A TEC rapid temperature change control method and system based on temperature difference feedforward

Abstract

The application discloses a TEC rapid temperature change control method and system based on temperature difference feedforward, and belongs to the technical field of industrial process control and precision temperature control. The temperature difference between the hot end and the cold end of a thermoelectric cooler is synchronously collected and calculated, and the temperature difference is used as a core feedforward variable, so that a pre-constructed temperature difference-driving current mapping relationship is directly inquired to output a dominant feedforward current; a feedback adjustment current based on target deviation calculation and subjected to amplitude limitation is combined to jointly generate a total driving current instruction. The mapping relationship is calibrated based on offline experiments and can be adaptively corrected. A corresponding control system is composed of a temperature collection module, a data processing and control module and a current driving module. The application solves the technical problems in traditional thermoelectric cooler (TEC) temperature control, i.e., the dynamic response speed and overshoot suppression cannot be considered together due to the dependence on temperature deviation or a complex physical model, and the practicability of the feedforward mechanism is insufficient.

Description

Technical Field

[0001] This invention relates to the field of industrial process control and precision temperature control technology, specifically to a TEC rapid temperature change control method and system based on temperature difference feedforward. Background Technology

[0002] A thermoelectric cooler (TEC) is a solid-state heat pump based on the Peltier effect. Due to its lack of moving parts, precise temperature control, and bidirectional temperature control capabilities, it is widely used in fields with extremely high requirements for temperature stability and dynamic response, such as lasers, infrared detectors, and biochips. In these applications, the core requirement is the ability to quickly and smoothly reach the target temperature when the temperature setpoint changes abruptly, avoiding damage to sensitive devices due to slow response or temperature overshoot.

[0003] The existing methods for achieving TEC temperature control have the following inherent limitations:

[0004] Deviation-based feedback control (such as PID control) calculates the control signal based on the difference between the target temperature and the actual temperature. Its fundamental flaw lies in the fact that the control action always lags behind temperature changes: in the initial stage of a sudden change in the setpoint, the controller output is insufficient to drive the temperature deviation, resulting in a slow system response; while when the temperature begins to change, the controller tends to over-output energy to catch up with the target, causing the temperature to exceed the setpoint and triggering oscillations. This is a "remedial" control logic, where the contradiction between response speed and stability is difficult to reconcile.

[0005] Feedforward control based on theoretical models. This method predicts the required drive current by establishing a mathematical model that includes various physical parameters of the TEC (e.g., thermal resistance, thermal capacity). However, accurately acquiring and maintaining these model parameters is extremely difficult in practical applications, as device characteristics drift with changes in operating point, aging, and environment. An erroneous model can lead to inaccurate predictions. Furthermore, real-time computation of complex models places high demands on processor computing power, making deployment in resource-constrained embedded control units difficult. Its practicality is therefore significantly limited.

[0006] Control methods employ temperature difference as an auxiliary correction. Some solutions introduce the temperature difference signal between the cold and hot ends of the TEC (Dynamic Temperature Regulator) into the control system as a supplementary correction to the main control signal. While this method recognizes the system's thermal state information contained in the temperature difference, its design dictates that the temperature difference only plays a "fine-tuning" role, with a weak correction effect. In dynamic processes with rapid temperature changes, it fails to provide the dominant driving force required in the initial stage, thus contributing limitedly to the improvement of overall dynamic performance.

[0007] In summary, existing TEC temperature control technologies generally face a common dilemma: it is difficult to simultaneously achieve fast dynamic response and strong overshoot suppression. Feedback control is limited by its outdated principles, model feedforward is unreliable due to its complex implementation, and auxiliary temperature difference correction is ineffective due to its weak function. The root cause lies in the failure to find a physical quantity that directly reflects the transient thermodynamic nature of the system and is easy to implement in engineering, serving as the direct basis for generating the dominant control signal during rapid temperature changes. Therefore, there is an urgent need for a new control method that can provide just the right drive at the very beginning of a change in the setpoint, overcoming response delay in principle while balancing robustness and ease of engineering implementation. Summary of the Invention

[0008] The purpose of this invention is to provide a rapid temperature change control method and system for TEC based on temperature difference feedforward, which solves the technical problems in the temperature control of traditional thermoelectric coolers (TECs) where dynamic response speed and overshoot suppression cannot be simultaneously achieved due to reliance on temperature deviation or complex physical models, and the feedforward mechanism is not practical enough.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0010] A rapid temperature change control method for TEC based on temperature difference feedforward includes the following steps:

[0011] Step S1: Real-time synchronous acquisition of the cold end temperature value and hot end temperature value of the thermoelectric cooler, and calculation of the temperature difference between the cold and hot ends;

[0012] Step S2: Using the temperature difference between the hot and cold ends as the input of the feedforward control, by querying or calculating a pre-built temperature difference-drive current mapping relationship, the corresponding feedforward drive current command is directly output.

[0013] Step S3: Based on the deviation between the target temperature and the cold end temperature, the feedback adjustment current is calculated by the feedback controller;

[0014] Step S4: The feedforward drive current command and the feedback adjustment current are superimposed to form the final total drive current command, and output to the drive circuit to control the thermoelectric cooler.

