A high-density molybdenum strip sintering temperature field self-adaptive compensation method and system

Abstract

The present application relates to the technical field of powder metallurgy process control, and specifically discloses a high-density molybdenum strip sintering temperature field self-adaptive compensation method and system. The present application first acquires the physical property data of the original molybdenum strip to generate a multi-dimensional feedforward compensation reference vector, analyzes the vector to construct a heat absorption characteristic prediction package representing the heat absorption characteristics of the micro-section heat absorption body, and performs a forward-looking prediction operation to plan a bottom-layer feedforward power instruction queue. At the same time, the sintering zone space-time temperature flow field is synchronously collected to extract a dynamic real-time deviation correction signal, and the dynamic phase and pulse density modulation parameters for responding to high-frequency disturbances are calculated. Finally, the feedforward fundamental frequency waveform and the feedback phase modulation mechanism are fused to generate a composite control instruction stream to drive the output of the predetermined correction heat of the execution element. The present application combines feedforward prediction with real-time feedback to compensate for the material heat absorption difference and external transient disturbance, thereby improving the density and uniformity of the finished molybdenum strip.

Description

Technical Field

[0001] This invention belongs to the field of powder metallurgy process control technology, and relates to an adaptive compensation method and system for the sintering temperature field of high-density molybdenum strips. Background Technology

[0002] In the production of high-density molybdenum materials, sintering is a crucial step determining the final product's performance. The core challenge lies in precisely controlling the temperature field within the high-temperature sintering furnace to ensure that the raw molybdenum bars receive uniform and sufficient energy throughout the heating process, achieving ideal densification. However, as pressed powder blanks, the raw molybdenum bars typically exhibit unavoidable slight fluctuations in their physical properties along their length, such as diameter, initial density, and surface condition. These inhomogeneities directly lead to differences in the thermophysical properties of different segments of the molybdenum bars during sintering, particularly in heat absorption and conduction capabilities, posing a fundamental challenge to achieving a uniform sintered structure.

[0003] Currently, the sintering temperature control schemes commonly used in the industry mainly rely on zoned constant temperature control or PID regulation mechanisms based on single-point temperature feedback. Zoned constant temperature control divides the sintering furnace into multiple temperature zones and sets a fixed power output for each zone to maintain the average temperature of that zone. The PID feedback-based scheme, on the other hand, uses thermocouples or infrared thermometers placed at specific locations within the furnace to monitor the surface temperature of the molybdenum strips in real time. The measured values ​​are compared with a preset target temperature, and the heating power is adjusted using a PID algorithm to eliminate temperature differences. This feedback control can, to some extent, cope with slow disturbances in the external environment and is a widely used technique in current industrial production.

[0004] While the aforementioned traditional methods can achieve some control over the sintering process, their limitations are quite prominent when pursuing higher product consistency. Zoned constant temperature control is an open-loop control strategy; it cannot detect and compensate for local differences in the physical properties of the original molybdenum strip itself, and is powerless to address changes in heat absorption capacity caused by fluctuations in diameter or density, potentially leading to over- or under-sintering in localized areas. Although PID control based on single-point feedback introduces closed-loop regulation, it is essentially a lag control. Due to the significant thermal inertia of the sintering system, when a temperature deviation is detected at the measuring point, the corresponding small segment of the molybdenum strip may have already moved from its optimal heating position. This lag in adjustment makes real-time and precise temperature control difficult, especially in continuous sintering processes with high feed rates, easily leading to temperature overshoot or oscillation. Existing technical solution A, such as CN114669758B, proposes a method for online temperature field control, but it is mainly applied to additive manufacturing scenarios such as powder bed melting. Its control object and disturbance source are different from those of continuous feed molybdenum strip sintering. Its solution still has room for improvement in dealing with the differences in heat absorption caused by the fluctuation of the continuous physical characteristics of the material itself.

[0005] Based on the above problems, the present invention aims to solve the problem that the uneven physical properties of the original molybdenum strips lead to differences in the heat absorption capacity of each micro segment during the sintering process, and that traditional control methods cannot achieve forward-looking, localized, and precise heat compensation, thus causing uneven density of the finished product. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background art, an adaptive compensation method and system for the sintering temperature field of high-density molybdenum strips is proposed.

[0007] The objective of this invention can be achieved through the following technical solution: The first aspect of this invention provides an adaptive compensation method for the sintering temperature field of high-density molybdenum strips, comprising: S1, acquiring physical characteristic data of the original molybdenum strips and generating a multidimensional feedforward compensation reference vector containing continuous spatial coordinates.

[0008] S2. Analyze the multidimensional feedforward compensation benchmark vector to construct a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip.

[0009] S3. Based on the thermal absorption characteristics prediction package, perform forward prediction matching operation and plan and generate the underlying feedforward power instruction queue based on the instantaneous power balance model.

[0010] S4. Synchronously collect spatiotemporal temperature and flow field information within the sintering area and extract the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism.

[0011] S5. Process dynamic real-time deviation correction signals and analyze and calculate dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters.

[0012] S6. The integrated feedforward fundamental frequency waveform locking and feedback phase modulation mechanism generates a composite control command stream, executes local compensation actions, and outputs high-density finished molybdenum bars.

[0013] The second aspect of the present invention provides an adaptive compensation system for the sintering temperature field of high-density molybdenum strips, comprising: a physical property data acquisition module, which acquires the physical property data of the original molybdenum strips being fed in, and generates a multidimensional feedforward compensation reference vector containing continuous spatial coordinates.

[0014] The thermal absorption characteristic analysis module analyzes the multi-dimensional feedforward compensation benchmark vector and constructs a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip.

[0015] The feedforward power planning module performs forward prediction matching operations based on the thermal absorption characteristic prediction package, and plans and generates the underlying feedforward power instruction queue based on the instantaneous power balance model.

[0016] The real-time deviation correction signal extraction module synchronously collects spatiotemporal temperature and flow field information within the sintering area and extracts the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism.

[0017] The dynamic modulation parameter calculation module processes dynamic real-time deviation correction signals and analyzes and calculates dynamic phase and pulse density modulation parameters to cope with high-frequency disturbances.

[0018] The composite instruction fusion and execution module integrates feedforward base frequency waveform locking and feedback phase modulation mechanisms to generate a composite control instruction stream, executes local compensation actions, and outputs high-density finished molybdenum bars.

[0019] Compared to existing technologies, the embodiments of the present invention have at least the following advantages or beneficial effects: By setting a non-contact physical property scanning unit at the sintering furnace inlet and acquiring continuous physical property data of the original molybdenum strip along its length before it enters the heating zone, the present invention can calculate the real-time pre-sintering density, surface emissivity, and other core thermophysical parameters of each micro-segment based on this data, and construct a thermal absorption characteristic prediction package containing these parameters. This method transforms the basis of control from lagging temperature measurements to proactive physical quantities characterizing the inherent properties of the material, enabling the control system to know the unique heat-absorbing properties of any micro-segment of the molybdenum strip before it enters the heating zone. This mechanism achieves root-cause prediction of sintering temperature fluctuations caused by the material's inherent inhomogeneity, thus providing a possibility for eliminating these intrinsic disturbances.

