outlet tension regulators to achieve a small pre-adjustment of the wire tension.

[0022] According to the present invention, a low-loss slicing method for single-crystal silicon wafers based on fluid pressure and thermal field control is provided. When the detected value of the bow state deviates from the preset safety range but the fluid pressure signal P(t) at the cutting point is within the target pressure window [Pmin(h,vf), Pmax(h,vf)], the wire mesh spatial attitude data is collected in real time by a high-speed CCD camera, and the bow curvature radius R(t) is extracted by a sub-pixel level edge detection algorithm. The central controller compares the bow curvature radius R(t) with a preset safety threshold [R]. min , R max In contrast, when R(t) < R min Or R(t)>R max When this occurs, it is determined to be an abnormal state of the bow line; The fluid pressure signal P(t) collected by the intelligent jet unit is acquired synchronously, and the signal noise is eliminated by Kalman filtering algorithm to verify whether P(t) is continuously within the target pressure window.

[0023] When the abnormal duration of the bow is t abnormalWhen the value is ≤0.2s, the central controller determines it as a transient disturbance and prioritizes maintaining the current tension setpoint T0; If 0.2s < t abnormal ≤1s, initiate visual feedback-driven tension adjustment, based on the bow curvature deviation ΔR(t) = R(t) - R target Calculate the visual accommodation amount ΔT visual , where R target Given the ideal arc curvature radius, the input variables ΔR(t) and dΔR / dt are determined using a fuzzy control algorithm, and the output variable ΔT is determined. visual The fuzzy set is divided into {negative large, negative small, zero, positive small, positive large}, and the membership function adopts the trigonometric function; When t abnormal When the change rate of the arc curvature is greater than 1 s or |dΔR / dt| > 0.5 mm / s, the fluid pressure signal P(t) is introduced as a constraint to construct a multi-objective optimization function:

[0024] in, , , For the weighting coefficients, ΔP(t) = P(t) - P center P center The target pressure window value; The optimization function is solved using the particle swarm optimization algorithm to obtain the cooperative adjustment amount ΔT. synergy .

[0025] Therefore, compared with the prior art, the low-loss slicing device and method for single-crystal silicon wafers based on fluid pressure and thermal field control proposed in this invention have the following beneficial effects: 1. Existing technologies mainly rely on macroscopic parameters such as the wire bow and overall temperature for indirect control, which suffers from significant lag and inability to reflect the true state of the cutting interface. This invention, for the first time, introduces the key physical quantity of fluid pressure at the cutting point into the closed-loop control circuit, directly reflecting the chip removal state and tribological behavior within the cutting kerf. This enables the system to sense and respond in real time to microscopic chip blockage or frictional abrupt changes, improving control precision and response speed by an order of magnitude, and truly achieving precise control over the material removal mechanism.

[0026] 2. This invention endows the coolant with dual functions of "sensing" and "damping," constructing a dynamic self-balancing mechanism for wire mesh tension, greatly improving cutting stability. This invention breaks through the traditional understanding that coolant is merely a heat exchange medium, utilizing the hydrodynamic pressure effect formed by the coolant during high-speed movement: as a sensor, fluctuations in fluid pressure directly reflect changes in the cutting load, achieving high-sensitivity state monitoring without the need for additional force sensors; as a damper, it utilizes the liquid film formed by hydrodynamic pressure to actively suppress the vibration of the wire mesh. This invention adjusts tension in real time through pressure signal feedback, ensuring that the wire mesh tension always maintains a dynamic balance with the cutting load, completely solving the problem of wire mesh jitter that easily occurs under variable loads with traditional constant tension control, significantly improving the stability of the cutting process.

[0027] 3. This invention utilizes localized supercooling technology to actively regulate the thermodynamic state of silicon material within a micro-domain at the cutting point, altering its critical condition for brittle-plastic transition. By opening a thermo-mechanical auxiliary window, the silicon material exhibits superior plastic deformation capability during the removal process, thereby significantly reducing cutting force from the source. This not only reduces abrasive wear but also prevents lattice fragmentation caused by excessive cutting force.

[0028] 4. This invention can significantly improve the quality of silicon wafer processing, achieving ultimate control of the subsurface damage layer and a leap in geometric precision, as well as ultimate control of force and heat at the cutting point: Damage layer control: The subsurface damage layer (SSD) of the silicon wafer is shallower and more uniformly distributed, which greatly reduces the amount of material removed during subsequent grinding and polishing, and improves material utilization. Geometric accuracy: Total thickness deviation (TTV) and warpage of silicon wafers have been fundamentally improved, reaching a new level of submicron level control; Improved yield: Due to the suppression of wire vibration and smooth chip removal, the wire breakage rate is reduced to an extremely low level, which not only reduces the cost of consumables, but also avoids the scrapping of crystal rods due to wire breakage, greatly improving production yield and economic benefits.

[0029] 5. This invention does not require radical modifications to existing diamond wire cutting machines; it only requires integration into the coolant supply system and thermal field control module to achieve enhanced control over the cutting process. This method is not only applicable to monocrystalline silicon wafers, but with slight adjustments, it can also be extended to low-loss slicing of other hard and brittle materials such as silicon carbide and sapphire, demonstrating broad industrial application prospects.

