Control method and device of low-noise thermoelectric refrigeration type detector

Abstract

The invention relates to the technical field of detector control, and discloses a low-noise thermoelectric refrigeration type detector control method and device, and the method comprises the steps: collecting the real-time temperature parameter of a detector, and calculating the temperature prediction data; detecting the current exposure time length of the detector and forming a control strategy; thermal electron noise calculation is carried out based on the temperature prediction data, and a chip temperature upper limit constraint value is determined; establishing a constraint optimization equation based on the control strategy and the chip temperature upper limit constraint value, and solving a driving current value; the thermoelectric refrigerating unit driving circuit adjusts the refrigerating power of the double-layer thermoelectric refrigerating structure according to the driving current value, chip temperature control is completed, collaborative optimization of temperature control and noise suppression is achieved, and the technical problem that temperature precision and noise performance conflict with each other is solved.

Description

Technical Field

[0001] The present invention relates to the field of detector control technology, and in particular to a control method and device for a low-noise thermoelectric cooling type detector. Background Art

[0002] The electronic components in detectors generate heat during operation, causing chip temperatures to rise, which in turn leads to a significant increase in thermal electron noise. As the detector's operating time increases and the exposure time of a single image increases, the level of thermal electron noise rises sharply, seriously affecting image quality and detection accuracy, and becoming a key technical bottleneck restricting the development of high-performance detectors.

[0003] The current mainstream detector temperature control technology relies primarily on traditional PID control algorithms to drive thermoelectric coolers for temperature regulation. However, this control approach suffers from issues such as response lag, insufficient control accuracy, and an inability to effectively suppress thermal electron noise. Traditional PID controllers only provide feedback based on current temperature deviations and lack the ability to predict future temperature trends, making them incapable of handling the dynamic thermal loads experienced during detector operation. Summary of the Invention

[0004] The present invention provides a control method and device for a low-noise thermoelectric cooling detector, which realizes the coordinated optimization of temperature control and noise suppression, and solves the technical problem of the conflict between temperature accuracy and noise performance.

[0005] In a first aspect, the present invention provides a control method for a low-noise thermoelectric cooling detector, the control method for the low-noise thermoelectric cooling detector comprising: Collect the real-time temperature parameters of the detector and calculate the temperature prediction data; Detect the current exposure time length of the detector and form a corresponding control strategy; Performing thermal electron noise calculation based on the temperature prediction data to determine an upper limit constraint value of the chip temperature that meets imaging quality requirements; Calculating a driving current value of a current cycle based on the control strategy and the chip temperature upper limit constraint value; The thermoelectric cooler driving circuit adjusts the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control.

[0006] In combination with the first aspect, in a first implementation of the first aspect of the present invention, collecting the real-time temperature parameters of the detector and calculating the temperature prediction data includes: Platinum resistance temperature sensors arranged on the surface of each chip and the hot and cold ends of the thermoelectric cooler sense temperature changes and output resistance value change signals; Based on the resistance value change signal, the platinum resistance temperature sensor is driven by a constant current source circuit to generate a voltage signal proportional to the temperature, and the voltage signal is amplified by a signal conditioning circuit and converted into an analog temperature signal within a standard voltage range; The analog temperature signals are synchronously collected by a multi-channel analog-to-digital converter and converted into corresponding digital temperature values, and real-time temperature parameters containing temperature information of each measuring point are generated according to the digital temperature values; The chip temperature variation trend in multiple future control cycles is calculated based on the real-time temperature parameters and the thermoelectric coupling heat transfer equation to generate temperature prediction data.

[0007] In combination with the first aspect, in a second implementation of the first aspect of the present invention, calculating the chip temperature change trend within multiple future control cycles based on the real-time temperature parameter and the thermoelectric coupling heat transfer equation to generate temperature prediction data includes: Substituting the chip temperature value and the hot and cold end temperature values ​​in the real-time temperature parameters into the Seebeck effect electromotive force equation to calculate the reverse electromotive force value generated by the temperature difference; Substituting the chip temperature value and the current thermoelectric cooler current into the Peltier effect cooling power equation and the Joule heating effect heating power equation to calculate the cooling power value and the heating power value respectively; Combining the back electromotive force value, the cooling power value, and the heating power value to establish a dynamic heat transfer model with the chip temperature as a state variable; creating a state space model including a system matrix and a control matrix based on the dynamic heat transfer model; The chip temperature response value at each moment in a plurality of future control cycles is calculated by recursive operation using the state space model, and the chip temperature response values ​​are combined into temperature prediction data in chronological order.