[0020] This invention, based on a constructed thermal absorption characteristic prediction package, uses heat transfer theory to solve equations to proactively calculate the theoretical total energy injection required for each micro-segment about to enter the heating zone to reach the target sintering temperature, and converts this into a basic feedforward power command. This feedforward mechanism can perform differentiated energy planning based on the volumetric heat capacity and radiation absorption coefficient of each micro-segment; for example, for micro-segments with larger diameters or higher densities, the system automatically plans a higher basic power command. This energy supply method based on the differentiated physical characteristics of micro-segments achieves personalized and localized precise distribution of sintering energy, fundamentally compensating for the endogenous thermal disturbances introduced by fluctuations in the physical characteristics of the original molybdenum strip, and helping to maintain a more consistent temperature distribution throughout the sintering process.

[0021] This invention employs a composite control strategy integrating feedforward control and feedback regulation. While the feedforward command provides a precise baseline for the main energy supply, a dynamic modulation mechanism based on real-time temperature deviation gradients is introduced. This mechanism can capture high-frequency, subtle temperature disturbances caused by unpredictable factors such as external environmental fluctuations by acquiring the actual sintering temperature flow at high frequencies, and instantly calculate microsecond-level fine-tuning phase offsets and pulse density disturbance values. These fine-tuning parameters are superimposed on the fundamental frequency waveform determined by the feedforward, achieving rapid response and precise suppression of disturbances by finely shaping the edges and internal structure of the power output pulse. This dual-channel control architecture enables the system to cope with both the macroscopic inhomogeneities of the material itself and transient disturbances during operation, enhancing the stability and robustness of temperature control. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0024] Figure 2 This is a distribution diagram of the instantaneous feedforward power planning and fundamental frequency duty cycle curves of the present invention.

[0025] Figure 3 This is a graph showing the monitoring and analysis of local abnormal temperature deviations and change gradients in this invention.

[0026] Figure 4 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 The first aspect of the present invention provides an adaptive compensation method for the sintering temperature field of high-density molybdenum strips, comprising: S1, acquiring physical characteristic data of the original molybdenum strips being fed, and generating a multidimensional feedforward compensation reference vector containing continuous spatial coordinates.

[0029] In a specific embodiment of the present invention, the physical characteristic data of the original molybdenum bar is acquired, and a multi-dimensional feedforward compensation reference vector containing continuous spatial coordinates is generated, including: activating a non-contact physical characteristic scanning unit to synchronously sense the continuous optical and geometric physical characteristics of the original molybdenum bar as it moves into the sintering furnace with the feeding mechanism.

[0030] A laser diameter measuring instrument was used to collect continuous diameter data streams of the original molybdenum strip along its length, and a multispectral analyzer was used to obtain the corresponding surface reflection and emission spectral streams.

[0031] By fusing continuous diameter data streams with surface reflection and emission spectral streams, a multidimensional feedforward compensation reference vector corresponding to the feed position is calculated.

[0032] Specifically, the first implementation step S1 of the method of the present invention involves acquiring the physical characteristic data of the original molybdenum bar being fed and generating a multidimensional feedforward compensation reference vector containing continuous spatial coordinates. This step is executed by a non-contact physical characteristic scanning unit deployed at the sintering furnace inlet under the command of the central control system. When the feeding mechanism is activated and the original molybdenum bar moves stably into the sintering furnace at a preset feed speed vector, the position encoder linked to the feeding mechanism begins to generate a continuous spatial coordinate signal that strictly corresponds to the longitudinal displacement of the original molybdenum bar. The laser diameter gauge and the multispectral analyzer in the non-contact physical characteristic scanning unit are synchronously activated and aligned with the same cross-section of the original molybdenum bar. The laser diameter gauge, such as a dual-axis scanning laser diameter gauge, continuously emits a fan-shaped laser beam and detects the width of the shadow formed after the original molybdenum bar is blocked. Through internal circuitry, it calculates the instantaneous diameter of the original molybdenum bar at each continuous spatial coordinate point in real time, thereby forming a continuous diameter data stream along its length direction. At the same time, the multispectral analyzer uses a built-in active broadband light source to illuminate the surface of the original molybdenum bar and captures the surface reflected light and thermal emission light from the same coordinate point through a high-sensitivity spectral receiver. The captured light signal is decomposed by a dispersive element into multiple discrete wavelength channels covering the visible to near-infrared bands, generating surface reflection and emission spectral streams that are precisely synchronized in time and space with the continuous diameter data stream. Finally, the data fusion processor inside the scanning unit, such as a field-programmable gate array (FPGA), associates and binds each spatial coordinate point, its diameter value, and the corresponding reflection and emission spectral data according to the synchronized timestamp, concatenating them into a structured data vector, namely the multidimensional feedforward compensation reference vector. This vector is generated point-by-point continuously, forming a data stream that provides the original physical basis for subsequent prediction of endothermic characteristics. In the specific implementation, the structure of the multidimensional feedforward compensation reference vector can be represented as:

[0033] The non-contact physical property scanning unit is an integrated hardware module located at the sintering furnace inlet. It integrates a laser diameter gauge, a multispectral analyzer, a position encoder interface, and data fusion processing circuitry, aiming to perform a non-invasive and comprehensive characterization of the physical parameters of the raw molybdenum strip before it enters the furnace. The continuous diameter data stream is a scalar time series with the original molybdenum strip length as the independent variable, denoted as... The unit is mm. Its measurement accuracy is typically set to better than ±0.005 mm. This accuracy is based on the current technological level of commercially available high-precision laser diameter gauges and is sufficient to capture minute geometric fluctuations that cause differences in subsequent sintering endothermic processes. Surface reflection and emission spectral flow are measured for each spatial coordinate point on the original molybdenum strip. A pair of spectral data vectors acquired and Each spectral vector is composed of... It consists of the light intensity values ​​at discrete wavelength points, for example .in, The value, such as 512, is selected to ensure sufficient spectral resolution to identify the characteristic spectral lines of molybdenum and its surface oxides at different temperatures. Wavelength range to The setting, for example, from 400nm to 2500nm, is based on the fact that this range fully covers the main thermal radiation bands and spectral reflectance regions sensitive to surface conditions of molybdenum materials during the pre-sintering stage. Multidimensional feedforward compensation reference vector. It is a composite data structure, and its internal components are defined as follows: It is a one-dimensional spatial coordinate along the original molybdenum strip feed direction provided by the position encoder, in mm; This is the original diameter value collected by the laser diameter measuring instrument at this coordinate point; and These are the surface reflectance spectral vector and surface emission spectral vector collected by the multispectral analyzer at this coordinate point, respectively. Their intensity values ​​are usually normalized and are dimensionless.

[0034] For example, assume the raw molybdenum strip enters the sintering furnace at a speed of 10 mm / s. At a certain sampling moment, the current scanning position determined by the position encoder linked to the feeding mechanism is... mm. At this time, the non-contact physical characteristic scanning unit simultaneously performs data acquisition. The diameter value output after calculating the shadow width measured by the laser diameter gauge is... mm. Simultaneously, after irradiating and acquiring data at this point, the multispectral analyzer outputs spectral data covering 512 channels in the range of 400 nm to 2500 nm. For simplicity, only the normalized intensity values ​​at three key wavelengths of 650 nm, 980 nm, and 1550 nm are listed; the key components of the obtained surface reflectance spectral vector are... The key component of the surface emission spectral vector is The data fusion processor binds these data and generates a multi-dimensional feedforward compensation reference vector corresponding to that location point. This vector is immediately pushed into the input buffer queue of subsequent processing steps to calculate the real-time endothermic mass of the molybdenum strip segment.

[0035] S2. Analyze the multidimensional feedforward compensation benchmark vector to construct a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip.

[0036] In a specific embodiment of the present invention, the multidimensional feedforward compensation reference vector is analyzed to construct a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip. This includes: aligning the spatial position coordinates of the multidimensional feedforward compensation reference vector and extracting the spectral features of the corresponding micro-segments to invert the real-time pre-sintering density and real-time surface emissivity.