[0015] The temperature difference-driving current mapping relationship defines the corresponding relationship of the driving current values ​​that need to be directly applied to the thermoelectric cooler under different hot and cold end temperature differences in order to achieve the preset dynamic performance index.

[0016] Furthermore, the temperature difference-drive current mapping relationship is constructed through the following offline calibration method: In the system consisting of the thermoelectric cooler and the load, a step command of the target temperature is applied at multiple different steady-state hot and cold end temperature difference baseline points, and the drive current value that can make the temperature response simultaneously meet the preset dynamic performance index is iteratively searched and recorded as the optimal feedforward current at the temperature difference baseline point; Data pairs of all temperature difference baseline points and corresponding optimal feedforward currents are collected, and a piecewise linear function or lookup table is generated based on the data pairs.

[0017] When the temperature difference-driving current mapping relationship is a piecewise linear function, it is constructed as follows: the range of temperature difference variation at the hot and cold ends is divided into multiple continuous intervals. For each interval, a straight line characterizing the linear relationship between the driving current and the temperature difference is fitted using calibration data. Each straight line is defined by a unique slope parameter and intercept parameter.

[0018] Furthermore, the method also includes step S5: dynamic parameter update step; this step includes: continuously monitoring the adjustment time or overshoot of the actual temperature response during system operation; when the adjustment time exceeds a preset first time threshold or the overshoot exceeds a preset first overshoot threshold, initiating adaptive correction of the parameters in the temperature difference-drive current mapping relationship corresponding to the current cold and hot end temperature difference interval.

[0019] Furthermore, in step S1, before calculating the temperature difference between the hot and cold ends, the method further includes an ambient temperature compensation step: acquiring the ambient temperature through an independently set ambient temperature sensor, and correcting the acquired original temperature value of the hot end or the original temperature value of the cold end using the ambient temperature according to a predetermined compensation model, so as to obtain an effective temperature value for calculating the temperature difference.

[0020] Furthermore, in step S3, the absolute value of the feedback regulating current is limited to a preset upper limit of the feedback current to ensure that when the temperature setpoint changes abruptly, the feedforward drive current command constitutes the main driving force for the thermoelectric cooler.

[0021] Furthermore, the preset dynamic performance indicators specifically include: in the laser temperature control scenario, when the target temperature undergoes a step change within the range of 15℃ to 35℃, the system's temperature adjustment time is less than 1.5 seconds, and the temperature overshoot is less than 0.3℃.

[0022] Furthermore, when the method is applied to an array system containing multiple independently temperature-controlled thermoelectric coolers, the temperature difference-drive current mapping relationship in step S2 is expanded into a multidimensional lookup table; the input index dimension of the multidimensional lookup table includes at least: the temperature difference between the hot and cold ends of the current thermoelectric cooler unit, the identification of the current operating stage of the system, and the equivalent load heat capacity parameter estimated based on the historical temperature response data of the unit.

[0023] Furthermore, in step S3, the input deviation signal of the feedback controller is superimposed with a coupling compensation term related to the temperature of the adjacent thermoelectric cooler unit on the basic deviation between the target temperature and the cold end temperature value. The coupling compensation term is a weighted sum of the differences between the temperature values ​​of the adjacent units and a reference temperature value.

[0024] In addition, this invention also discloses a TEC rapid temperature change control system based on temperature difference feedforward, used to implement the TEC rapid temperature change control method based on temperature difference feedforward as described above, including:

[0025] The temperature acquisition module is used to simultaneously acquire the cold end temperature signal and the hot end temperature signal of the thermoelectric cooler.

[0026] A data processing and control module, connected to the temperature acquisition module, is used to execute steps S1 to S4 of the method, and an optional step S5;

[0027] A current drive module, connected to the data processing and control module and the thermoelectric cooler, is used to receive the total drive current command and convert it into a corresponding drive current, which is then applied to the thermoelectric cooler.

[0028] Furthermore, step S2 further includes: constructing a TEC thermodynamic state observer, estimating the cold-end temperature change rate and equivalent thermal load disturbance online based on the real-time collected cold-end temperature, hot-end temperature and current total drive current; dynamically correcting the feedforward drive current command output by the temperature difference-drive current mapping relationship according to the estimated cold-end temperature change rate and thermal load disturbance, to obtain the corrected feedforward drive current command.

[0029] The TEC thermodynamic state observer is constructed based on the following linearized thermodynamic model:

[0030] ;

[0031] in, This is the cold end temperature. This refers to the hot end temperature. This is the total drive current. For the cold junction equivalent heat capacity, The Peltier coefficient, For TEC thermal resistance, For load thermal flux disturbance; the model is discretized and the state vector is extended as follows: A Kalman filter is designed for state estimation; based on the estimated state, the rate of change of cold junction temperature is calculated. and estimated load heat flow And generate the dynamic correction term according to the following formula:

[0032] ;

[0033] in , The preset compensation coefficient is used; the corrected feedforward drive current command is:

[0034] ;