[0008] In combination with the first aspect, in a third implementation of the first aspect of the present invention, recursively calculating the chip temperature response value at each moment in multiple future control cycles using the state space model, and organizing the chip temperature response values ​​into temperature prediction data in chronological order, includes: Setting the chip temperature value in the real-time temperature parameter as the initial state vector of the state space model, and setting the prediction time domain length to a preset number of control cycles; Based on the system matrix and the control matrix in the state space model, recursively calculate the state vector evolution value of each future control cycle starting from the current moment, wherein each recursive operation uses the state vector at the previous moment and the current control input; The temperature component in the state vector evolution value is extracted to obtain a chip temperature value sequence corresponding to each prediction moment, and the chip temperature value sequence is paired and combined with the corresponding time tag to generate temperature prediction data.

[0009] In combination with the first aspect, in a fourth implementation of the first aspect of the present invention, detecting the current exposure time length of the detector and forming a corresponding control strategy includes: The detector working status monitoring module reads the exposure time setting value of the current imaging task in real time to obtain the current exposure time length; Substituting the current exposure time length into the prediction time domain adjustment algorithm, and calculating the prediction time domain range by a linear combination of the basic prediction time domain length and the exposure time increment; Querying a preset weight adjustment lookup table according to the current exposure time length, obtaining an output weight coefficient corresponding to the long exposure mode and a control weight coefficient corresponding to the short exposure mode, and generating a temperature tracking weight parameter; Based on the combination of the prediction time domain range and the temperature tracking weight parameter, a control strategy including time domain configuration and weight configuration is formed.

[0010] In combination with the first aspect, in a fifth implementation of the first aspect of the present invention, performing thermal electron noise calculation based on the temperature prediction data to determine an upper limit constraint value of the chip temperature that meets imaging quality requirements includes: Substituting the chip temperature values ​​at each prediction moment in the temperature prediction data into a thermal electron noise calculation model including a temperature coefficient and a reference noise parameter, and calculating the thermal electron noise intensity value at the corresponding moment; Based on the thermal electron noise intensity value, a noise response curve reflecting the noise variation with temperature is constructed to obtain corresponding relationship data between temperature and noise level; The preset maximum allowable noise level threshold is queried according to the detector imaging quality standard, and a reverse lookup operation is performed in combination with the noise response curve to determine the chip temperature upper limit constraint value that meets the imaging quality requirements.

[0011] In combination with the first aspect, in a sixth implementation of the first aspect of the present invention, solving the driving current value of the current cycle based on the control strategy and the chip temperature upper limit constraint value includes: Constructing an objective function including a temperature deviation term and a control increment term based on the temperature tracking weight parameter and the prediction time domain range in the control strategy; Converting the chip temperature upper limit constraint value into an inequality constraint condition for the chip temperature at each moment in the prediction time domain, and adding the thermoelectric cooler current amplitude limit and current change rate limit to form a complete constraint condition group; The objective function is combined with the constraint condition group to form a constrained optimization equation, and the constrained optimization equation is numerically solved by an effective set algorithm to obtain the driving current value of the current cycle.

[0012] In combination with the first aspect, in a seventh implementation of the first aspect of the present invention, combining the objective function with the constraint condition group to form a constrained optimization equation, and numerically solving the constrained optimization equation using an active set algorithm to obtain the driving current value of the current cycle includes: Based on the temperature deviation term and the control increment term in the objective function and the inequality constraints and equality constraints in the constraint condition group, constructing a constrained optimization equation including decision variables, an objective coefficient matrix, and a constraint coefficient matrix; Selecting an initial feasible point based on the current state of the thermoelectric cooler, identifying active constraints of the initial feasible point to establish an initial valid constraint set, and calculating a corresponding Lagrange multiplier vector; By solving the linear equations, the search direction under the current valid constraint set is calculated to determine whether the search direction meets the optimality condition. If not, the decision variables are updated along the search direction and the valid constraint set is adjusted for the next round of iterative operations. When all Lagrange multiplier vectors are non-negative and the search direction is a zero vector during the iteration process, the iteration is terminated, and the corresponding thermoelectric cooler current component in the decision variable at this time is output as the driving current value of the current cycle.