[0030] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0031] Figure 1This is a schematic diagram of the architecture of an embodiment of a low-loss slicing device for monocrystalline silicon wafers based on fluid pressure and thermal field control according to the present invention.

[0032] Figure 2 This is a schematic diagram illustrating the implementation principle of an embodiment of a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control according to the present invention.

[0033] Figure 3 This is a schematic diagram illustrating the rules for dynamically adjusting the allocation coefficient α according to the cutting stage in an embodiment of a low-loss slicing device for monocrystalline silicon wafers based on fluid pressure and thermal field control according to the present invention.

[0034] Figure 4 This is a schematic diagram of the actuator in an embodiment of a low-loss slicing device for monocrystalline silicon wafers based on fluid pressure and thermal field control according to the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0036] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0037] An embodiment of a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control. See Figures 1 to 4 This embodiment provides a low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control, comprising: The intelligent jet cooling and sensing unit is equipped with a high-precision nozzle array and a miniature pressure sensor integrated at the nozzle outlet or on the path close to the cutting point. This sensor is used to detect the dynamic fluid pressure formed near the cutting point after the coolant is injected into the cutting slit in real time, so as to reflect the silicon powder discharge, the cutting friction state, and the coolant penetration depth. The central controller employs an improved incremental control algorithm, using the cutting point fluid pressure signal P(t)(t) collected in real time by the intelligent jet cooling and sensing unit as the core feedback quantity. Based on the current cutting depth h(t) and feed rate vf(t), it determines the target pressure window [Pmin(h,vf), Pmax(h,vf)] through a built-in process database or empirical formulas, and defines the real-time pressure deviation e(t) = P(t). Pt (h,vf), where Pt can be the center value of the window or dynamically adjusted according to the process stage; the tension compensation amount ΔT(t) acting on the tension regulator at the outlet end or the tension regulator at the inlet end is dynamically calculated. The tension compensation amount ΔT(t) consists of a proportional term, an integral term and a differential term, and a nonlinear gain factor is introduced to adapt to the strong nonlinearity of the cutting process. The tension regulator, including the inlet tension regulator and the outlet tension regulator, adjusts the tension according to the tension compensation amount ΔT(t) calculated by the central controller to achieve instantaneous matching between the cutting resistance and the wire mesh support force; The actuator, according to the instructions of the central controller, adjusts the jet parameters of the intelligent jet cooling and sensing unit. The jet parameters include at least jet pressure, angle, temperature or flow rate.

[0038] In this embodiment, the tension compensation amount ΔT(t) consists of a proportional term, an integral term, and a differential term, and is expressed by the following formula:

[0039] Among them, the proportional term adopts In form, To determine the real-time pressure deviation, a piecewise proportional gain method is used. Value: When the absolute value of the pressure deviation | When | is less than the preset first threshold, take = , A smaller proportional gain value is used to avoid frequent fine-tuning due to small pressure deviations, thus ensuring system stability; when the absolute value of the pressure deviation is | When | is greater than the preset second threshold, take = , A larger proportional gain value is used to achieve a rapid response to large pressure deviations and timely adjustment of the tension compensation; when the absolute value of the pressure deviation is | | When the value is in the intermediate region between the first and second thresholds, linear interpolation is used to determine the value. The value of makes the proportional gain at and Smooth transition between them.

[0040] The integral term is used to eliminate steady-state error by adjusting the pressure deviation. To prevent integral saturation, an integral limit is set to restrict the output value of the integral term, ensuring it does not exceed the preset maximum and minimum values. Furthermore, at the beginning and end of the cutting process, the integral coefficient of the integral term is appropriately reduced or the integral operation is paused based on the actual cutting situation. This is to avoid the integral term from having an adverse effect on the system due to the instability of the initial cutting state and the special working conditions at the end of the cutting process. The differential term is used to reflect the pressure deviation. The pressure change trend, through pressure deviation To reduce the impact of noise in the pressure signal on the differential term, a first-order low-pass filter is connected in series before the differential operation to filter the pressure signal. The time constant of the first-order low-pass filter is... T f The sampling frequency and noise level are selected to ensure that the pressure deviation trend can be accurately reflected while effectively filtering out noise. The calculated proportional, integral, and differential terms are added together to obtain the final tension compensation amount. .

[0041] In this embodiment, the central controller employs the following strategy to allocate the total tension compensation amount ΔT(t) in a dual-regulator independent control scenario: The total compensation amount ΔT(t) is proportionally allocated to the inlet tension regulator and the outlet tension regulator, and the specific allocation formula is as follows: Inlet tension compensation amount:

[0042] Outlet tension compensation: .