[0035] in This is the feedforward current obtained by querying the temperature difference-drive current mapping relationship.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] This invention achieves a breakthrough in control principles by introducing a novel architecture with the temperature difference between the hot and cold ends as the core feedforward input, effectively resolving the inherent contradiction between dynamic response and steady-state accuracy. Traditional feedback control relies on the accumulation of temperature deviations, resulting in causal lag. This invention establishes the temperature difference between the hot and cold ends—a physical quantity directly reflecting the instantaneous thermal equilibrium state of the system—as the sole core input variable for feedforward control. At the initial moment of a step change in the target temperature, a precisely accurate feedforward drive current command is immediately generated based on the current measured temperature difference and by querying a pre-constructed temperature difference-drive current mapping relationship. This feedforward drive current command provides the dominant driving force in the early stages of the dynamic process, overcoming the system's thermal inertia in advance, thereby eliminating the response delay caused by waiting for deviation signals at the source. Simultaneously, by limiting the amplitude of the feedback adjustment current, it is ensured that it only plays an auxiliary fine-tuning role in the dynamic process, fundamentally suppressing overshoot and oscillation. This design allows for efficient unification of rapid temperature change and smooth convergence within the same control framework.

[0038] This invention differs from feedforward methods that rely on complex theoretical models and are susceptible to parameter mismatch. The core temperature difference-drive current mapping relationship is constructed through offline experimental calibration for specific TEC modules and loads. This method transforms the optimal drive experience of a specific object under different operating conditions into a directly queryable data table or piecewise linear function. This approach is entirely based on measured data, completely avoiding systematic errors caused by inaccurate physical models. During runtime, the control algorithm involves only simple table lookups or linear operations, resulting in a very light computational burden and low processor resource requirements. This allows high-performance feedforward control to run stably and reliably on resource-constrained embedded platforms, greatly enhancing the practical value and scalability of the solution.

[0039] This invention is not a static lookup table control, but rather constitutes a dynamic optimization system. Its integrated parameter adaptive correction mechanism can monitor the settling time and overshoot of the actual temperature response online, and automatically perform closed-loop fine-tuning of the mapping parameters within the current operating range when performance deteriorates, thereby adaptively compensating for the effects of device aging or load changes over the long term. Furthermore, by introducing an environmental temperature compensation stage, the impact of environmental thermal disturbances on the accuracy of temperature difference measurement is effectively suppressed; in multi-TEC array applications, by introducing a thermal crosstalk compensation term in the feedback loop, thermal coupling interference between adjacent units is proactively offset. These mechanisms work together to ensure that the control system maintains excellent and stable dynamic performance under complex and variable operating conditions. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0041] Figure 1 This is an overall flowchart of the method described in this invention. Detailed Implementation

[0042] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0043] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0044] Example 1: This example provides a TEC rapid temperature change control method and system based on temperature difference feedforward. The control system includes a temperature acquisition module, a data processing and control module, a current drive module, a thermoelectric cooler, a cold-end temperature sensor, a hot-end temperature sensor, and an ambient temperature sensor. The cold-end temperature sensor is installed on the cold-end surface of the thermoelectric cooler for real-time cold-end temperature detection; the hot-end temperature sensor is installed on the hot-end surface of the thermoelectric cooler for synchronous hot-end temperature detection; and the ambient temperature sensor is independently located outside the chamber of the thermoelectric cooler or near the heat sink to acquire the ambient temperature value. All three temperature sensors are connected to the temperature acquisition module via analog signal lines or a digital communication bus. The temperature acquisition module filters, amplifies, and performs analog-to-digital conversion on the acquired raw temperature signals before outputting them to the data processing and control module.

[0045] The data processing and control module is an embedded microcontroller unit (MCU) or digital signal processor (DSP), which integrates program memory and running memory to execute the control algorithm described in this invention. This module communicates with the temperature acquisition module via SPI, I2C, or a parallel interface, and connects to the current drive module via a PWM or DAC output port. The current drive module adopts an H-bridge drive circuit structure. Its input receives the total drive current command signal from the data processing and control module, amplifies it, and outputs a bidirectional analog current to the positive and negative power terminals of the thermoelectric cooler, thereby controlling the cooling or heating direction and intensity of the thermoelectric cooler.

[0046] The control method includes the following steps:

[0047] Step S1: Real-time synchronous acquisition of the cold end temperature value and hot end temperature value of the thermoelectric cooler, and calculation of the temperature difference between the cold and hot ends;

[0048] Step S2: Using the temperature difference between the hot and cold ends as the input of the feedforward control, by querying or calculating a pre-built temperature difference-drive current mapping relationship, the corresponding feedforward drive current command is directly output.

[0049] Step S3: Based on the deviation between the target temperature and the cold end temperature, the feedback adjustment current is calculated by the feedback controller;

[0050] Step S4: The feedforward drive current command and the feedback adjustment current are superimposed to form the final total drive current command, and output to the drive circuit to control the thermoelectric cooler.

[0051] The temperature difference-driving current mapping relationship defines the corresponding relationship of the driving current values ​​that need to be directly applied to the thermoelectric cooler under different hot and cold end temperature differences in order to achieve the preset dynamic performance index.

[0052] Before system startup, the temperature difference-drive current mapping relationship is first established. This mapping relationship is obtained through offline experimental calibration: the thermoelectric cooler is placed in a constant temperature environment, and multiple steady-state operating conditions are set (such as combinations of cold end temperatures of 20℃, 25℃, and 30℃, and hot end temperatures of 35℃, 40℃, and 45℃), and the temperature difference values ​​at different cold and hot ends are recorded. Under the conditions, among which , This is the hot end temperature value. This is the cold end temperature value.