[0013] In combination with the first aspect, in an eighth implementation of the first aspect of the present invention, the thermoelectric cooler driving circuit adjusts the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete chip temperature control, including: Performing linear conversion from digital to analog on the driving current value to generate a corresponding reference voltage control signal; Inputting the reference voltage control signal into the thermoelectric cooler drive circuit for current regulation to generate a stable thermoelectric cooler drive current; The stable thermoelectric cooler driving current is simultaneously delivered to the first layer of thermoelectric cooler in close contact with the chip surface and the second layer of thermoelectric cooler in close contact with the detector housing in the double-layer thermoelectric cooling structure according to a preset distribution ratio, so that the cold end of the first layer of thermoelectric cooler directly absorbs the heat generated by the chip and conducts it to the hot end, and the second layer of thermoelectric cooler receives the heat conducted by the first layer and dissipates it to the external environment; The chip temperature is precisely adjusted to the target temperature range through the double-layer thermoelectric cooling structure, and the working current and temperature status of each layer of thermoelectric cooler are monitored to ensure stable operation of the system and complete chip temperature control.

[0014] In a second aspect, the present invention provides a control device for a low-noise thermoelectric cooling type detector, the control device for the low-noise thermoelectric cooling type detector comprising: The acquisition module is used to collect the real-time temperature parameters of the detector and calculate the temperature prediction data; The adjustment module is used to detect the current exposure time length of the detector and form a corresponding control strategy; a calculation module, configured to calculate thermal electron noise based on the temperature prediction data and determine an upper limit constraint value of the chip temperature that meets imaging quality requirements; A solution module, configured to solve a driving current value of a current cycle based on the control strategy and the chip temperature upper limit constraint value; The output module is used for the thermoelectric cooler driving circuit to adjust the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control.

[0015] The technical solution provided by the present invention adopts a rolling time domain model predictive control algorithm, which can predict the chip temperature change trend within multiple future control cycles. Compared with traditional PID control, it has active preventive control capabilities and significantly improves the dynamic response performance of temperature control. By comprehensively considering the physical coupling relationship between the Seebeck effect, Peltier effect and Joule heating effect, an accurate dynamic heat transfer model is established, which more accurately describes the operating characteristics of the thermoelectric cooler than the simplified modeling of the existing technology. The thermal electron noise level is incorporated into the control optimization process as a hard constraint condition to achieve the coordinated optimization of temperature control and noise suppression, solving the technical problem of the conflict between temperature accuracy and noise performance. The system can dynamically adjust the prediction time domain and control weight parameters according to different exposure modes, and has stronger adaptability than fixed parameter control strategies. The use of a cascaded double-layer thermoelectric cooling structure and a precise current control mechanism ensures that the output of the control algorithm is accurately converted into the actual cooling effect, improving the control accuracy and reliability of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 Schematic diagram of the steps of a control method for a low-noise thermoelectric cooling detector according to an embodiment of the present invention; Figure 2 Schematic diagram of the structure of the control device of the low-noise thermoelectric cooling detector in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The embodiments of the present application provide a control method and device of a low-noise thermoelectric refrigeration type detector. The terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be exchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" or "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0019] For ease of understanding, the specific flow of the embodiments of the present application is described below. Please refer to Figure 1 One embodiment of the control method of the low-noise thermoelectric refrigeration type detector in the embodiments of the present application includes: Step S1, collecting real-time temperature parameters of the detector and calculating temperature prediction data; It can be understood that the execution subject of the present application can be a control device of a low-noise thermoelectric refrigeration type detector, and can also be a terminal or a server, and the specific place is not limited. The embodiments of the present application take the server as the execution subject for example.

[0020] Specifically, high-precision temperature sensing devices are placed at key heat transfer locations in the detector. Platinum resistance temperature sensors (such as PT1000) with high linearity and stability are selected and mounted on the detector chip surface, at the hot and cold interfaces of the thermoelectric cooler, and at external environmental monitoring points. This ensures that the sensors accurately sense minute temperature changes at the corresponding locations within each control cycle and output a resistance signal reflecting the thermal state change in real time. A constant current source circuit applies a stable current between 1mA and 5mA to the platinum resistance to prevent self-heating errors when heated and to ensure that the output voltage signal is linear and controllable. Because the resistance of the platinum resistance exhibits an approximately linear relationship with temperature, constant current excitation ensures that the output voltage signal varies proportionally to the sensed temperature. To enhance the system's ability to resolve minute temperature changes, a high-precision signal conditioning circuit is introduced to amplify, filter, and offset the raw voltage signal to ensure it meets the standard voltage input range required for analog-to-digital conversion. The amplified signal amplitude is typically between 0 and 3.3V, and it also features 50Hz or 60Hz power frequency noise suppression. By configuring a multi-channel synchronously sampling analog-to-digital converter (ADC) module, such as the 12-bit ADC channels integrated in the STM32H750 microprocessor, the analog temperature signals from multiple sensing points are sampled and digitized in parallel at a sampling frequency of 100Hz. This yields a set of high-precision digital temperature values ​​for the current chip temperature, the TEC cold-end and hot-end temperatures, and the ambient temperature at the same moment. This set of values ​​is then aggregated into a set of real-time temperature parameters. During the control algorithm processing phase, this set of temperature parameters serves as the observation input for state estimation. Combined with the previously established thermoelectric coupling dynamic heat transfer equation, specifically considering dynamic thermal processes such as heat absorption due to the Peltier effect, current dissipation due to Joule heating, and the heat conduction path through the TEC structure, a state-space model is constructed with the chip temperature as the controlled object and the TEC drive current as the control input. This model numerically simulates and recursively predicts chip temperature trends over multiple future control cycles. Within the prediction time domain, the expected chip temperature at each future moment is iteratively calculated based on the current state x(k) and the control variable u(k+i), generating temperature prediction data on the time axis.