[0043] Among them, such as Figure 3 As shown, the allocation coefficient α is dynamically adjusted according to the cutting stage, and the adjustment rule is as follows: In the early stage of cutting (near the silicon rod inlet), α > 1. At this time, the focus is on tension adjustment at the inlet end to suppress the initial wire bowing by using the larger tension compensation at the inlet end, ensuring the stability of the initial cutting stage. In the middle stage of cutting (stable cutting section), α = 1. The inlet and outlet ends are balanced to keep the wire mesh in a stable tension state during the cutting process, ensuring the consistency of cutting quality. In the later stage of cutting (near the silicon rod outlet), α < 1. The focus is on tension adjustment at the outlet end to compensate for the tension changes caused by the weakening of the silicon rod material support, avoiding problems such as silicon wafer damage at the end of the cutting stage. Furthermore, the control period Ts is set to 30~50 ms. This control period is set to match the sampling rate of the pressure sensor and the response time of the actuator, ensuring that the system can obtain the feedback information of the pressure sensor in a timely manner and quickly adjust the actuator according to the calculated tension compensation amount, so as to realize real-time and precise control of the cutting process and improve the stability of cutting and the quality of slices.

[0044] In this embodiment, as Figure 4 As shown, the actuator includes a jet parameter adjustment component group, and the specific implementation methods include: The jet pressure regulating component adopts a combination structure of a high-pressure proportional pump and an electro-hydraulic proportional pressure valve. The input end of the high-pressure proportional pump is connected to the coolant storage tank, and the output end of the high-pressure proportional pump is connected to the high-precision nozzle array through a pipeline. The electro-hydraulic proportional pressure valve is installed on the pipeline between the pump and the nozzle. The jet angle adjustment component includes an electric rotary nozzle holder and an angle sensor. A high-precision nozzle array is fixedly installed on the electric rotary nozzle holder, and the angle sensor provides real-time feedback on the angle between the nozzle outlet and the cutting line feed direction. The jet temperature regulating component consists of a coolant precooling heat exchanger, a cryogenic medium storage tank, and a three-way proportional mixing valve. The precooling heat exchanger precools the coolant to 0~5℃ through circulating cooling water. The cryogenic medium storage tank stores liquid nitrogen or ethylene glycol cryogenic medium. The three ports of the three-way proportional mixing valve are respectively connected to the precooling coolant pipeline, the ambient temperature coolant pipeline, and the cryogenic medium pipeline. The jet flow rate regulating component uses a frequency-controlled magnetic circulation pump, which is installed on the coolant return line to regulate the coolant circulation flow rate.

[0045] The jet pressure regulating component is used to receive pressure regulating commands from the central controller. By changing the output pressure of the proportional pump or the opening of the proportional valve, the pressure of the coolant injected into the cutting slit is adjusted in real time to match the coolant pressure with the fluid pressure target window [Pmin, Pmax] at the cutting point. When the pressure deviation e(t) is large, the controller commands the proportional pump to increase the pressure or the proportional valve to increase the opening to enhance the penetration ability of the coolant into the cutting slit and discharge silicon powder. The jet angle adjustment component is used to receive the angle adjustment command from the central controller. It drives the nozzle seat to rotate around the vertical axis of the cutting line through the electric drive mechanism, and adjusts the relative angle between the nozzle outlet and the cutting point. When the cutting depth h(t) increases, the controller commands the nozzle seat to tilt towards the depth of the cutting slit to ensure that the coolant accurately covers the leading edge of the cutting point. The three-way proportional mixing valve is used to receive temperature regulation commands from the central controller and mix precooling coolant and low-temperature medium in a set ratio, so that the temperature of the coolant injected into the cutting point drops to the target low-temperature range, forming a local supercooled zone in front of the cutting point, and reducing cutting resistance by utilizing the low-temperature brittleness of the material. The magnetic circulation pump is used to receive flow regulation commands from the central controller. By changing the motor speed, it adjusts the flow rate of coolant injected into the cutting slit per unit time. When the pressure deviation e(t) is positive (the pressure is too high and chip removal is obstructed), the controller commands the pump to speed up and increase the flow rate, thereby enhancing the flushing effect of the coolant on the silicon powder. As can be seen, the above-mentioned adjustment components communicate with the central controller through the fieldbus, receive and execute the coordinated adjustment commands generated by the controller in real time based on the fluid pressure signal P(t) at the cutting point, the cutting depth h(t) and the feed speed vf(t), so as to realize the dynamic matching of jet pressure, angle, temperature and flow rate, so as to optimize the cooling, chip removal and thermal field control effect at the cutting point.

[0046] In this embodiment, to address the strong nonlinear characteristics of the cutting process, a nonlinear gain factor is introduced to optimize the calculation of the tension compensation amount ΔT(t). The specific implementation method is as follows: Define nonlinear gain factor function This function uses real-time pressure deviation. e ( t ) and its rate of change The input variable is a gain coefficient used to adjust the tension compensation amount.

[0047] The nonlinear gain factor function employs a piecewise nonlinear mapping relationship, dynamically adjusting the gain coefficient based on the different ranges of pressure deviation and its rate of change during the cutting process. For example: When pressure deviation | e ( t |Smaller and rate of change| When | is also small, the gain coefficient G Choose a smaller value to avoid over-adjustment of the system due to small fluctuations and ensure the stability of the cutting process.

[0048] When pressure deviation | e ( t | Larger or greater rate of change | When | is large, the gain coefficient G A larger value is selected to quickly respond to strong nonlinear changes during the cutting process, adjust the tension compensation amount in a timely manner, and prevent faults such as bow loss of control or wire breakage.

[0049] In the intermediate transition region of pressure deviation and its rate of change, the gain coefficient G The transition is smoothed by nonlinear interpolation, ensuring the continuity and stability of system regulation.