[0037] The cross-sectional area of ​​a small segment is calculated based on a continuous diameter data stream, and the volumetric heat capacity is dynamically calculated by combining the real-time pre-sintering density.

[0038] The real-time surface emissivity is extracted and coupled with the pre-stored spectral physical properties of the furnace heating element to calculate the effective radiation absorption coefficient of the environment.

[0039] In a specific embodiment of the present invention, the real-time surface emissivity is extracted and coupled with the pre-stored spectral physical properties of the furnace heating element to calculate the effective radiation absorption coefficient of the environment, including: based on Kirchhoff's thermal radiation law and under thermal equilibrium setting, the extracted real-time surface emissivity is equivalently converted into the basic radiation absorption rate of a small segment.

[0040] The emissivity curve of the radiation spectrum of a specific heating element is retrieved from the pre-stored data and integrated with the basic radiation absorptivity in the frequency domain to obtain the effective radiation absorptivity coefficient of the environment.

[0041] A solidified thermal absorption characteristic prediction package is constructed by encrypting the location coordinates of tiny segments with unique spatiotemporal timestamps, along with volumetric heat capacity and radiation absorption coefficient.

[0042] Specifically, the second implementation step S2 of the method of the present invention involves parsing the multidimensional feedforward compensation reference vector to construct a thermal absorption characteristic prediction package characterizing the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip. This step is executed by a dedicated data processing module, which receives the multidimensional feedforward compensation reference vector stream from step S1. For each multidimensional feedforward compensation reference vector in the data stream, the processing module first aligns it according to its continuous spatial coordinates and extracts the surface reflection and emission spectral streams corresponding to that coordinate point. These two spectral vectors are input into a pre-established physical-spectral inverse solution model, which establishes a mapping relationship between spectral features and material physical properties based on a large amount of offline calibration experimental data. Through the calculation of this model, the real-time pre-sintering density and real-time surface emissivity of the micro-segment are inverted. Subsequently, the data processing module extracts the diameter value from the continuous diameter data stream from the same multidimensional feedforward compensation reference vector and uses the diameter value to calculate the cross-sectional area of ​​the micro-segment. The processing module further calls upon the molybdenum metal specific heat capacity constant stored in the material property database and multiplies it by the previously calculated real-time pre-sintering density to dynamically calculate the volumetric heat capacity characterizing the heat storage capacity of the micro-segment. Next, using the acquired real-time surface emissivity as input, combined with pre-stored parameters describing the spectral characteristics of the furnace heating elements, and applying Kirchhoff's thermal radiation law, the effective radiation absorption coefficient of the micro-segment under the current sintering furnace environment is calculated. Finally, the data processing module structurally encapsulates the continuous spatial coordinates uniquely corresponding to the micro-segment, along with the calculated volumetric heat capacity and radiation absorption coefficient. To ensure the integrity and immutability of the data during subsequent transmission and processing, this encapsulation structure uses a preset key to calculate a hash message authentication code, generating a digital signature, which is then solidified, ultimately forming an encrypted thermal absorption characteristic prediction package containing a unique spatiotemporal stamp. The generation process of the thermal absorption characteristic prediction package is continuous, and its data stream is pushed to the downstream prediction matching calculation module in real time.

[0043] The construction process of the thermal absorption characteristic prediction package involves the following calculations: First, real-time pre-sintering density. and real-time surface emissivity Through a reverse solution function The following was obtained from the spectral data:

[0044] Secondly, volumetric heat capacity The calculation formula is:

[0045] Finally, the heat absorption characteristic prediction package The structured expression is:

[0046] Among them, real-time pre-sintering density This is the initial density, measured in g / cm³, obtained by inverting the ratio of characteristic absorption peaks or reflectance in the surface reflection and emission spectral streams using a physics-spectral inverse solution model. It characterizes the initial density of the original molybdenum strip after being pressed into this small segment. This is a dimensionless parameter describing the ability of a small surface segment to radiate heat at the sintering temperature, calculated based on the spectral intensity distribution using a physics-spectral inverse solution model. Its value ranges from 0 to 1. Volumetric heat capacity. It is the product of the real-time pre-sintered density and the inherent specific heat capacity of the molybdenum material. It quantifies the energy required to raise the temperature of a tiny segment per unit volume by 1 K, and is expressed in J / (cm³·K). Specific heat capacity As a physical constant, its value is obtained from standard material handbooks and is set based on its average value over the target sintering temperature range. Radiation absorption coefficient This is derived from Kirchhoff's law of thermal radiation, under the assumptions of thermal equilibrium and gray body, which states that the absorptivity of a material to thermal radiation equals its emissivity. This coefficient characterizes the efficiency of a small segment in absorbing radiant energy from surrounding heating elements and is a dimensionless value. (Thermal absorption characteristic prediction package...) It is a digitally signed data structure used to transmit tiny segments of core thermophysical parameters downstream. (Function) Indicates the use of a preset key Signing operations are performed on the core data within the package to ensure data integrity.

[0047] Specifically, real-time pre-sintering density characterizes the local physical compactness of the molybdenum powder after pressing, while real-time surface emissivity reflects the local surface's ability to absorb and dissipate heat. The system obtains these two values ​​not through uninterpretable black-box algorithms, but based on a clear "light-matter correlation" physical mechanism: differences in the compactness of the pressed molybdenum powder cause variations in surface microporosity, leading to a regular change in the ratio of diffuse to specular reflection (i.e., reflectivity ratio) of light at a specific wavelength; simultaneously, surface micro-roughness or slight oxidation state directly alters the overall spectral intensity distribution curve. Therefore, the so-called "physical-spectral inverse solution model" is essentially a fundamental mapping database established by engineers before system deployment, through offline optical scanning and thermal calibration of a large number of real molybdenum strip samples with known densities and surface states. This database contains a one-to-one correspondence between "spectral features and real physical parameters." In actual production millisecond-level control, the system only needs to use the detector to capture the optical signal characteristics of the molybdenum strip surface in real time, and use this as an index to compare and perform linear interpolation calculations in the aforementioned pre-set database. This will accurately "translate" the true initial density and dimensionless surface emissivity of the current tiny segment, thereby transforming the abstract optical signal into a solid physical basis for subsequent control and heat compensation.

[0048] For example, in connection with step S1, the data processing module receives a multi-dimensional feedforward compensation reference vector. The module extracts the spectral data vector. and Input these two vectors into the physics-spectrum inverse solution model. The model was pre-established by spectral calibration of molybdenum strip samples with known density and surface condition. After calculation, the model outputs the real-time pre-sintering density at that location. g / cm³, real-time surface emissivity Next, the module extracts the diameter value. mm. It retrieves the specific heat constant of molybdenum from an internal database, assuming the value is... J / (g·K). The volumetric heat capacity was then calculated as follows: At the same time, according to Based on the principle, the radiation absorption coefficient is determined to be Finally, the module will output the location coordinates. mm, volumetric heat capacity J / (cm³·K) and radiation absorption coefficient Encapsulate the data and use the key to calculate a digital signature, for example... The final generated thermal absorption characteristic prediction package is as follows: This data packet is then sent to the next step.

[0049] Please see Figure 2 S3. Based on the thermal absorption characteristics prediction package, perform forward prediction matching operation and plan and generate the underlying feedforward power instruction queue based on the instantaneous power balance model.