[0021] Step S2: Detect the current exposure time length of the detector and form a corresponding control strategy; Specifically, based on the working status monitoring module, the imaging task currently being run by the detector is monitored in real time. The module reads the task setting parameters including exposure time through the communication interface with the image processing unit or the task scheduling system. The exposure time refers to the time that the photosensitive element is continuously exposed to light during a single image acquisition process. The length of the exposure time determines the chip's thermal load level and the demand for temperature control response speed. The control system obtains the current exposure time length and substitutes it as an input parameter into the prediction time domain adjustment algorithm. The algorithm adopts a dynamic time domain adjustment mechanism based on linear mapping, that is, the prediction interval is adjusted by a proportional factor between a basic prediction time domain length and the current exposure time increment. When the exposure time becomes longer, the prediction time domain is lengthened accordingly to cover the thermal response of the chip temperature for a longer period of time. When the exposure time is short, the prediction time domain is shortened to reduce the computational burden and improve system response speed. Simultaneously, based on the specific numerical range of the current exposure time, a pre-set weight adjustment lookup table is queried. This lookup table, an empirical mapping table derived from imaging experimental data and system control characteristics, outputs two weight parameters corresponding to different exposure times: an output tracking weight and a control action weight. For long exposure modes, the output tracking weight is set larger to enhance the ability to accurately track the target temperature trajectory, thereby reducing the level of thermal electron noise caused by chip heating. In short exposure modes, the control action weight is increased to reduce the adjustment impact caused by excessively frequent control, thereby avoiding system oscillation and excessive energy consumption. The predicted time domain length obtained from the exposure time and the two weight coefficients extracted from the lookup table are combined to form a control strategy that includes time window configuration and control optimization bias configuration.

[0022] Step S3: performing thermal electron noise calculation based on the temperature prediction data to determine an upper limit constraint value of the chip temperature that meets the imaging quality requirements; Specifically, each temperature value in the temperature prediction trajectory is sequentially substituted into a thermal electron noise calculation model. This model comprehensively considers parameters such as reference temperature, temperature coefficient, and reference noise intensity to quantitatively characterize the impact of temperature changes on thermal electron noise. The model describes how increasing temperature leads to enhanced thermal excitation, which in turn exacerbates the random fluctuations in charge carrier generation, resulting in a nonlinear growth relationship in noise intensity. Each temperature value substituted into the calculation model at each prediction moment produces a corresponding noise intensity output in the calculation model, thus forming a series of thermal electron noise intensity values. Based on this sequence, the system plots or fits a noise response curve that reflects the change in noise with chip temperature. This response curve numerically forms a one-to-one mapping between temperature and noise, showing a trend of accelerating noise increase with increasing temperature. Based on this temperature-noise correspondence, a key constraint parameter, the maximum allowable noise level threshold, is read from the detector imaging quality specification or industry standard. This threshold represents the maximum noise limit that the imaging system can tolerate in actual operation. Exceeding this threshold will inevitably have irreversible negative effects on the imaging signal-to-noise ratio, resolution, and contrast. Therefore, based on the noise response curve, a reverse search operation is performed, that is, the first temperature point corresponding to the noise intensity equal to or slightly lower than the maximum allowable noise threshold is found in the curve, and the value of this temperature point is used as the chip temperature upper limit constraint value that meets the imaging quality requirements.