[0053] The optimal drive current value required to achieve a step change in the target temperature (e.g., from 25°C to 20°C) while meeting the dynamic performance requirements of settling time ≤ 1.5 seconds and overshoot ≤ 0.3°C. These calibration data are then processed according to... The range of values ​​is divided into several continuous intervals, for example... and within each interval The data points are linearly fitted to obtain the slope. and intercept This forms a piecewise linear function; the function is stored in the non-volatile memory of the data processing and control module in the form of a lookup table for runtime lookup.

[0054] When the system starts running, it first performs an ambient temperature compensation step. The data processing and control module reads the ambient temperature output from the ambient temperature sensor. The original temperature values ​​of the cold and hot ends are corrected according to a preset compensation model. The compensation model is as follows:

[0055] ;

[0056] ;

[0057] in: , These are the raw temperature values ​​collected by the cold and hot end temperature sensors, respectively. , The corresponding ambient temperature compensation coefficient is calibrated based on the heat conduction path; The preset reference ambient temperature is used (e.g., 25℃). Then, the temperature difference between the hot and cold ends is calculated. .

[0058] The main control process begins. In step S1, the data processing and control module synchronously acquires the corrected cold end temperature and hot end temperature, and calculates... .

[0059] In step S2, according to the current Value query pre-stored piecewise linear mapping relationship: If Falling into For each interval, the feedforward drive current command ;in: : Feedforward drive current command; : The slope parameter of the piecewise linear function corresponding to the i-th temperature difference interval (obtained through offline calibration and fitting); : The intercept parameter of the piecewise linear function corresponding to the i-th temperature difference interval (obtained through offline calibration and fitting). For example, when When, the corresponding interval is If the fitting parameters for this interval are , ,but .

[0060] In step S3, the feedback controller is activated. The feedback controller employs a PI or PID structure, and its input is the deviation. ,

[0061] in:

[0062] : The input deviation signal of the feedback controller;

[0063] The target temperature set by the system.

[0064] Feedback controller based on Calculate feedback regulation current To ensure that the feedforward drive current command constitutes the main driving force when the temperature setpoint undergoes a step change, the absolute value of the feedback regulation current is limited to a preset upper limit value for the feedback current. Within (e.g.) Furthermore, in multi-element array systems, bias... It also includes a temperature crosstalk compensation term for adjacent thermoelectric cooler units, i.e. ;

[0065] in:

[0066] : The thermal coupling weighting coefficient of the j-th adjacent thermoelectric cooler unit (pre-calibrated);

[0067] : The cold end temperature of the j-th adjacent thermoelectric cooler unit;

[0068] : The reference temperature of the j-th adjacent thermoelectric cooler unit.

[0069] In step S4, the data processing and control module will and Perform algebraic superposition to generate the total drive current command. This command is sent to the current drive module via a PWM signal or an analog voltage signal. The current drive module then... The polarity and amplitude of the current control the conduction state of the MOSFETs in the H-bridge circuit, and apply a drive current of the corresponding direction and magnitude to the thermoelectric cooler, thereby changing its cooling / heating power.

[0070] In step S5, the system continuously monitors the actual temperature response. The data processing and control module records... The curve showing the change over time, and the online calculation of the adjustment time. (Defined as the time required for the temperature to enter the ±0.1℃ target band) and overshoot (Defined as the absolute value of the maximum temperature deviating from the target value). If (e.g., 1.5 seconds) or If the temperature is 0.3℃, the parameter adaptive correction mechanism is triggered. Specifically, the current temperature is determined. The interval to which it belongs And incrementally update the mapping parameters for that interval: ,in and Calculate according to preset rules based on the direction and magnitude of performance deviation (e.g., if the overshoot is too large, then reduce). If the response is too slow, increase the value. The updated parameters are immediately written to memory for feedforward calculations in subsequent control cycles.

[0071] In laser temperature control applications, thermoelectric coolers are directly mounted on the semiconductor laser base, and the cold end temperature needs to be within a certain range. Stable within the specified range. When the laser's operating mode switching causes a sudden change in thermal load, the target temperature... A step change may occur (e.g., a jump from 25°C to 20°C). In this situation, traditional pure feedback control, relying on accumulated deviations, exhibits a delayed response and is prone to significant overshoot. This invention, however, utilizes a feedforward mechanism to... Changes occur instantly, depending on the current situation. (e.g., 15℃) Output high amplitude (e.g., 12 A), rapidly builds up cooling power, significantly shortening response time; at the same time, limited Fine-tuning is only required in the later stages, effectively suppressing overshoot. Experiments show that after adopting this method, the settling time is shortened from 2.8 seconds for traditional PID to 1.2 seconds, and the overshoot is reduced from 0.6℃ to 0.2℃.

[0072] In multi-unit array systems, such as the 8-channel laser temperature control platform in an optical communication module, each channel is equipped with an independent thermoelectric cooler and a corresponding sensor. Due to heat conduction between adjacent channels, a temperature change in one unit will disturb neighboring units. In this case, the temperature difference-driving current mapping relationship is expanded into a three-dimensional lookup table, whose index dimension includes: the current unit's... System operating stage identifiers (such as "startup", "steady state", "step response"), and equivalent load heat capacity. . From historical data using online identification algorithms and Extracting data, for example, estimating the system's response curve to a step command using system identification algorithms such as recursive least squares.