[0050] In a specific embodiment of the present invention, a forward-looking prediction matching operation is performed based on the thermal absorption characteristic prediction package, and a feedforward power instruction queue is generated based on the instantaneous power balance model. This includes: detecting the spatial coordinate data corresponding to the thermal absorption characteristic prediction package and mapping the physical range association model of each independent heating module in the sintering furnace.

[0051] The obtained volumetric heat capacity and radiation absorption coefficient are input into the instantaneous power equation of the heat transfer theory to calculate the dynamic heating power requirement for the micro-segment to reach the target sintering temperature. Combined with the steady-state heat dissipation constant of the module, the target total feedforward power is obtained.

[0052] The operating system on the chip uses the standard architecture of pulse width modulation to divide the total target feedforward power by the maximum rated power of the module, and converts it into the base frequency duty cycle used by the control unit.

[0053] The feed rate vector along the original molybdenum bar is arranged with the time baseline to form a fundamental frequency duty cycle sequence to construct a feedforward power command queue covering all heating zones.

[0054] Specifically, the third implementation step S3 of the method of the present invention involves performing a forward-looking prediction matching operation based on the thermal absorption characteristic prediction package, and planning and generating a feedforward power instruction queue based on the instantaneous power balance model. This step is executed by a forward-looking prediction matching operation module, which continuously receives the data stream of the thermal absorption characteristic prediction package from the upstream data processing module and stores it in a first-in-first-out buffer queue. The core task of this module is to pre-plan the required energy injection amount for each small segment of the original molybdenum bar before it enters each of the downstream independent heating modules. The execution process is as follows: when a thermal absorption characteristic prediction package, along with the physical displacement of the original molybdenum bar, is about to enter the control range of a target independent heating module, the module is triggered. First, the module detects the continuous spatial coordinates and queries the internally stored physical range association model of each zone of independent heating modules in the sintering furnace. This is a static lookup table that maps the travel coordinates of the original molybdenum bar to the specific heating module number. Once the target independent heating module is determined, the module extracts the volumetric heat capacity and radiation absorption coefficient from the thermal absorption characteristic prediction package. These two key physical parameters are used as inputs and substituted into a built-in heat transfer theory solution equation. Based on the principle of energy conservation, this equation comprehensively considers the sensible heat increment required for a small segment to heat from its initial temperature upon entering the heating module to the target sintering temperature, as well as the heat loss required to maintain the temperature during its stay within the heating module, calculating the theoretical total energy injection that must be provided by the target independent heating module. Subsequently, the on-chip system embedded in the module, according to the standard architecture of pulse width modulation, converts this scalar form of the theoretical total energy injection into a dimensionless parameter for controlling the power execution unit. Specifically, it divides the theoretical total energy injection by the maximum output energy of the target independent heating module during the same dwell time, thus obtaining a precise percentage value, i.e., the fundamental frequency duty cycle. Finally, the module binds the calculated fundamental frequency duty cycle with the corresponding target independent heating module number and execution timestamp, and, according to the time baseline determined by the original molybdenum bar feed velocity vector, inserts this bound instruction into a queue specifically maintained for the target independent heating module, thereby dynamically constructing and updating the feedforward power instruction queue covering all heating zones.

[0055] The fundamental frequency duty cycle is calculated using the following instantaneous power balance model: First, the time interval between small segments passing through a specific physical location is calculated. :

[0056] in, The feed velocity vector of the original molybdenum bar. Indicates length is The time span during which a tiny segment enters the heating zone. It defines the discrete length of a tiny segment, which is a system-preset constant, for example, set to 1 mm.

[0057] Secondly, calculate the dynamic energy that must be injected to achieve the target temperature rise in this tiny segment. :

[0058] in, It is the theoretical temperature rise energy specific to that tiny segment itself. It is the heat transfer efficiency, which is a factor related to the radiation absorption coefficient. A positively correlated function. This function, calibrated through offline experiments, reflects the difference in the ability of molybdenum strips with different surface absorptivity to receive radiant heat from the heating element. The lower the radiative absorptivity, the lower the heat transfer efficiency, and the higher the required injection energy. It is the target sintering temperature that the target independent heating module needs to achieve. It is a key process parameter, and its setting is based on the metallographic transformation temperature required for molybdenum material to obtain high density. It is usually set in the range of 1800℃ to 2100℃. It is the initial temperature when the tiny segment enters the target independent heating module, and its value is determined by the outlet temperature of the previous heating module. The corresponding coordinates in the continuous diameter data stream obtained in step S1 The diameter value.

[0059] Subsequently, the dynamic power required to meet the heating demand of this tiny segment is calculated. and the module's target total feedforward power :

[0060]

[0061] in, This characterizes the instantaneous dynamic power that the heating module must output to ensure that the continuously fed molybdenum bar heats up at a set rate. This constitutes the total feedforward power that the heating module needs to provide at a certain moment. It is the inherent power loss of the target independent heating module in maintaining the target sintering temperature during no-load steady-state operation. It is a macroscopic constant determined in advance through experiments, and its unit is W.

[0062] Finally, this is converted to the dimensionless fundamental frequency duty cycle:

[0063] Among them, the fundamental frequency duty cycle It accurately represents the basic power output level that matches the physical properties of the current minute segment. This is the rated maximum output power of the heating element, which is a hardware specification parameter of the heating element, and the unit is W.

[0064] Furthermore, to achieve simplified calculations and accurate mapping in engineering, and with the radiation absorption coefficient... Positive correlation function Essentially, it is an empirical polynomial fitted using the least squares method from offline experimental data (usually simplified to a first-order linear function expression, i.e.) Where k is the positive proportionality coefficient and c is the basic heat transfer intercept. Before the system is put into operation, technicians can easily fit the specific coefficients of this function by inputting a standard test sample with a known gradient absorption coefficient into the furnace and recording the actual energy consumed to reach the target temperature rise. On the other hand, the power loss during no-load steady-state operation... The measurement method is very intuitive: simply heat the target independent heating module to the set target sintering temperature (e.g., 1900℃) under no-molybdenum strip (no load) conditions. After the temperature field is completely stable, the sustaining power output by the power supply system is then determined. In a typical industrial-grade continuous molybdenum bar sintering furnace, for an independent heating zone with a single effective heating length of approximately 200 mm, this constant is limited by the heat dissipation characteristics of the furnace insulation material. Typical values ​​are usually in the range of 3000W to 4500W (e.g., a typical value of 3500W), which provides a real energy benchmark for feedforward power planning.

[0065] For example, in the transition step S2, the prospective prediction matching operation module receives the prediction packet. The parameters for the upcoming module 2 are: target temperature. Power loss Maximum power Molybdenum bar feed rate , length of tiny segment The initial temperature is assumed to be... First, calculate the time interval of the tiny segment. Its volume is approximately The energy required for the temperature rise of this tiny segment is According to the absorption coefficient Query efficiency function Therefore, the actual dynamic energy that needs to be injected Therefore, the dynamic heating power demand triggered by this tiny segment can be derived. At this point, the total target feedforward power that the module needs to provide is Finally, the fundamental frequency duty cycle is calculated. (i.e., 65.06%). The module generation instruction pushes a 65.06% duty cycle into the queue.

[0066] Please see Figure 3 S4. Synchronously collect spatiotemporal temperature and flow field information within the sintering area and extract the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism.

[0067] In a specific embodiment of the present invention, the spatiotemporal temperature flow field information within the sintering area is collected synchronously, and the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism is extracted, including: driving a non-contact infrared temperature sensor array deployed along the transmission path to collect the actual sintering temperature flow of the corresponding small segment within the independent heating module.