[0050] In the original tension compensation amount Δ T ( t Based on the calculation, a nonlinear gain factor is introduced. GThe proportional, integral, and derivative terms are weighted and adjusted. The optimized formula for calculating the tension compensation is:

[0051] in, For segmented proportional gain, For integral gain, For the differential gain, the calculation methods for the integral and differential terms are consistent with the aforementioned formulas. This is achieved by introducing a nonlinear gain factor. To make the tension compensation amount It can better adapt to the strong nonlinear characteristics of pressure deviation and its rate of change during the cutting process, achieve a more precise match between cutting resistance and wire mesh support force, and further improve the stability of the cutting process and the quality of the slices.

[0052] Furthermore, the central controller uses real-time collected pressure deviation data... e ( t ) and its rate of change According to the preset nonlinear gain factor function Dynamically update gain coefficient G .

[0053] Furthermore, in each control cycle Ts Inside, the central controller recalculates the nonlinear gain factor. G This information is then applied to the calculation of tension compensation in the current cycle to ensure that the system can track the nonlinear changes in the cutting process in real time and optimize closed-loop control.

[0054] An Example of a Low-Loss Slicing Method for Single-Crystal Silicon Wafers Based on Fluid Pressure and Thermal Field Control This embodiment provides a low-loss slicing method for single-crystal silicon wafers based on fluid pressure and thermal field control, comprising the following steps: Real-time monitoring of fluid pressure signals near the cutting point; using a miniature pressure sensor integrated in the intelligent jet cooling and sensing unit to sense the dynamic fluid pressure formed near the cutting point after the coolant is injected into the cutting slit. To determine whether the fluid pressure signal deviates from the preset target range, the target pressure window [Pmin(h,vf), Pmax(h,vf)] is determined based on the current cutting depth h(t) and feed rate vf(t) using the built-in process database or empirical formulas. This window serves as the preset target range, and the real-time pressure deviation e(t) = P(t) is defined. Pt(h,vf) is used to determine whether e(t) is within a reasonable range. When the fluid pressure signal deviates from the preset target range, perform the following operations simultaneously: Adjusting the jet parameters of the intelligent jet cooling unit: If the P value is too high, it indicates that the silicon powder discharge is obstructed and friction is aggravated. The controller command unit will instantly increase the jet pressure and / or adjust the angle to open the chip removal channel and reduce the P value. If the P value is too low, it may indicate that the jet has failed to effectively enter the cutting point or that the cutting is in an abnormally easy state. The controller will then work together to reduce the tension and adjust the jet to ensure effective energy input. The controller starts the "high-pressure chip removal + tension following" mode, and the unit increases the jet pressure to directionally impact the cutting point and flush away the silicon powder. Based on the deviation in fluid pressure, the compensation amount for the inlet and / or outlet tension regulators is calculated and executed in real time: The controller interprets any increase or decrease in the P value as a signal of changing cutting resistance and immediately instructs the outlet tension regulator to perform a micro-incremental tension compensation proportional to the magnitude of the P value change (e.g., ΔT = K * ΔP). cut This incremental tension helps the wire mesh resist increased or decreased resistance, stabilizes the wire bow, and slightly alters the contact state between the diamond wire and the silicon material by utilizing minute changes in tension, thus aiding in chip breaking. The tension distribution coefficient α is dynamically adjusted according to the cutting stage, distributing the total compensation ΔT(t) to the inlet and outlet ends: in the early cutting stage (near the silicon rod inlet), α>1, focusing on tension adjustment at the inlet end; in the middle cutting stage (stable cutting section), α=1, with balanced adjustment; in the later cutting stage (near the silicon rod outlet), α<1, focusing on tension adjustment at the outlet end.

[0055] If P is within the range, but the temperature is too high: The controller activates the low-temperature cooling + tension fine-tuning mode. The unit adjusts the jet temperature (through pre-cooling or mixing in a low-temperature medium) or flow rate to create a small local supercooled zone in front of the cutting point, reducing cutting resistance by utilizing the principle of material low-temperature brittleness. At the same time, the tension is slightly pre-adjusted according to the temperature change trend to prepare for the upcoming resistance change. Specifically, when the fluid pressure signal P(t) at the cutting point is within the preset target range [Pmin(h,vf), Pmax(h,vf)], but the temperature detection value is higher than the ideal temperature threshold for the corresponding process stage, the controller activates the low-temperature cooling and tension fine-tuning mode. The specific implementation method is as follows: The intelligent jet cooling and sensing unit actuator receives instructions from the central controller and adjusts the jet parameters through one or a combination of the following two methods: Pre-cooling regulation: A plate heat exchanger is added to the coolant supply line, with its hot end connected to an independent refrigeration cycle system, including a compressor, condenser, and expansion valve, and its cold end exchanging heat with the coolant line; the central controller adjusts the temperature deviation ΔT based on the temperature. temp The compressor frequency of the refrigeration cycle system is dynamically adjusted to pre-cool the coolant to 5-15°C below the target temperature before it enters the nozzle array, forming a low-temperature jet; where ΔT temp = Tactual - T target T actual For real-time temperature, T target The target temperature corresponds to the cutting depth h(t) and feed rate vf(t).