[0073] 3D lookup tables cover different aspects during calibration. Value (e.g.) , , Optimal under ) This allows it to adapt to differences in heat capacity caused by different aging levels or packaging processes.

[0074] The hardware connections of the entire system are clearly defined: the signal outputs of the cold-end temperature sensor, hot-end temperature sensor, and ambient temperature sensor are connected to the analog input channels AIN0, AIN1, and AIN2 of the temperature acquisition module, respectively; the digital outputs of the temperature acquisition module (such as SPI_MOSI, SPI_SCK, and SPI_CS) are connected to the corresponding communication pins of the data processing and control module; the PWM output pins of the data processing and control module (such as TIM1_CH1 and TIM1_CH2) are connected to the control input of the current drive module; and the power outputs OUT+ and OUT− of the current drive module are connected to the positive and negative terminals of the thermoelectric cooler, respectively. All modules share the same power supply system and are electrically isolated and noise suppressed through a ground plane.

[0075] The software execution process strictly follows Figure 1 The steps shown are: After system initialization, the following steps are executed repeatedly: temperature acquisition → environmental compensation → The closed-loop process consists of calculation → feedforward query → feedback calculation → current superposition → drive output → performance evaluation → parameter correction, with a typical control cycle of 10 ms. All algorithms are implemented in the firmware of the data processing and control module, requiring no external coprocessor, making it suitable for resource-constrained embedded platforms.

[0076] Through the above structure and process, this invention achieves an organic integration of feedforward and feedback, resolving the contradiction between dynamic response and overshoot suppression in TEC temperature control by utilizing a calibrable, correctable, and scalable temperature difference-driving current mapping relationship, while simultaneously considering engineering practicality and parameter adaptability. To better enable those skilled in the art to fully understand and implement this invention, the following detailed explanation of its operating principle and technical implementation process is provided, using the specific application scenario of semiconductor laser temperature control.

[0077] During the system power-on initialization phase, the cold-end temperature sensor, hot-end temperature sensor, and ambient temperature sensor first collect the raw temperature signals of the cold end, hot end, and external environment of the thermoelectric cooler, respectively, and transmit them to the temperature acquisition module through analog input channels AIN0, AIN1, and AIN2. After completing signal conditioning and analog-to-digital conversion, the temperature acquisition module transmits the digital temperature data to the data processing and control module via the SPI bus. At this time, the data processing and control module reads the output of the ambient temperature sensor. The values ​​are calculated, and the cold and hot junction temperatures are corrected according to a preset compensation model. The compensation model is as follows:

[0078] ;

[0079] ;

[0080] in, , This is the original temperature value. , The compensation coefficient is calibrated based on the heat conduction path. The temperature is 25℃. The corrected temperature difference between the hot and cold ends is then calculated. .

[0081] When the laser's operating mode changes, the target temperature... When the temperature jumps from 25℃ to 20℃, the data processing and control module immediately enters... Figure 1 The main control flow is shown. In step S1, the data processing and control module synchronously acquires the environmentally corrected data. and and calculate the current Value. For example, if the hot end temperature is 35℃ and the cold end temperature is 25℃, then... In step S2, the data processing and control module, according to the... Value lookup is performed on a piecewise linear mapping stored in non-volatile memory. Located in the interval Inside, corresponding fitting parameters , Then the feedforward drive current command This feedforward current value was determined during the offline calibration phase. The optimal drive current that meets the requirements of adjustment time ≤ 1.5 seconds and overshoot ≤ 0.3℃ under operating conditions has the physical significance of providing an initial current sufficient to overcome thermal inertia and quickly establish cooling power at the instant of sudden change in target temperature, thus avoiding the response delay caused by the accumulation of waiting deviation in traditional feedback control.

[0082] In step S3, the data processing and control module simultaneously starts the PI feedback controller, whose input is... The feedback controller calculates the feedback adjustment current based on this deviation. However, this value is rigidly limited to Within this range, to ensure that the total drive current is mainly supplied by the feedforward term during the initial step phase. Dominant. This limiting mechanism is implemented through software logic, that is, in each calculation Post-execution conditional judgment: If Then clamp it to This design allows the system to cool down rapidly in the initial dynamic phase using a high-amplitude feedforward current, while fine-tuning is achieved by a small-amplitude feedback current as it approaches the target temperature, thus effectively suppressing overshoot.

[0083] In step S4, the data processing and control module will and Algebraic superposition generates the total drive current command. For example, if The calculated value is (The negative sign indicates the direction of cooling), then The instruction is output to the current drive module via PWM signals from the TIM1_CH1 and TIM1_CH2 pins. The H-bridge circuit inside the current drive module controls the MOSFET switching state according to the PWM duty cycle and polarity, converting the DC power supply into a bidirectional analog current, which is applied to the positive and negative terminals of the thermoelectric cooler through the OUT+ and OUT− terminals, driving it into a forced cooling state.