[0068] The current value of the received real sintering temperature flow is compared with the preset optimized sintering temperature target two-dimensional field at high frequency in real time to obtain the current temperature deviation value.

[0069] The system assesses and determines the gradient of changes in the current minute temperature deviation and outputs a dynamic real-time deviation correction signal that indicates local temperature fluctuations.

[0070] In a specific embodiment of the present invention, evaluating and determining the gradient of change of the current small temperature deviation value and outputting a dynamic real-time deviation correction signal indicating local temperature fluctuations includes: using a first-order time derivative algorithm to calculate the rate of change of the current small temperature deviation value over time to obtain the temperature deviation change gradient.

[0071] The temperature deviation gradient is compared with the set dead zone threshold. If the absolute value exceeds the dead zone threshold, a dynamic real-time deviation correction signal containing amplitude and sign information is generated.

[0072] Specifically, the fourth implementation step S4 of the method of the present invention involves synchronously acquiring spatiotemporal temperature flow field information within the sintering area and extracting the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism. This step is executed in parallel with the aforementioned feedforward control path and is completed by a temperature monitoring and processing system deployed inside the sintering furnace. When a small segment of the original molybdenum strip physically enters a target independent heating module under the drive of the feeding mechanism, a non-contact infrared temperature sensor array deployed along the inside of the heating module and parallel to the original molybdenum strip's transmission path is activated. This array consists of multiple point or line infrared probes, whose field of view precisely covers the surface area of ​​the original molybdenum strip. At a preset high sampling frequency, such as 100Hz, the sensor array continuously acquires the infrared radiation intensity at different points on the surface of the small segment as it moves through the heating module, and uses a built-in calibration curve to convert it into a series of discrete real sintering temperature values, thereby forming a real sintering temperature flow. At the same time, the control system retrieves the optimized sintering temperature target two-dimensional field corresponding to the independent heating module from the process database. This two-dimensional field represents a preset ideal temperature distribution, defining not only the target temperature curve along the feed direction but also potentially including lateral temperature distribution requirements. The temperature processing unit performs a point-to-point, high-frequency real-time comparison between the received current value of the actual sintering temperature flow and the target temperature value at the same spatial coordinate point in the optimized sintering temperature target two-dimensional field. This comparison is essentially a subtraction operation, the result of which is the current temperature deviation value. Finally, the system further evaluates and judges a series of continuously acquired current temperature deviation values. This process is achieved by calculating its first-order time derivative, i.e., determining the gradient of the deviation value. If this gradient exceeds a preset dead zone threshold, indicating non-stationary fluctuations in the local temperature, the system will immediately output a non-zero dynamic real-time deviation correction signal containing information on the deviation amplitude and trend. This signal will be used as input to the downstream dynamic modulation engine to trigger subtle feedback adjustment actions.

[0073] The non-contact infrared temperature sensor array consists of one or more infrared focal plane arrays or high-speed scanning point thermometers. Its operating wavelength is typically selected in the near-infrared band, which is insensitive to molybdenum oxide, such as 1.6 μm, to reduce measurement errors caused by changes in surface condition. Its spatial resolution is designed to be able to resolve discrete segments of at least a tiny length. Temperature changes. The target two-dimensional field for optimizing sintering temperature is a static data matrix. ,in Let y be the coordinate along the feed direction and y be the horizontal coordinate. For elongated molybdenum bars, this is usually simplified to a one-dimensional target temperature curve. The design is based on ensuring that molybdenum powder particles complete the full physicochemical processes of preheating, debinding, sintering shrinkage, and grain growth in different temperature zones to achieve the best densification effect.

[0074] Current temperature slight deviation value The calculation formula is:

[0075] in It is the sensor array at any time ,Location The instantaneous value of the actual sintering temperature flow collected.

[0076] Dynamic real-time deviation correction signal It is generated after assessing the gradient of changes in a small deviation from the current temperature. Its generation logic is as follows:

[0077] in, This is the gradient dead zone threshold, a parameter set to prevent the system from overreacting to background noise. It is based on statistical analysis of natural temperature fluctuations during normal steady-state sintering and is typically set to a minimum value, such as 0.1℃ / s. Only when the rate of change of temperature deviation exceeds this threshold is a localized temperature fluctuation considered to require immediate correction.

[0078] For example, in the example of connecting step S3, a small segment carrying a 65.06% baseband duty cycle command is in The system immediately enters independent heating module number 2. Assume the target temperature curve for this module is... The temperature was set to a constant 1900℃ within this region. At that moment, the tiny segment moved to a certain position within module 2. The actual sintering temperature value collected by the infrared temperature sensor located above this position is... ℃. The control system immediately performs a comparison calculation to obtain the current minute temperature deviation value. ℃. At the next sampling time The temperature value collected by the sensor is ℃, the corresponding deviation value is ℃. The system evaluates and determines the gradient of change of the current small temperature deviation value, and calculates it as follows: Assuming a set gradient dead zone threshold... ,because The system determines that this is a significant temperature increase trend. Therefore, the system immediately outputs a dynamic real-time deviation correction signal. The amplitude of this signal is +30. The positive sign indicates that the temperature is rising, and the value quantifies the severity of this rise. This signal will be immediately sent to step S5 for processing.

[0079] S5. Process dynamic real-time deviation correction signals and analyze and calculate dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters.

[0080] In a specific embodiment of the present invention, processing the dynamic real-time deviation correction signal and analyzing and calculating the dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters includes: reading the dynamic real-time deviation correction signal into the dynamic modulation engine, and substituting the dynamic real-time deviation correction signal into the extended trend prediction model configured based on the proportional-integral mechanism for calculation.

[0081] Based on the energy adjustment command calculated by the model, a microsecond-level time offset value is determined for advancing or delaying the edge of the power pulse, and a fine-tuning phase offset is generated to smooth out thermal response hysteresis.

[0082] The sudden increase in heat loss is quantified by using dynamic real-time deviation correction signals and directly calculated into pulse density perturbation values ​​with microsecond time widths to directly adjust the total energy within the cycle.

[0083] The aggregated fine-tuning phase offset and pulse density perturbation values ​​are integrated according to the communication interaction protocol to establish a precise correction modulation parameter sub-format.

[0084] Specifically, the fifth implementation step S5 of the method of the present invention involves processing the dynamic real-time deviation correction signal and analyzing and calculating the dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters. This step is executed by a dynamic modulation engine integrated in the control system. Its core function is to convert the temperature change rate information from step S4 into a fine perturbation command for the underlying power output waveform. When the dynamic modulation engine receives a non-zero dynamic real-time deviation correction signal, it first initiates prediction and deduction calculations. The engine internally contains a trend prediction model based on the thermal inertia of the sintering process. Based on the amplitude and sign of the dynamic real-time deviation correction signal, this model pre-judges whether the current trend of temperature decrease or abnormal increase will continue to expand. Based on this judgment result, the engine deduces a fine-tuning phase offset. For example, a positive signal (overheating trend) will calculate a negative phase offset to delay the triggering timing of the carrier waveform edge as a whole, using thermal inertia to suppress the phase overshoot of temperature, rather than directly changing the energy integral within the pulse period. Meanwhile, another path of the dynamic modulation engine calculates the pulse density perturbation value, directly converting it into the total width of narrow pulses implanted or removed within the high level of the baseband waveform. This instantaneously reduces or increases the injected total energy by breaking down or filling the waveform. Simultaneously, another processing path of the dynamic modulation engine quantizes the absolute value of the dynamic real-time deviation correction signal. This path treats it as a direct reflection of system disturbances caused by sudden heat loss or gain, and, based on a preset scaling factor, directly calculates and converts the quantized value of the deviation signal into an increment or decrement of a time width, i.e., the pulse density perturbation value. This value is typically in microseconds and represents the total width of the narrow pulses that need to be implanted or removed from the baseband waveform. Finally, the dynamic modulation engine aggregates the calculated fine-tuning phase offset and pulse density perturbation value, and encapsulates these two parameters into a structured data packet, i.e., the modified modulation parameter sub-format, according to a predefined communication protocol. This sub-format is immediately sent to the downstream signal fusion module as the basis for real-time correction of the feedforward baseband waveform.