[0056] Cryogenic medium mixing: A dual-medium mixing chamber is installed in the coolant storage tank. The main coolant (such as deionized water) and the cryogenic auxiliary medium (such as liquid nitrogen or Freon) are mixed proportionally via an electromagnetic proportional valve; the central controller adjusts the mixing ratio based on ΔT. temp Calculate the required proportion of cryogenic medium to be mixed in, and instruct the electromagnetic proportional valve to adjust the opening so that the temperature of the mixed coolant meets the local subcooling requirements. A high-precision nozzle array injects regulated low-temperature coolant into the cutting slit at a specified angle (15°~45° to the cutting direction), creating a tiny localized supercooled zone with a diameter of 2~5mm 1~3mm in front of the cutting point. Utilizing the low-temperature brittleness of silicon in the temperature range of -50℃ to 0℃, the material removal mode in the cutting area changes from mainly plastic deformation to mainly brittle fracture, reducing cutting resistance by 20%~40%. The central controller uses a temperature change trend prediction module to predict temperature deviation ΔT. temp By performing differentiation, the rate of temperature change is obtained. Among them, a first-order inertial element model is adopted. The time constant τ is determined by fitting historical data, and the gain K is related to the cutting parameter.

[0057] according to The sign and magnitude are determined by the central controller through an improved incremental control algorithm (the algorithm structure is consistent with the above-mentioned tension compensation calculation logic, but the proportional gain Kp is different). temp Integral gain Ki temp Differential gain Kd temp (Individual setting) Calculate the tension pre-adjustment amount ΔT pre : when When ΔT > 0 (temperature continues to rise), pre Take a positive value to increase the wire mesh tension in advance to resist the increasing cutting resistance; when When ΔT < 0 (temperature continues to decrease), pre Taking a negative value reduces the wire tension in advance to avoid excessive constraint leading to wire breakage. Tension pre-adjustment amount ΔT pre According to the dynamic allocation coefficient α temp (α) temp= 0.6~1.0, adjusted according to the cutting stage: 0.6~0.8 in the early stage of cutting, 0.9~1.0 in the middle stage of cutting, and 0.6~0.7 in the later stage of cutting) and distributed to the inlet tension regulator and outlet tension regulator to achieve micro-pre-adjustment of the wire mesh tension.

[0058] After the low temperature cooling and tension fine-tuning mode is activated, the central controller continuously monitors the fluid pressure P and temperature T at the cutting point. When the temperature drops below the target threshold or the pressure P deviates from the target range, the mode is automatically exited and switched to the normal control mode. During operation in low-temperature cooling and tension fine-tuning modes, if a risk signal of wire breakage is detected, such as the wire network vibration frequency exceeding the threshold or the tension fluctuation amplitude exceeding the set value, the central controller will immediately suspend the mixing of low-temperature medium or pre-cooling adjustment to prioritize the stability of the wire network.

[0059] If the coil is abnormal but P is normal: it may be caused by the shape of the silicon rod or external vibration. In this case, tension adjustment is mainly based on visual feedback, but the adjustment amplitude and timing will still refer to the P signal to ensure that the tension adjustment will not interfere with the established stable hydrodynamic pressure state.

[0060] Specifically, when the detected value of the bow state deviates from the preset safety range but the fluid pressure signal P(t) at the cutting point is within the target pressure window [Pmin(h,vf), Pmax(h,vf)], the system executes the vision-pressure coordinated tension adjustment mode, which is implemented as follows: The wire mesh spatial attitude data is acquired in real time by a high-speed CCD camera, and the arc curvature radius R(t) is extracted by a sub-pixel level edge detection algorithm. The central controller compares the bow curvature radius R(t) with a preset safety threshold [R]. min , R max In contrast, when R(t) < R min Or R(t)>R max When this occurs, it is determined to be an abnormal state of the bow line; The fluid pressure signal P(t) collected by the intelligent jet unit is acquired synchronously. The signal noise is eliminated by Kalman filtering algorithm (state transition matrix A=[1 Δt;0 1], observation matrix H=[1 0]) to verify whether P(t) is continuously within the target pressure window.

[0061] When the abnormal duration of the bow is t abnormal When the value is ≤0.2s, the central controller determines it as a transient disturbance and prioritizes maintaining the current tension setpoint T0; If 0.2s < t abnormal ≤1s, initiate visual feedback-driven tension adjustment, based on the bow curvature deviation ΔR(t) = R(t) - R target (R) target(where ΔR(t) is the ideal arc curvature radius), and the fuzzy control algorithm is used (input variables ΔR(t) and dΔR / dt, output variable ΔT). visual ) Calculate the visual accommodation amount ΔT visual The fuzzy set is divided into {negative large, negative small, zero, positive small, positive large}, and the membership function adopts the trigonometric function; When t abnormal When the change rate of the arc curvature is greater than 1 s or |dΔR / dt| > 0.5 mm / s, the fluid pressure signal P(t) is introduced as a constraint to construct a multi-objective optimization function:

[0062] in, =0.6、 =0.3、 =0.1 is the weighting coefficient, ΔP(t)=P(t)-P center P center The target pressure window value; The optimization function was solved using a particle swarm optimization algorithm (particle number N=20, inertia weight ω=0.729, learning factor c1=c2=1.494) to obtain the cooperative adjustment amount ΔT. synergy Its value range is limited to ±5%T0.