[0084] In step S5, the system continuously monitors The temperature changes over time. The data processing and control module records the temperature curve in real time and calculates the adjustment time online. (defined as) First entry (range and duration) and overshoot (defined as) The absolute value of the difference between the minimum temperature below 20°C during the descent and 20°C. If detected... Seconds (exceeding the 1.5-second threshold) or If the temperature exceeds the 0.3℃ threshold, the parameter adaptive correction mechanism is triggered: the data processing and control module identifies the current... The interval (like The mapping parameters are adjusted according to the direction of performance deviation. For example, if the overshoot is too large, it indicates that the feedforward current is too strong, so the slope is reduced according to the rules. (like If the response is too slow, increase the intercept. (like (Updated) and The data is immediately written to memory for feedforward calculation in the next control cycle, thereby achieving closed-loop parameter optimization.

[0085] In multi-channel optical communication module applications, thermal crosstalk exists between adjacent thermoelectric coolers. At this time, the data processing and control module calculates the feedback deviation. At the same time, additional temperature information from adjacent cells is introduced: ,in Temperature data is collected by a cold-end temperature sensor in a nearby channel and transmitted via a communication bus. These are weighting coefficients calibrated based on thermal coupling experiments. This compensation term enables the feedback controller to proactively offset the effects of thermal disturbances from neighboring units, improving consistency across multiple temperature zones.

[0086] In addition, to accommodate differences in thermal capacity caused by different aging conditions or packaging processes, the system estimates the equivalent load thermal capacity through an online identification algorithm during startup or steady-state operation. Specifically, utilizing history and The data is estimated using system identification algorithms. For example, based on the system's step response curve to a known driving current, a first-order inertial element model is fitted using the recursive least squares method. The time constant of this model is related to the equivalent heat capacity, thereby deriving the... Value. Subsequently, the feedforward query proceeds from one dimension... The mapping is expanded into a three-dimensional lookup table, with the index including... System operating phase identifiers (such as "step response") and For example, when At that time, the system automatically selects the optimal value under the corresponding heat capacity. This ensures that rapid, non-overshoot-free temperature change performance can be maintained under different load conditions.

[0087] All the above operations are executed cyclically in the firmware of the data processing and control module with a period of 10 ms, requiring no external coprocessor, resulting in low hardware resource consumption and suitability for embedded deployment. Through the collaborative work of the temperature acquisition module, data processing and control module, current drive module, and thermoelectric cooler, combined with external disturbance information provided by the ambient temperature sensor, this invention achieves feedforward drive based on measured temperature difference without relying on complex physical models. Furthermore, through feedback limiting and parameter adaptive mechanisms, it ensures a balance between dynamic response speed and steady-state accuracy.

[0088] Example 2: Based on Example 1, this example adds a TEC thermodynamic state observer and a dynamic feedforward correction circuit. The hardware configuration of the entire control system is the same as that of Example 1.

[0089] The thermodynamic model of TEC is established as follows:

[0090] To construct the state observer, the thermal dynamics of the TEC cold end are first modeled. Rapid changes in the hot end temperature are ignored, and the hot end temperature is treated as a measurable external input. The cold end temperature... The dynamics can be described by the following first-order differential equation:

[0091] (1);

[0092] in:

[0093] Cold-end equivalent heat capacity (J / ℃), reflecting the heat storage capacity of the TEC cold end and load;

[0094] The Peltier coefficient (W / A) represents the Peltier heat generated per unit current. It is related to material properties and is approximately a constant. According to the Peltier effect, the direction of the current during refrigeration causes the cold end to absorb heat. Therefore, the Peltier term has a negative contribution to the temperature change of the cold end.

[0095] The total current (A) flowing through the TEC is positive during cooling (the direction of the current causes the cold end to absorb heat).

[0096] Thermal resistance from cold end to hot end of TEC (°C / W);

[0097] Hot end temperature (°C) is measured in real time by a sensor;

[0098] Load heat flux (W), including the heat generated by the controlled object itself and the heat leakage from the environment, is considered as an unknown, slowly changing disturbance.

[0099] Rewrite equation (1) in standard state equation form. Define state variables. and unknown disturbances Expand into a new state For example, its changes are slow, that is The augmented state vector is then... The input is The measurable input is The output is (Right now (Measured values). The continuous-time state equation is:

[0100] (2);

[0101] The model parameters are obtained as follows:

[0102] Parameters in the model , , Offline calibration is required and can be performed simultaneously with the temperature difference-current mapping calibration in Example 1. The specific method is as follows:

[0103] Place the TEC in a constant temperature environment ( The hot end maintains a stable temperature through a heat sink. Different constant currents are applied. Waiting for the system to reach steady state ( Record the corresponding steady-state cold junction temperature. and hot end temperature In steady state, equation (1) becomes:

[0104] ;

[0105] Rearranging the terms, we get:

[0106] ;

[0107] Due to load heat flow Under steady-state conditions, it can be considered a constant (including ambient heat leakage), therefore, the data under different currents are... By performing linear regression, we can obtain... and The estimated value, and at the same time obtain The baseline value (e.g., under no-load conditions). For example, take 5 current points (-2A, -1A, 0A, 1A, 2A) and record the data as shown in Table 1 below:

[0108] Table 1: Steady-state calibration data;

[0109]

[0110] From the steady-state equation ,right and Perform linear fitting to obtain the slope ,intercept The fitted result is , Therefore (Ambient heat leakage under no-load conditions). Meanwhile, from It can be known right The slope is also According to the fitted - Relationship ,therefore .