[0085] Among them, fine-tuning phase offset The calculation can be performed by a proportional-integral controller:

[0086] in, and These are the proportional and integral coefficients, which are set through systematic identification and simulation optimization of the sintering furnace thermal response model, aiming to ensure the speed and overshoot-free nature of phase modulation. Pulse density perturbation value. The calculation is a direct proportional conversion:

[0087] in, It is the pulse density conversion coefficient, and its dimensions are... The coefficient is set based on the relationship between the Joule energy that a single microsecond-level pulse can inject and the instantaneous temperature rise it can cause, as determined by experiments, ensuring that the energy level of the disturbance value matches the magnitude of the thermal imbalance reflected by the deviation signal. The modified modulation parameter sub-format is a data structure containing two fields, in the form of... This is used to efficiently transmit correction instructions downstream. The communication interaction protocol refers to the data exchange specifications between the dynamic modulation engine and the downstream signal fusion module, such as using a shared memory-based or high-speed serial bus approach to ensure low-latency transmission of modulation parameters.

[0088] Specifically, to ensure full transparency and feasibility for engineering applications, the three coefficients mentioned above are not abstract mathematical variables, but rather engineering adjustment parameters with an intuitive physical meaning of "feedback suppression." Since this is a dynamic correction mechanism to address abnormal temperature fluctuations, when an abnormal temperature rise trend is detected (i.e., the deviation signal is positive), the system must reduce the instantaneous energy injection. Therefore, these three feedback coefficients are logically set to negative values ​​to achieve precise reverse suppression. Among them, the proportional coefficient... This determines the system's "instantaneous push-pull strength" in response to temperature gradient changes, i.e., how many microseconds the power pulse waveform needs to lead or lag for each abnormal temperature change rate of 1℃ / s. Considering the thermal inertia of molybdenum strip sintering, its typical engineering guideline value is usually between -0.10 and -0.01 (e.g., preferably -0.05); integral coefficient This is used for "accounting correction," addressing trend drift that has persisted for a long time by gradually accumulating small offsets to force it back to zero. Its value is typically on a smaller order of magnitude, such as between -0.01 and -0.001. The pulse density conversion coefficient... It acts as a direct "energy-time translator," directly answering the physical question of "to offset a temperature fluctuation of 1℃ / s, a few microseconds of power supply pulses need to be instantaneously cut off from the base waveform." In engineering practice, technicians only need to calculate the excess joules required to generate the corresponding temperature fluctuation in a tiny segment, divide it by the joules released by the heating element energized for 1 microsecond, and take the opposite sign to obtain the constant. In a typical high-frequency PWM control loop, the typical value range of this coefficient is between -0.05 and -0.01 (preferably -0.02). By clearly defining the physical logic and reference value range of its reverse suppression, this system transforms the complex control algorithm into engineering parameters that can be intuitively calibrated and tuned by field personnel.

[0089] For example, in connection with step S4, the dynamic modulation engine receives a dynamic real-time deviation correction signal. ℃ / s. First, this signal is input to the phase offset calculation path. Assuming this path uses pure proportional control for simplification, the proportional coefficient is set. The calculated fine-tuning phase offset is then... This negative value indicates that the rising edge of the upcoming PWM waveform needs to be delayed by 1.5 microseconds to slightly reduce energy injection. Simultaneously, this signal is also input to the pulse density calculation path, setting the pulse density conversion coefficient. The calculated pulse density perturbation value is This negative value indicates that a narrow pulse with a total width of 0.6 microseconds needs to be removed during the high-level segment of the current PWM cycle through logical operations, thereby instantly reducing the injected power. Finally, the dynamic modulation engine aggregates these two calculation results to form the corrected modulation parameter sub-format. The data packet was immediately sent to the signal fusion module via the internal bus to fine-tune the power waveform that was about to be applied to heating module No. 2 in real time, in order to suppress the observed excessively rapid temperature rise of +30℃ / s.

[0090] S6. The integrated feedforward fundamental frequency waveform locking and feedback phase modulation mechanism generates a composite control command stream, executes local compensation actions, and outputs high-density finished molybdenum bars.

[0091] In a specific embodiment of the present invention, a composite control command stream is generated by integrating feedforward baseband waveform locking and feedback phase modulation mechanism, and a local compensation action is performed to output a high-density finished molybdenum bar. This includes: reading the baseband duty cycle of a dedicated micro-segment in the feedforward power command queue corresponding to the current action execution time as the basic energy injection command carrier waveform.

[0092] The fine-tuned phase offset covered in the modified modulation parameter sub-format is shifted and superimposed onto the edge of the carrier waveform, and the narrow pulse breakdown point corresponding to the pulse density perturbation value is implanted into the high-level segment of the carrier waveform to form a composite broadband pulse modulation signal.

[0093] The composite broadband pulse modulation signal is guided to the control relay terminal in the target furnace to drive the silicon carbide rod actuator to open and respond to output a predetermined compensation heat flow covering a small segment, thereby dynamically eliminating the internal heat absorption difference and external temperature fluctuation to output a dense and uniform high-density finished molybdenum bar.

[0094] Specifically, the sixth implementation step S6 of the method of the present invention involves generating a composite control command stream by integrating the feedforward baseband waveform locking and feedback phase modulation mechanisms, and executing local compensation actions to output high-density finished molybdenum bars. This step is completed by the signal fusion and actuator within the underlying power drive unit of each independent heating module, and is the final execution link of the entire control closed loop. In each control clock cycle, the unit first reads and schedules the baseband duty cycle specific to the current moment and the current small segment from the feedforward power command queue constructed in step S3, based on the current execution time. This baseband duty cycle, for example, a value of 65.06%, is used as the command carrier for generating a basic PWM waveform, defining the basic energy injection level within a standard cycle, and constituting the basic energy injection command carrier waveform. Next, the signal fusion unit checks whether there is a modified modulation parameter sub-format from step S5. If it exists, it unpacks the sub-format and extracts the fine-tuning phase offset and pulse density perturbation value. The signal fusion unit directly shifts and superimposes the fine-tuning phase offset onto the rising or falling edge of the carrier waveform. For example, a negative offset will cause the rising edge of the waveform to lag by a microsecond. Simultaneously, it converts the total time width corresponding to the pulse density perturbation value into one or more extremely narrow pulses, which are then inserted in a point-like manner within the high-level segment of the carrier waveform using logic NAND gates to achieve breakdown, thereby fine-tuning the average power without changing the fundamental frequency duty cycle. Through the aforementioned phase shift and pulse insertion operations, a composite wideband pulse modulation signal integrating the feedforward setpoint and feedback fine-tuning is generated. Finally, the unit guides this high-precision composite wideband pulse modulation signal, transmitting it directly to the control relay terminal, such as a high-speed solid-state relay, associated with the independent heating module within the target furnace. This relay, based on the precise timing of the composite wideband pulse modulation signal, drives the connected heating actuators, such as silicon carbide rods, to activate and respond by outputting a precisely corrected heat flow. This heat flow precisely covers the corresponding micro-segment. Because its energy level includes both a forward-looking compensation for the inherent heat absorption characteristics of the micro-segment and a dynamic correction for real-time fluctuations in the external environment, it can dynamically eliminate internal heat absorption gaps and external temperature fluctuations. This ensures that each micro-segment of the original molybdenum bar can undergo a near-ideal sintering temperature curve, ultimately outputting a high-density finished molybdenum bar with uniform structure and few defects.