[0023] Step S4: solving the driving current value of the current cycle based on the control strategy and the chip temperature upper limit constraint value; Specifically, during the control strategy phase, a data set consisting of temperature tracking weight coefficients, control response weight coefficients, and prediction time step size is generated through a combination of table lookup and calculation, based on the current exposure time and imaging mode set by the imaging task parameters. This data set reflects the system's tendency to adjust between temperature control accuracy and control action smoothness within the current control cycle. The temperature tracking weight determines the control system's attention to the deviation between the predicted temperature value and the desired temperature when optimizing the objective function, while the control response weight determines the strength of the system's constraint on the amplitude of controller output changes. The prediction time step size specifies the number of future control cycles covered by the prediction range. Together, these three factors define the structure of the objective function and the width of the prediction window. Based on this, a state prediction algorithm, based on the dynamic heat transfer model of a thermoelectric cooler, recursively estimates the response trajectory of the chip temperature to input current changes within each future control step, forming a set of prediction paths that reflect the impact of the current control variables on the evolution of future state variables. Based on this, an objective function is constructed. Its mathematical structure consists of two parts: the sum of squared deviations between the predicted chip temperature and the target temperature at all times within the prediction time domain, which characterizes the system's optimization objective for temperature tracking performance. The other part is the sum of squared changes in the control current sequence between adjacent control cycles, which is used to control the intensity of current fluctuations to balance hardware response speed and stability. The weight coefficients of this objective function are determined by the output weights and control weights in the control strategy, respectively. By minimizing this function, the controller can track the temperature target while avoiding system oscillation or overshoot caused by excessive current changes. After constructing the objective function, the aforementioned chip temperature upper limit constraint is introduced into the prediction path. This upper limit is expanded into an inequality constraint that holds at all times within the prediction time domain. This constraint requires that the chip temperature within any prediction step must not exceed the upper limit. This condition ensures that the thermal electron noise level is always controlled within the range allowed by imaging quality, thereby indirectly protecting the imaging signal-to-noise ratio. To prevent the controller output current from exceeding the current handling capacity of the thermoelectric cooler, a current amplitude restriction condition is added to the constraint set. This condition stipulates that the current value must remain between the preset maximum and minimum driving capabilities within any prediction step to avoid overload operation or cooling failure. At the same time, a current change rate restriction is imposed to ensure that the current change between two adjacent cycles does not exceed the physical response upper limit of the drive circuit, thereby avoiding unstable driver output or abnormal system heating due to excessive current jumps.The above objective function is integrated with three constraints, namely the temperature upper limit constraint, the current amplitude constraint, and the current change rate constraint, to form a constrained optimization problem. This optimization problem is a quadratic programming problem with inequality constraints. To solve this problem, an optimization engine built based on an iterative solver is called. This solver uses the active set method as the solution strategy. In each control cycle, the prediction starting point is refreshed in real time based on the current state measurement value. By iteratively selecting the control input sequence that minimizes the objective function value in the feasible solution domain that meets the current constraints, a set of optimal control current solutions is obtained, and only the first current value in the current control cycle is extracted as the target drive current output value actually executed.

[0024] Step S5: The thermoelectric cooler driving circuit adjusts the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control.

[0025] Specifically, a high-speed digital-to-analog converter integrated within the microprocessor performs a linear proportional conversion on the drive current value, converting it into a reference voltage control signal of corresponding amplitude. The amplitude of the voltage signal represents the current required cooling power adjustment instruction in the analog domain. The reference voltage control signal is input into the thermoelectric cooler drive circuit, which is equipped with a set of fast-response, high-precision voltage-to-current conversion modules. For example, a dedicated constant current source control chip is used, which can automatically adjust the output current based on the input voltage within milliseconds. This ensures that the output thermoelectric cooler drive current has extremely high current stability and dynamic tracking capabilities, as well as low noise and high linearity characteristics, ensuring that the cooler input response accurately reflects even small changes in the control instruction. The stable output driving current will be simultaneously delivered to the double-layer thermoelectric cooling structure according to the system's preset shunt control ratio. The first layer of thermoelectric cooler is installed close to the lower surface of the detector chip, and its cold end is in direct contact with the bottom surface of the chip. It can efficiently capture and absorb the heat generated by the operation of the electronic device during the operation of the chip, and conduct the absorbed heat energy to its hot end along the heat flow direction inside the device; the second layer of thermoelectric cooler covers the metal shell part of the detector with a larger size, and its cold end is connected to the hot end of the first layer, acting as an intermediate heat buffer and secondary heat extraction. Its hot end faces the external casing or heat dissipation system, so that the waste heat generated by the system can eventually be discharged into the environment through the structural boundary, thus forming a stable chip heat energy transmission path, which is organized in a double-layer series manner. At the same time, to ensure the safety and closed-loop regulation capability of the thermal control system during actual operation, the operating status of each level of the double-layer thermoelectric cooler is monitored in real time through multi-channel high-precision current sampling modules and temperature sensors. Current sampling is used to determine whether the drive output is stable and whether there is overcurrent or short circuit. Temperature sensors are placed at the contact surface of the hot and cold ends of the cooler to monitor the heat exchange efficiency and temperature difference response at any time. When the system detects that the actual temperature of a certain level of cooler deviates from the target control range or the current output is abnormal, the controller's protection mechanism is immediately triggered, such as output current limiting, fault alarm, or automatic power supply cut-off, to prevent chip damage due to thermal runaway. During the continuous temperature control closed-loop regulation process, the target current output by the optimized control algorithm is used to drive the double-layer thermoelectric cooling structure to continuously adjust the heat flow conduction path, so that the chip temperature converges stably to within the preset target temperature range and remains within the high-precision control tolerance for a long time.