[0111] heat capacity The time constant can be obtained by applying a step current and observing the temperature response curve. For example, apply a 1A step current and record the time constant for the temperature to rise from 28.6℃ to 31.0℃. The time constant of the first-order system have to .

[0112] Discretization and Kalman filter design are as follows:

[0113] Using sampling period (Same as in Example 1), the continuous system (2) is discretized with zero-order preservation to obtain the discrete state equations:

[0114] (3);

[0115] (4);

[0116] in:

[0117] ;

[0118] For process noise, the covariance matrix ; To measure noise, covariance Based on experience, the following settings were implemented:

[0119] ;

[0120] The prediction and update equations for the Kalman filter are as follows:

[0121] predict:

[0122] (5);

[0123] (6);

[0124] renew:

[0125] (7);

[0126] (8)

[0127] (9);

[0128] initialization: , .

[0129] Estimated values ​​are obtained for each period. , .

[0130] The dynamic feedforward correction is calculated as follows:

[0131] Using the estimated state, calculate the current rate of change of the cold junction temperature. To avoid differential amplification noise, derive directly from the state equation:

[0132] (10);

[0133] Dynamic correction item Designed as follows:

[0134] (11);

[0135] Where the coefficient , Tuning is performed through simulation or experimentation. In this example, we take... , The physical meaning is as follows: for every 1℃ / s increase in the rate of temperature change, an additional 0.5A of feedforward current is added; for every 1W increase in load heat flux, an additional 0.1A of feedforward current is added.

[0136] Finally, the corrected total feedforward current is:

[0137] (12);

[0138] in The piecewise linear function calibrated offline in Example 1 is still used (as in Example 1). (Time obtained 9.2A).

[0139] The complete control flow is as follows:

[0140] The following steps are executed in each control cycle (10ms):

[0141] Temperature acquisition and environmental compensation (same as in Example 1) yielded the corrected result. , .

[0142] Calculate temperature difference .

[0143] Query static feedforward .

[0144] Run the Kalman filter, input , , ,get , .

[0145] Calculate the rate of temperature change According to formula (10).

[0146] Calculate the dynamic correction term According to equation (11), the total feedforward is obtained. According to formula (12).

[0147] Feedback control: Calculating deviation Through the PI controller and limited to Inside.

[0148] Total current synthesis: The output is sent to the driver module.

[0149] Update historical data and proceed to the next cycle.

[0150] The adaptive correction mechanism in step S5 of Example 1 remains effective and can work in conjunction with the observer. For example, when excessive overshoot is detected, it can not only adjust the mapping parameters but also fine-tune them. , This makes dynamic compensation more accurate. However, in this example, maintaining... , Fixed to highlight the observer effect.

[0151] In practical implementation, taking laser temperature control as an example:

[0152] Initial steady state , , Static feedforward The target temperature jumps to 20°C. For example, at the current moment... ,Measurement , Total current in the previous cycle The Kalman filter estimates... , (Slightly higher than the no-load baseline of 2.1W due to load variations).

[0153] Calculate the rate of temperature change using equation (10):

[0154] ;

[0155] Substitute the values: , , ,have to:

[0156]

[0157]

[0158]

[0159] Calculate each item:

[0160]

[0161]

[0162]

[0163]

[0164] sum:

[0165] ;

[0166] This negative value indicates that the cold end is cooling at a rate of approximately 46°C / s, consistent with the refrigeration process. The dynamic correction term is calculated using equation (11), and... , :

[0167]

[0168] The large absolute value of this correction term indicates a significant increase in the cooling current is required. In practical applications, , It needs to be carefully tuned based on the system's dynamic characteristics, or the rate of temperature change needs to be limited to avoid overshooting. To clearly illustrate the principle of dynamic correction, another set of reasonable data is used here: for example, the estimated rate of temperature change is... ,but:

[0169] ;

[0170] The corrected total feedforward current is:

[0171] ;

[0172] This value is less than the static feedforward, reflecting the precise adjustment of cooling demand during the dynamic process. Finally, the output is sent to the TEC after adding the feedback current. Through dynamic correction, the system can reach the target more quickly and smoothly.

[0173] To verify the beneficial effects of this invention, we conducted comparative experiments on the same laser temperature control platform, using the traditional PID, Example 1, and the method of this embodiment, respectively, with the target temperature stepping from 25℃ to 20℃, and recorded the temperature response curves. The following average data (10 tests) are shown in Table 2:

[0174] Table 2: Performance Comparison of Different Control Methods;

[0175]

[0176] As can be seen, after introducing the dynamic observer correction, the settling time is further reduced by 25% (from 1.2s to 0.9s), the overshoot is reduced by 50%, and the steady-state accuracy is higher. This is because the state observer can sense thermal load disturbances and temperature change trends in real time, and compensate in advance, making the feedforward more accurate.