[0095] Among them, composite wideband pulse modulation signal The generation logic can be formally described as follows:

[0096] Here This indicates the generation of a pulse width modulation signal with a specific duty cycle. The function converts the time-width perturbation value into a logic gate operation within the high-level segment of the PWM. The function then performs a time-axis translation of the entire waveform. Operators This indicates the signal fusion process. High-density finished molybdenum bars refer to products whose relative density exceeds 99.5% of the theoretical density, and which, when observed under a metallographic microscope, exhibit uniform grain distribution without obvious pores, cracks, or inclusions. All mechanical properties meet or exceed the national standards for high-quality molybdenum materials.

[0097] For example, continuing from the previous example, the power drive unit of heating module 2 reads the base frequency duty cycle as 65.06% from the queue. The waveform generator generation cycle is... The carrier waveform, its original rising edge is at The falling edge at At this point, the correction parameters are received. The signal fusion unit first... The phase offset acts on the entire waveform, causing the rising edge to be delayed until... The falling edge is translated to (The overall translation does not change the original total energy area, only the phase timing of energy injection). Next, process... The pulse density perturbation value is passed through a logic gate during the delayed high-level segment (such as the first...). (place) forcibly inserted a A wide low-level signal "breaks it down". The actual total effective time of the high level in the final composite broadband pulse modulation signal is: Furthermore, the phase initiation of energy injection was precisely delayed. This high-precision waveform, after time-domain shifting and internal energy shaping, perfectly cancels out... The trend of excessively rapid warming.

[0098] Reference Figure 4 The second aspect of the present invention provides an adaptive compensation system for the sintering temperature field of high-density molybdenum strips, comprising: a physical property data acquisition module, a thermal absorption characteristic analysis module, a feedforward power planning module, a real-time deviation correction signal extraction module, a dynamic modulation parameter calculation module, and a composite instruction fusion and execution module.

[0099] The physical characteristic data acquisition module is connected to the thermal absorption characteristic analysis module, the thermal absorption characteristic analysis module is connected to the feedforward power planning module, the feedforward power planning module is connected to the real-time deviation correction signal extraction module, the real-time deviation correction signal extraction module is connected to the dynamic modulation parameter calculation module, and the dynamic modulation parameter calculation module is connected to the composite instruction fusion and execution module.

[0100] The physical property data acquisition module acquires the physical property data of the original molybdenum bar and generates a multi-dimensional feedforward compensation reference vector containing continuous spatial coordinates.

[0101] The thermal absorption characteristic analysis module analyzes the multi-dimensional feedforward compensation benchmark vector and constructs a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip.

[0102] The feedforward power planning module performs forward prediction matching operations based on the thermal absorption characteristic prediction package, and plans and generates the underlying feedforward power instruction queue based on the instantaneous power balance model.

[0103] The real-time deviation correction signal extraction module synchronously collects spatiotemporal temperature and flow field information within the sintering area and extracts the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism.

[0104] The dynamic modulation parameter calculation module processes dynamic real-time deviation correction signals and analyzes and calculates dynamic phase and pulse density modulation parameters to cope with high-frequency disturbances.

[0105] The composite instruction fusion and execution module integrates feedforward base frequency waveform locking and feedback phase modulation mechanisms to generate a composite control instruction stream, executes local compensation actions, and outputs high-density finished molybdenum bars.

[0106] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. An adaptive compensation method for the sintering temperature field of high-density molybdenum strips, characterized in that, include: S1. Obtain the physical property data of the original molybdenum bar and generate a multi-dimensional feedforward compensation reference vector containing continuous spatial coordinates. The process of acquiring the physical property data of the original molybdenum bar and generating a multi-dimensional feedforward compensation reference vector containing continuous spatial coordinates includes: activating a non-contact physical property scanning unit to synchronously sense the continuous optical and geometric physical characteristics of the original molybdenum bar as it moves into the sintering furnace with the feeding mechanism. The laser diameter measuring instrument was used to collect continuous diameter data streams along the length of the original molybdenum strip, and the corresponding surface reflection and emission spectra were obtained using a multispectral analyzer. By fusing continuous diameter data streams with surface reflection and emission spectral streams, a multidimensional feedforward compensation reference vector corresponding to the feed position is calculated. S2. Analyze the multidimensional feedforward compensation benchmark vector and construct a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip. S3. Based on the thermal absorption characteristics prediction package, perform forward prediction matching operation, and plan and generate the underlying feedforward power instruction queue based on the instantaneous power balance model. S4. Synchronously collect spatiotemporal temperature and flow field information within the sintering area and extract the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism. The synchronous acquisition of spatiotemporal temperature flow field information within the sintering area and the extraction of dynamic real-time deviation correction signals that trigger the feedback adjustment mechanism include: driving a non-contact infrared temperature sensor array deployed along the transmission path to acquire the actual sintering temperature flow of the corresponding tiny segment within the independent heating module. The current value of the received real sintering temperature flow is compared with the preset optimized sintering temperature target two-dimensional field at high frequency in real time to obtain the current temperature deviation value. The system assesses and determines the gradient of changes in the current minute temperature deviation and outputs a dynamic real-time deviation correction signal that indicates local temperature fluctuations. S5. Process dynamic real-time deviation correction signals and analyze and calculate dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters; The process of processing the dynamic real-time deviation correction signal and analyzing and calculating the dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters includes: reading the dynamic real-time deviation correction signal into the dynamic modulation engine, and substituting the dynamic real-time deviation correction signal into the extended trend prediction model configured based on the proportional-integral mechanism for calculation. Based on the energy adjustment command calculated by the model, a microsecond-level time offset value is determined for advancing or delaying the edge of the power pulse, and a fine-tuning phase offset is generated to smooth out thermal response hysteresis. The sudden increase in heat loss is quantified by using dynamic real-time deviation correction signals and directly calculated into pulse density perturbation values ​​with microsecond time widths to directly adjust the total energy within the cycle. The aggregated fine-tuning phase offset and pulse density perturbation values ​​are integrated according to the communication interaction protocol to establish a precise correction modulation parameter sub-format; S6. The integrated feedforward fundamental frequency waveform locking and feedback phase modulation mechanism generates a composite control command stream, executes local compensation actions, and outputs high-density finished molybdenum bars. The integrated feedforward fundamental frequency waveform locking and feedback phase modulation mechanism generates a composite control command stream, executes local compensation actions to output high-density finished molybdenum bars, including: reading the fundamental frequency duty cycle of a dedicated micro-segment in the feedforward power command queue corresponding to the current action execution time as the basic energy injection command carrier waveform; The fine-tuned phase offset covered in the modified modulation parameter sub-format is shifted and superimposed onto the edge of the carrier waveform, and the narrow pulse breakdown point corresponding to the pulse density perturbation value is implanted into the high-level segment of the carrier waveform to form a composite wideband pulse modulation signal. The composite broadband pulse modulation signal is guided to the control relay terminal in the target furnace to drive the silicon carbide rod actuator to open and respond to output a predetermined compensation heat flow covering a small segment, thereby dynamically eliminating the internal heat absorption difference and external temperature fluctuation to output a dense and uniform high-density finished molybdenum bar.