[0063] A tension adjustment delay compensation mechanism is configured to dynamically adjust the response timing based on the type of pantograph anomaly (inlet / outlet / overall). When the anomaly originates from the shape deviation of the silicon rod, such as when the radial runout of the silicon rod is detected by the laser displacement sensor as >0.1mm, a lead amount Δt_lead=0.5Δt is added to the tension regulator at the inlet end, where Δt is the system response delay time. When the abnormality is caused by external vibration, such as when the vibration frequency is detected by the acceleration sensor to be >50Hz, a damping coefficient K_damp=0.8K_original is applied to the tension regulator at the outlet end, where K_original is the original tension regulation gain; All tension adjustment commands are synchronously sent to the inlet / outlet tension regulators via EtherCAT industrial Ethernet to ensure the timing consistency of tension adjustments at both ends.

[0064] In the vision-pressure coordinated regulation mode, the fluctuation range of the fluid pressure signal P(t) is continuously monitored. When P(t) exceeds the target pressure window for three consecutive sampling cycles, the mode is immediately terminated and switched to the pressure-dominated emergency regulation mode. After each adjustment, the curvature radius R(t), fluid pressure P(t), and tension setpoint T(t) are recorded to the process database for subsequent self-optimization learning of cutting parameters. An LSTM neural network model is used with 32 hidden layer nodes and 1000 training cycles.

[0065] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0066] The above embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-substantial changes and substitutions made by those skilled in the art based on the present invention shall fall within the scope of protection claimed by the present invention.

Claims

1. A low-loss slicing device for single-crystal silicon wafers based on fluid pressure and thermal field control, characterized in that, include: The intelligent jet cooling and sensing unit is equipped with a high-precision nozzle array and a miniature pressure sensor integrated at the nozzle outlet or on the path close to the cutting point. This sensor is used to detect the dynamic fluid pressure formed near the cutting point after the coolant is injected into the cutting slit in real time, so as to reflect the silicon powder discharge, the cutting friction state, and the coolant penetration depth. The central controller employs an improved incremental control algorithm, using the cutting point fluid pressure signal P(t)(t) collected in real time by the intelligent jet cooling and sensing unit as the core feedback quantity. Based on the current cutting depth h(t) and feed rate vf(t), it determines the target pressure window [Pmin(h,vf), Pmax(h,vf)] through a built-in process database or empirical formulas, and defines the real-time pressure deviation e(t) = P(t). Pt (h,vf), where Pt can be the center value of the window or dynamically adjusted according to the process stage; the tension compensation amount ΔT(t) acting on the tension regulator at the outlet end or the tension regulator at the inlet end is dynamically calculated. The tension compensation amount ΔT(t) consists of a proportional term, an integral term and a differential term, and a nonlinear gain factor is introduced to adapt to the strong nonlinearity of the cutting process. The tension regulator, including the inlet tension regulator and the outlet tension regulator, adjusts the tension according to the tension compensation amount ΔT(t) calculated by the central controller to achieve instantaneous matching between the cutting resistance and the wire mesh support force; The actuator, according to the instructions of the central controller, adjusts the jet parameters of the intelligent jet cooling and sensing unit. The jet parameters include at least jet pressure, angle, temperature or flow rate.

2. The apparatus according to claim 1, characterized in that: The tension compensation amount ΔT(t) consists of a proportional term, an integral term, and a differential term, and is expressed by the following formula: Among them, the proportional term adopts In form, To determine the real-time pressure deviation, a piecewise proportional gain method is used. Value: When the absolute value of the pressure deviation | When | is less than the preset first threshold, take = When the absolute value of the pressure deviation is | When | is greater than the preset second threshold, take = When the absolute value of the pressure deviation is | | When the value is in the intermediate region between the first and second thresholds, linear interpolation is used to determine the value. The value of makes the proportional gain at and Smooth transition between them; The integral term is used to eliminate steady-state error by adjusting the pressure deviation. The integral is obtained by performing an integral operation; the integral limit is set, that is, the output value of the integral term is restricted so that it does not exceed the preset maximum and minimum values, and the integral coefficient of the integral term is reduced or the integral operation is paused at the beginning and end of the cutting stage, depending on the actual cutting situation. The differential term is used to reflect the pressure deviation. The pressure change trend, through pressure deviation The result is obtained by performing differentiation; a first-order low-pass filter is connected in series before the differentiation operation to filter the pressure signal. The time constant of the first-order low-pass filter is... T f Selected based on sampling frequency and noise level; The calculated proportional, integral, and differential terms are added together to obtain the final tension compensation amount. .

3. The apparatus according to claim 1, characterized in that, In a scenario with independent control of two regulators, the central controller's allocation strategy for the total tension compensation ΔT(t) includes: The total compensation amount ΔT(t) is proportionally allocated to the inlet tension regulator and the outlet tension regulator, as expressed as: Inlet tension compensation amount: Outlet tension compensation: .