[0026] In a specific embodiment, the process of executing step S1 may specifically include the following steps: Platinum resistance temperature sensors arranged on the surface of each chip and the hot and cold ends of the thermoelectric cooler sense temperature changes and output resistance value change signals; Based on the resistance value change signal, the constant current source circuit drives the platinum resistance temperature sensor to generate a voltage signal proportional to the temperature. The voltage signal is amplified by the signal conditioning circuit and converted into an analog temperature signal in the standard voltage range. The analog temperature signal is synchronously collected by a multi-channel analog-to-digital converter and converted into the corresponding digital temperature value, and the real-time temperature parameters containing the temperature information of each measuring point are generated according to the digital temperature value; Based on real-time temperature parameters and thermoelectric coupling heat transfer equations, the chip temperature change trend in multiple future control cycles is calculated to generate temperature prediction data.

[0027] Specifically, by arranging multiple groups of platinum resistance temperature sensors on the surface of the detector chip and at key locations on the hot and cold ends of the thermoelectric cooler, the system can sense temperature changes along the entire thermal path in a spatially covered manner. The platinum resistors used on the chip surface are arranged on the back of the chip package or under the substrate to achieve direct perception of the core heat source point, while the sensors used for the cold and hot ends of the thermoelectric cooler are installed close to the contact interface between the radiator and the casing, thereby reflecting the actual heat flow state at both ends of the cooler. These platinum resistance sensors use PT1000 specifications and have good linear response characteristics and low temperature drift errors, and can operate stably for a long time over a wide temperature range. In order to convert the temperature changes sensed in the platinum resistance element into a measurable signal, a constant current source circuit is constructed to apply a stable and constant excitation current to each sensor channel. Since the resistance value of the platinum resistor increases linearly with temperature, under constant current excitation conditions, its output voltage signal directly shows a change characteristic proportional to the temperature, enabling the voltage acquisition circuit to indirectly reflect the measured temperature in the form of voltage. The voltage signal is then amplified and filtered with high precision by a signal conditioning circuit. This circuit integrates a low-noise amplifier, a bandpass filter, and an offset voltage compensation unit. This amplifies the signal to the standard input voltage range required by the analog-to-digital converter (e.g., 0 to 3.3 volts) and filters out power-frequency noise and electromagnetic interference, improving the temperature signal's effective resolution and anti-interference capability. After analog signal conditioning, the system utilizes a highly integrated multi-channel analog-to-digital conversion module (ADC) to simultaneously sample multiple temperature channels. This ADC, integrated within the main control microprocessor, uses a 12-bit resolution ADC channel for continuous sampling at 100 Hz, ensuring that all hot node data is synchronously updated within each control cycle, avoiding temperature estimation errors caused by sampling delays. The digital temperature data output by the ADC is stored in the temperature control system's status buffer, and the processing unit generates a complete set of real-time temperature parameters containing the temperature values ​​at each measurement point. This parameter set includes key variables such as chip temperature, thermoelectric cooler cold-end temperature, hot-end temperature, and ambient background temperature. Based on the real-time temperature parameter, the control system enters the predictive calculation stage and calls the pre-established thermoelectric coupling heat transfer equation model. This model simultaneously takes into account multiple thermal effect mechanisms, including Joule heating generated by current flowing through thermoelectric materials, the Peltier effect under the combined action of current and temperature difference, and the Seebeck reverse influence effect caused by the electromotive force generated by the temperature difference. Combined with the specific structure and heat capacity parameters of the detector, the response relationship between heat flow transmission and control input is constructed in the state space expression form.By inputting the current temperature values ​​of each node as the initial state into the state equation system, the control algorithm recursively calculates the dynamic evolution trajectory of the chip temperature as the cooling current changes in multiple future control cycles in the prediction time domain. During the recursive process at each moment, the system combines factors such as heat conduction delay, cooling efficiency changes and external environmental fluctuations to correct the predicted temperature value at each step, and finally generates temperature prediction data.