2. The adaptive compensation method for the sintering temperature field of high-density molybdenum strips according to claim 1, characterized in that, The analytical multidimensional feedforward compensation benchmark vector is used to construct a thermal absorption characteristic prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip, including: Align the spatial position coordinates of the multidimensional feedforward compensation reference vector and extract the spectral features of the corresponding small segments to invert the real-time pre-sintering density and real-time surface emissivity; The cross-sectional area of ​​a small segment is calculated based on a continuous diameter data stream, and the volumetric heat capacity is dynamically calculated in conjunction with the real-time pre-sintering density. The real-time surface emissivity is extracted and coupled with the pre-stored spectral physical properties of the furnace heating element to calculate the effective radiation absorption coefficient of the environment; A solidified thermal absorption characteristic prediction package is constructed by encrypting the location coordinates of tiny segments with unique spatiotemporal timestamps, along with volumetric heat capacity and radiation absorption coefficient.

3. The adaptive compensation method for the sintering temperature field of high-density molybdenum strips according to claim 2, characterized in that, The process of extracting real-time surface emissivity and coupling it with pre-stored spectral physical properties of in-furnace heating elements to calculate the effective environmental radiation absorption coefficient includes: Based on Kirchhoff's thermal radiation law and under thermal equilibrium settings, the extracted real-time surface emissivity is equivalently converted into the basic radiation absorptivity of a tiny segment. The emissivity curve of the radiation spectrum of a specific heating element is retrieved from the pre-stored data and integrated with the basic radiation absorptivity in the frequency domain to obtain the effective radiation absorptivity of the environment.

4. The adaptive compensation method for the sintering temperature field of high-density molybdenum strips according to claim 1, characterized in that, The prediction package based on thermal absorption characteristics performs forward prediction matching operations and generates a feedforward power instruction queue based on the instantaneous power balance model, including: The spatial coordinate data corresponding to the heat absorption characteristic prediction package is detected and mapped to the physical range association model of each independent heating module in the sintering furnace. The obtained volumetric heat capacity and radiation absorption coefficient are input into the instantaneous power equation of the heat transfer theory to calculate the dynamic heating power requirement for the micro-segment to reach the target sintering temperature. Combined with the steady-state heat dissipation constant of the module, the target total feedforward power is obtained. The operating system on the chip divides the total target feedforward power by the module's maximum rated power according to the standard architecture of pulse width modulation, and converts it into the base frequency duty cycle used by the control unit. The feed rate vector along the original molybdenum bar is arranged with the time baseline to form a fundamental frequency duty cycle sequence to construct a feedforward power command queue covering all heating zones.

5. The adaptive compensation method for the sintering temperature field of high-density molybdenum strips according to claim 1, characterized in that, The process of evaluating and determining the gradient of changes in the current minute temperature deviation value and immediately outputting a dynamic real-time deviation correction signal indicating local temperature fluctuations includes: The rate of change of the current small temperature deviation value with time is calculated using the first-order time derivative algorithm to obtain the temperature deviation gradient. The temperature deviation gradient is compared with the set dead zone threshold. If the absolute value exceeds the dead zone threshold, a dynamic real-time deviation correction signal containing amplitude and sign information is generated.

6. An adaptive compensation system for the sintering temperature field of high-density molybdenum strips, characterized in that, include: The physical property data acquisition module acquires the physical property data of the fed raw molybdenum bar and generates a multi-dimensional feedforward compensation reference vector containing continuous spatial coordinates. The process of acquiring the physical property data of the original molybdenum bar and generating a multi-dimensional feedforward compensation reference vector containing continuous spatial coordinates includes: activating a non-contact physical property scanning unit to synchronously sense the continuous optical and geometric physical characteristics of the original molybdenum bar as it moves into the sintering furnace with the feeding mechanism. The laser diameter measuring instrument was used to collect continuous diameter data streams along the length of the original molybdenum strip, and the corresponding surface reflection and emission spectra were obtained using a multispectral analyzer. By fusing continuous diameter data streams with surface reflection and emission spectral streams, a multidimensional feedforward compensation reference vector corresponding to the feed position is calculated. The heat absorption characteristics analysis module analyzes the multi-dimensional feedforward compensation benchmark vector and constructs a heat absorption characteristics prediction package that characterizes the real-time heat-absorbing properties of different micro-segments of the original molybdenum strip. The feedforward power planning module performs forward prediction matching operations based on the thermal absorption characteristic prediction package, and plans and generates the underlying feedforward power instruction queue based on the instantaneous power balance model. The real-time deviation correction signal extraction module synchronously collects spatiotemporal temperature and flow field information within the sintering area and extracts the dynamic real-time deviation correction signal that triggers the feedback adjustment mechanism. The synchronous acquisition of spatiotemporal temperature flow field information within the sintering area and the extraction of dynamic real-time deviation correction signals that trigger the feedback adjustment mechanism include: driving a non-contact infrared temperature sensor array deployed along the transmission path to acquire the actual sintering temperature flow of the corresponding tiny segment within the independent heating module. The current value of the received real sintering temperature flow is compared with the preset optimized sintering temperature target two-dimensional field at high frequency in real time to obtain the current temperature deviation value. The system assesses and determines the gradient of changes in the current minute temperature deviation and outputs a dynamic real-time deviation correction signal that indicates local temperature fluctuations. The dynamic modulation parameter calculation module processes dynamic real-time deviation correction signals and analyzes and calculates dynamic phase and pulse density modulation parameters to cope with high-frequency disturbances. The process of processing the dynamic real-time deviation correction signal and analyzing and calculating the dynamic phase and pulse density modulation parameters to cope with high-frequency disturbance parameters includes: reading the dynamic real-time deviation correction signal into the dynamic modulation engine, and substituting the dynamic real-time deviation correction signal into the extended trend prediction model configured based on the proportional-integral mechanism for calculation. Based on the energy adjustment command calculated by the model, a microsecond-level time offset value is determined for advancing or delaying the edge of the power pulse, and a fine-tuning phase offset is generated to smooth out thermal response hysteresis. The sudden increase in heat loss is quantified by using dynamic real-time deviation correction signals and directly calculated into pulse density perturbation values ​​with microsecond time widths to directly adjust the total energy within the cycle. The aggregated fine-tuning phase offset and pulse density perturbation values ​​are integrated according to the communication interaction protocol to establish a precise correction modulation parameter sub-format; The composite instruction fusion and execution module integrates feedforward base frequency waveform locking and feedback phase modulation mechanisms to generate a composite control instruction stream, executes local compensation actions, and outputs high-density finished molybdenum bars. The integrated feedforward fundamental frequency waveform locking and feedback phase modulation mechanism generates a composite control command stream, executes local compensation actions to output high-density finished molybdenum bars, including: reading the fundamental frequency duty cycle of a dedicated micro-segment in the feedforward power command queue corresponding to the current action execution time as the basic energy injection command carrier waveform; The fine-tuned phase offset covered in the modified modulation parameter sub-format is shifted and superimposed onto the edge of the carrier waveform, and the narrow pulse breakdown point corresponding to the pulse density perturbation value is implanted into the high-level segment of the carrier waveform to form a composite wideband pulse modulation signal. The composite broadband pulse modulation signal is guided to the control relay terminal in the target furnace to drive the silicon carbide rod actuator to open and respond to output a predetermined compensation heat flow covering a small segment, thereby dynamically eliminating the internal heat absorption difference and external temperature fluctuation to output a dense and uniform high-density finished molybdenum bar.

Images