4. The apparatus according to claim 3, characterized in that, The allocation coefficient α is dynamically adjusted according to the cutting stage, and the adjustment rule is as follows: In the early stage of cutting, α>1, the focus is on tension adjustment at the inlet end, using tension compensation at the inlet end to suppress the initial wire bow and ensure stability in the initial stage of cutting; in the middle stage of cutting, α=1, the inlet and outlet ends are balanced to keep the wire mesh in a stable tension state during the cutting process. In the later stages of cutting, when α < 1, the focus is on adjusting the tension at the exit end to compensate for the tension changes caused by the weakening of the silicon rod material support.

5. The apparatus according to claim 1, characterized in that, The actuator includes a jet parameter adjustment component group, and the specific implementation methods include: The jet pressure regulating component adopts a combination structure of a high-pressure proportional pump and an electro-hydraulic proportional pressure valve. The input end of the high-pressure proportional pump is connected to the coolant storage tank, and the output end of the high-pressure proportional pump is connected to the high-precision nozzle array through a pipeline. The electro-hydraulic proportional pressure valve is installed on the pipeline between the pump and the nozzle. The jet angle adjustment component includes an electric rotary nozzle holder and an angle sensor. A high-precision nozzle array is fixedly installed on the electric rotary nozzle holder, and the angle sensor provides real-time feedback on the angle between the nozzle outlet and the cutting line feed direction. The jet temperature regulating component consists of a coolant precooling heat exchanger, a cryogenic medium storage tank, and a three-way proportional mixing valve. The precooling heat exchanger precools the coolant to 0~5℃ through circulating cooling water. The cryogenic medium storage tank stores liquid nitrogen or ethylene glycol cryogenic medium. The three ports of the three-way proportional mixing valve are respectively connected to the precooling coolant pipeline, the ambient temperature coolant pipeline, and the cryogenic medium pipeline. The jet flow rate regulating component uses a frequency-controlled magnetic circulation pump, which is installed on the coolant return line to regulate the coolant circulation flow rate.

6. The apparatus according to claim 5, characterized in that: The jet pressure regulating component is used to receive pressure regulating commands from the central controller. By changing the output pressure of the proportional pump or the opening of the proportional valve, the pressure of the coolant injected into the cutting slit is adjusted in real time to match the coolant pressure with the fluid pressure target window [Pmin, Pmax] at the cutting point. When the pressure deviation e(t) is large, the controller commands the proportional pump to increase the pressure or the proportional valve to increase the opening to enhance the penetration ability of the coolant into the cutting slit and discharge silicon powder. The jet angle adjustment component is used to receive the angle adjustment command from the central controller. It drives the nozzle seat to rotate around the vertical axis of the cutting line through the electric drive mechanism, and adjusts the relative angle between the nozzle outlet and the cutting point. When the cutting depth h(t) increases, the controller commands the nozzle seat to tilt deeper into the cutting slit. The three-way proportional mixing valve is used to receive temperature regulation commands from the central controller and mix precooling coolant and low-temperature medium in a set ratio, so that the temperature of the coolant injected into the cutting point drops to the target low-temperature range, forming a local supercooled zone in front of the cutting point, and reducing cutting resistance by utilizing the low-temperature brittleness of the material. The magnetic circulation pump is used to receive flow regulation commands from the central controller. By changing the motor speed, it adjusts the flow rate of coolant injected into the cutting slit per unit time. When the pressure deviation e(t) is positive, the controller commands the pump to speed up and increase the flow rate, thereby enhancing the flushing effect of the coolant on the silicon powder.

7. The apparatus according to any one of claims 1 to 6, characterized in that, To address the strong nonlinear characteristics of the cutting process, a nonlinear gain factor is introduced to optimize the calculation of the tension compensation amount ΔT(t), including: Define nonlinear gain factor function This function uses real-time pressure deviation. e ( t ) and its rate of change As input variables, output a gain coefficient used to adjust the tension compensation amount; The nonlinear gain factor function adopts a piecewise nonlinear mapping relationship and dynamically adjusts the gain coefficient according to the different ranges of pressure deviation and its rate of change during the cutting process. In tension compensation amount Δ T ( t Based on the calculation, a nonlinear gain factor is introduced. G After weighting and adjusting the proportional, integral, and differential terms, the optimized formula for calculating the tension compensation is as follows: in, For segmented proportional gain, For integral gain, This is the differential gain; The central controller uses real-time collected pressure deviation data. e ( t ) and its rate of change According to the preset nonlinear gain factor function Dynamically update nonlinear gain factor G In each control cycle Ts Inside, the central controller recalculates the nonlinear gain factor. G And apply it to the calculation of tension compensation for the current cycle.