[0028] In a specific embodiment, the step of calculating the chip temperature variation trend in multiple future control cycles based on the real-time temperature parameters and the thermoelectric coupling heat transfer equation to generate temperature prediction data may specifically include the following steps: Substitute the chip temperature value and hot and cold end temperature values ​​in the real-time temperature parameters into the Seebeck effect electromotive force equation to calculate the reverse electromotive force value generated by the temperature difference; Based on the chip temperature and the current current of the thermoelectric cooler, the cooling power and heating power are calculated by substituting the Peltier effect cooling power equation and the Joule heating effect heating power equation respectively. The back electromotive force value, cooling power value and heating power value are combined to establish a dynamic heat transfer model with chip temperature as the state variable; Create a state space model based on the dynamic heat transfer model, including the system matrix and the control matrix; The chip temperature response value at each moment in multiple future control cycles is calculated through recursive operation of the state space model, and the chip temperature response values ​​are combined into temperature prediction data in chronological order.

[0029] Specifically, three temperature values ​​are extracted from the real-time temperature parameters, namely the chip temperature, the cold-end temperature of the thermoelectric cooler, and the hot-end temperature. These three temperature values ​​represent the starting point of the heat source, the heat extraction node, and the heat discharge outlet in the heat transfer path, respectively. The temperature difference between them has a decisive influence on the thermoelectric effect. The difference between the cold-end temperature and the hot-end temperature is substituted into the electromotive force model constructed based on the Seebeck effect. Considering the constancy of the Seebeck coefficient in thermoelectric materials, the model describes the reverse electromotive force generated by the temperature difference. The electromotive force is proportional to the current temperature difference between the hot end and the cold end, and its direction is opposite to the direction of heat flow. Its essence is the reverse driving voltage that inhibits the thermoelectric cooler from continuing to work, so it is corrected in the power estimation. After completing the electromotive force estimation, the system uses the chip temperature value collected in the current cycle and the thermoelectric cooler driving current output by the controller, and substitutes the two into the cooling power calculation model describing the Peltier effect and the heating power model describing the Joule heating effect. The former reflects the amount of cold absorbed from the chip by the thermoelectric cooler due to the interaction between current and temperature, and the latter reflects the heat generated by the current passing through the resistance of the cooling material. The two are expressed as a positive and negative energy flux direction in the heat transfer model. The Peltier term injects cold power into the system, and the Joule term injects cold power into the system. Thermal interference; the three key thermal energy terms mentioned above - Seebeck back electromotive force, Peltier cooling power and Joule heating power - are unified and integrated into a dynamic heat transfer model based on the energy conservation and heat transfer rate equations. In this model, the chip temperature is the only controlled state variable, and the thermal conductivity and heat capacity parameters between the chip and the external heat path are considered. The differential dynamic expression of the chip temperature is constructed by establishing a functional relationship between the chip temperature change rate and the input current and environmental disturbances; based on this thermal dynamic transmission model, the system converts it into a standard form state space model, which not only contains a system matrix reflecting the internal evolution law of the chip temperature, but also introduces a control matrix reflecting the path of the control variable (i.e., the thermoelectric cooler driving current) on the chip temperature. At the same time, the external thermal disturbance term is used as the system noise or disturbance input term, and the model parameters are calibrated and reduced in combination with the actual thermal inertia and thermal resistance characteristics of the system to achieve a balance between computational complexity and modeling accuracy. A recursive calculation is performed using a state-space model. This involves using the real-time temperature parameters collected at the current moment as the model's initial state input, and the current and future drive current predictions output by the optimization solver as control input variables. Using a step-by-step time progression approach, the chip temperature response over multiple future control cycles is iteratively predicted. At each prediction step, the model substitutes the current state variables and input variables into the system matrix and control matrix to perform matrix operations, obtaining an estimated chip temperature at the next moment. This value is then fed back to the model for the next prediction, until the entire prediction time domain is covered. The system combines the chip temperature response values ​​at each prediction moment in chronological order to form a temperature prediction trajectory with a continuous time dimension.

[0030] In a specific embodiment, the step of performing recursive calculations on the state space model to calculate the chip temperature response value at each moment in multiple future control cycles and composing the chip temperature response values ​​into temperature prediction data in chronological order may specifically include the following steps: The chip temperature value in the real-time temperature parameter is set as the initial state vector of the state space model, and the prediction time domain length is set to the preset number of control cycles; Based on the system matrix and control matrix in the state space model, the state vector evolution value of each future control cycle is recursively calculated step by step starting from the current moment, where each recursive operation uses the state vector of the previous moment and the current control input; The temperature component in the state vector evolution value is extracted to obtain the chip temperature value sequence corresponding to each prediction moment, and the chip temperature value sequence is paired and combined with the corresponding time label to generate temperature prediction data.