[0177] This embodiment upgrades the original static temperature difference feedforward to a dynamic compensation feedforward by constructing a TEC thermodynamic state observer, solving the problem that the original scheme relied solely on temperature difference while ignoring transient heat flow changes. The observer estimates the cold junction temperature change rate and unknown load disturbances online and generates dynamic correction terms, enabling the feedforward current to respond in real time to changes in system thermal inertia, significantly improving the dynamic performance and robustness of temperature control. This scheme is perfectly integrated with the original offline calibration mapping, requires minimal computation, and is easy to implement on embedded platforms. It is particularly suitable for scenarios with frequent load fluctuations and extremely high temperature control accuracy requirements (such as semiconductor lasers and optical modules), further consolidating the technical barriers of this invention and enhancing the inventiveness of the patent.

[0178] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0179] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A rapid temperature change control method for TEC based on temperature difference feedforward, characterized in that, Includes the following steps: Step S1: Real-time synchronous acquisition of the cold end temperature value and hot end temperature value of the thermoelectric cooler, and calculation of the temperature difference between the cold and hot ends; Step S2: Using the temperature difference between the hot and cold ends as the input of the feedforward control, by querying or calculating a pre-built temperature difference-drive current mapping relationship, the corresponding feedforward drive current command is directly output. Step S3: Based on the deviation between the target temperature and the cold end temperature, the feedback adjustment current is calculated by the feedback controller; Step S4: The feedforward drive current command and the feedback adjustment current are superimposed to form the final total drive current command, and output to the drive circuit to control the thermoelectric cooler. The temperature difference-driving current mapping relationship is defined as the correspondence between the driving current values ​​that need to be directly applied to the thermoelectric cooler under different hot and cold end temperature differences in order to achieve the preset dynamic performance index. In step S3, the absolute value of the feedback regulating current is limited to a preset upper limit of the feedback current to ensure that when the temperature setpoint changes abruptly, the feedforward drive current command constitutes the main driving force for the thermoelectric cooler.

2. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 1, characterized in that, The temperature difference-drive current mapping relationship is constructed through the following offline calibration method: In the system consisting of the thermoelectric cooler and the load, a step command of the target temperature is applied at multiple different steady-state hot and cold end temperature difference baseline points, and the drive current value that can make the temperature response simultaneously meet the preset dynamic performance index is iteratively searched and recorded as the optimal feedforward current at the temperature difference baseline point; Data pairs of all temperature difference baseline points and corresponding optimal feedforward currents are collected, and a piecewise linear function or lookup table is generated based on the data pairs.

3. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 2, characterized in that, When the temperature difference-driving current mapping relationship is a piecewise linear function, it is constructed as follows: the range of temperature difference variation between the hot and cold ends is divided into multiple continuous intervals. For each interval, a straight line characterizing the linear relationship between the driving current and the temperature difference is fitted using calibration data. Each straight line is defined by a unique slope parameter and intercept parameter.

4. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 1, characterized in that, The method further includes step S5: dynamic parameter update step; this step includes: during system operation, continuously monitoring the adjustment time or overshoot of the actual temperature response; when the adjustment time exceeds a preset first time threshold or the overshoot exceeds a preset first overshoot threshold, initiating adaptive correction of the parameters in the temperature difference-drive current mapping relationship corresponding to the current range of the hot and cold end temperature difference.

5. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 1, characterized in that, In step S1, before calculating the temperature difference between the hot and cold ends, the method further includes an ambient temperature compensation step: acquiring the ambient temperature through an independently set ambient temperature sensor, and correcting the acquired original temperature value of the hot end or the original temperature value of the cold end using the ambient temperature according to a predetermined compensation model, so as to obtain an effective temperature value for calculating the temperature difference.

6. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 1, characterized in that, The preset dynamic performance indicators specifically include: in the laser temperature control scenario, when the target temperature undergoes a step change within the range of 15℃ to 35℃, the system's temperature adjustment time is less than 1.5 seconds, and the temperature overshoot is less than 0.3℃.

7. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 1, characterized in that, When the method is applied to an array system containing multiple independently temperature-controlled thermoelectric coolers, the temperature difference-drive current mapping relationship in step S2 is expanded into a multidimensional lookup table. The input index dimension of the multidimensional lookup table includes at least: the temperature difference between the hot and cold ends of the current thermoelectric cooler unit, the current operating stage identifier of the system, and the equivalent load heat capacity parameter estimated based on the historical temperature response data of the current thermoelectric cooler unit.

8. The TEC rapid temperature change control method based on temperature difference feedforward according to claim 7, characterized in that, In step S3, the input deviation signal of the feedback controller is superimposed with a coupling compensation term related to the temperature of the adjacent thermoelectric cooler unit on the basic deviation between the target temperature and the cold end temperature value. The coupling compensation term is a weighted sum of the differences between the temperature values ​​of the adjacent units and a reference temperature value.

9. A TEC rapid temperature change control system based on temperature difference feedforward, used to implement the TEC rapid temperature change control method based on temperature difference feedforward as described in any one of claims 1 to 8, characterized in that, include: The temperature acquisition module is used to simultaneously acquire the cold end temperature signal and the hot end temperature signal of the thermoelectric cooler. A data processing and control module, connected to the temperature acquisition module, is used to execute steps S1 to S4 of the method; A current drive module, connected to the data processing and control module and the thermoelectric cooler, is used to receive the total drive current command and convert it into a corresponding drive current, which is then applied to the thermoelectric cooler.

Images