8. A low-loss slicing method for single-crystal silicon wafers based on fluid pressure and thermal field control, characterized in that, Includes the following steps: Real-time monitoring of fluid pressure signals near the cutting point; using a miniature pressure sensor integrated in the intelligent jet cooling and sensing unit to sense the dynamic fluid pressure formed near the cutting point after the coolant is injected into the cutting slit. To determine whether the fluid pressure signal deviates from the preset target range, the target pressure window [Pmin(h,vf), Pmax(h,vf)] is determined based on the current cutting depth h(t) and feed rate vf(t) using the built-in process database or empirical formulas. This window serves as the preset target range, and the real-time pressure deviation e(t) = P(t) is defined. Pt(h,vf) is used to determine whether e(t) is within a reasonable range. When the fluid pressure signal deviates from the preset target range, perform the following operations simultaneously: Adjusting the jet parameters of the intelligent jet cooling unit: If the P value is too high, it indicates that the silicon powder discharge is obstructed and friction is aggravated. According to the instructions of the central controller, the jet pressure is increased instantaneously and / or the angle is adjusted to open the chip removal channel and reduce the P value. If the P value is too low, it indicates that the jet has failed to effectively enter the cutting point or the cutting is in an abnormally easy state. The central controller will work together to reduce the tension and adjust the jet to ensure effective energy input. The central controller starts the high-pressure chip removal and tension following mode, increases the jet pressure, and directionally impacts the cutting point to flush away the silicon powder. Based on the deviation in fluid pressure, the compensation amount for the inlet and / or outlet tension regulators is calculated and executed in real time: The central controller treats the increase or decrease of the P value as a signal of change in cutting resistance and immediately instructs the outlet tension regulator to perform a micro-incremental tension compensation proportional to the change in the P value; according to the dynamic adjustment of the tension distribution coefficient α in the cutting stage, the total compensation amount ΔT(t) is distributed to the inlet and outlet ends.

9. The method according to claim 8, characterized in that: When the fluid pressure signal P(t) at the cutting point is within the preset target range [Pmin(h,vf), Pmax(h,vf)], but the temperature detection value is higher than the ideal temperature threshold for the corresponding process stage, the controller activates the low-temperature cooling and tension fine-tuning mode. The actuator receives instructions from the central controller and adjusts the jet parameters through one or a combination of the following two methods: Pre-cooling regulation: A plate heat exchanger is added to the coolant supply pipeline, with its hot end connected to an independent refrigeration cycle system and its cold end exchanging heat with the coolant pipeline; the central controller adjusts the temperature deviation ΔT according to the temperature. temp The compressor frequency of the refrigeration cycle system is dynamically adjusted so that the coolant temperature is pre-cooled to 5~15°C below the target temperature before entering the nozzle array, forming a low-temperature jet. Cryogenic medium mixing: A dual-medium mixing chamber is set in the coolant storage tank, and the main coolant and the cryogenic auxiliary medium are mixed in proportion through an electromagnetic proportional valve; The central controller is based on ΔT temp Calculate the required proportion of cryogenic medium to be mixed in, and instruct the electromagnetic proportional valve to adjust the opening so that the temperature of the mixed coolant meets the local subcooling requirements. The central controller uses a temperature change trend prediction module to predict temperature deviation ΔT. temp By performing differentiation, the rate of temperature change is obtained. ; according to The sign and size are determined by the central controller, which calculates the tension pre-adjustment amount ΔT using an improved incremental control algorithm. pre : when When ΔT > 0, pre Take a positive value to increase the wire mesh tension in advance to resist the increasing cutting resistance; when When <0, ΔT pre Taking a negative value reduces the wire tension in advance to avoid excessive constraint leading to wire breakage. Tension pre-adjustment amount ΔT pre According to the dynamic allocation coefficient α temp The tension is distributed to the inlet and outlet tension regulators to achieve a small pre-adjustment of the wire tension.

10. The method according to claim 8, characterized in that: When the detected value of the bow state deviates from the preset safety range but the fluid pressure signal P(t) at the cutting point is within the target pressure window [Pmin(h,vf), Pmax(h,vf)], the spatial attitude data of the wire mesh is collected in real time by a high-speed CCD camera, and the bow curvature radius R(t) is extracted by a sub-pixel level edge detection algorithm. The central controller compares the bow curvature radius R(t) with a preset safety threshold [R]. min , R max In contrast, when R(t) < R min Or R(t)>R max When this occurs, it is determined to be an abnormal state of the bow line; The fluid pressure signal P(t) collected by the intelligent jet unit is acquired synchronously, and the signal noise is eliminated by Kalman filtering algorithm to verify whether P(t) is continuously within the target pressure window; When the abnormal duration of the bow is t abnormal When the value is ≤0.2s, the central controller determines it as a transient disturbance and prioritizes maintaining the current tension setpoint T0; If 0.2s < t abnormal ≤1s, initiate visual feedback-driven tension adjustment, based on the bow curvature deviation ΔR(t) = R(t) - R target Calculate the visual accommodation amount ΔT visual , where R target Given the ideal arc curvature radius, the input variables ΔR(t) and dΔR / dt are determined using a fuzzy control algorithm, and the output variable ΔT is determined. visual The fuzzy set is divided into {negative large, negative small, zero, positive small, positive large}, and the membership function adopts the trigonometric function; When t abnormal When the change rate of the arc curvature is greater than 1 s or |dΔR / dt| > 0.5 mm / s, the fluid pressure signal P(t) is introduced as a constraint to construct a multi-objective optimization function: in, , , For the weighting coefficients, ΔP(t) = P(t) - P center P center The target pressure window value; The optimization function is solved using the particle swarm optimization algorithm to obtain the cooperative adjustment amount ΔT. synergy .

Images