[0031] Specifically, the current chip temperature value is extracted from the real-time temperature parameters of the thermoelectric cooling detector and used as the temperature state variable of the state-space model. Together with the remaining temperature information involved in the modeling (such as ambient temperature, cold-end or hot-end temperature), it forms the starting value of the state vector. This initial state vector represents a precise physical description of the system's current thermal state and serves as the starting point for predictive recursion input into the model solver. To clarify the time boundaries of predictive control, the system sets a fixed-length prediction horizon based on the task type and control strategy. This length is defined in units of control cycles. For example, if the prediction horizon length is set to twenty cycles, the system recursively predicts the temperature response behavior within twenty cycles based on the current moment. The system then initiates the temporal recursive operation process of the state-space model. In this model, the system matrix describes the system's state evolution in the absence of control inputs—that is, the natural heat conduction process of the chip under the current temperature and ambient conditions. The control matrix describes the direct regulatory effects of external input variables, particularly the thermoelectric cooler's drive current, on the system's state. Therefore, within each prediction cycle, the system uses the current control input value (preset by the previous optimization controller or iteratively updated) and the state vector at the previous moment to perform a recursive calculation to derive the state value for the next moment. This recursive process is implemented in discrete time. Within each control cycle, based on the current temperature state and control input, matrix multiplication and vector superposition are used to calculate the chip temperature and other relevant state indicators for the next cycle, such as the response to external thermal disturbances or the heat conduction state at the cooler. Each iteration uses the new state vector as input to initiate the next prediction, forming a progressively extended chain of state evolution. After completing the state recursion for all cycles within the entire prediction time domain, the system obtains a data sequence consisting of state vectors at multiple consecutive moments, reflecting the evolutionary trend of the system's thermal behavior under external control. Because chip temperature is a core component of the state vector, the control system extracts the corresponding component value of the chip temperature from the state vector of each prediction cycle, forming a set of temperature value sequences. This sequence describes the quantitative process of chip temperature change from the current moment as the control strategy continues to execute. To make this temperature data readable and traceable in the time dimension, the system pairs each temperature value with its corresponding prediction time tag. This time tag is based on the control cycle number or absolute timestamp, forming temperature prediction data with temporal logic.

[0032] In a specific embodiment, the process of executing step S2 may specifically include the following steps: The detector working status monitoring module reads the exposure time setting value of the current imaging task in real time to obtain the current exposure time length; Substitute the current exposure time length into the prediction time domain adjustment algorithm, and calculate the prediction time domain range through the linear combination of the basic prediction time domain length and the exposure time increment; According to the current exposure time length, the preset weight adjustment lookup table is queried to obtain the output weight coefficient in the long exposure mode and the control weight coefficient in the short exposure mode, respectively, and generate the temperature tracking weight parameter; Based on the combination of the prediction time domain range and the temperature tracking weight parameters, a control strategy including time domain configuration and weight configuration is formed.

[0033] Specifically, a real-time detector operating status monitoring module is constructed. This module maintains synchronous communication with the imaging system's main control logic, user task command interface, or task scheduling system, and is capable of periodically and proactively acquiring imaging configuration parameters. When an imaging task is about to start or the current imaging cycle has not yet ended, this module actively reads the exposure time setting for the current task. Exposure time, a core variable in the imaging control system, determines the duration of the detector's light exposure and indirectly reflects the thermal load environment of the detector chip. For example, long exposures increase the duration of continuous operation, resulting in greater chip heat generation and heat dissipation pressure; short exposures shorten the heat generation period, reduce the heat dissipation window, and require faster control response. The exposure time is input into a predictive time-domain adjustment algorithm. This algorithm, built on the basic predictive model, uses a predictive controller to dynamically adjust the time-domain length of the chip's future temperature trajectory to match the task's thermal intensity. In specific implementation, the system sets a basic prediction time domain length as a reference constant. This basic length is suitable for standard or medium thermal load scenarios. The current exposure time is then subtracted from this baseline value to obtain an exposure time increment. This increment is then linearly amplified based on a preset time domain expansion factor and added to the base length to obtain the actual prediction time domain range suitable for the current task. For example, if the exposure time is much longer than the baseline, the system's prediction time domain is significantly lengthened to more comprehensively cover the temperature evolution process of the chip under long-term operation. Conversely, if the exposure time is very short, the prediction time domain will be compressed to improve the controller's computing efficiency and response speed. Based on the current exposure time length, a preset weight adjustment lookup table is queried. This lookup table is constructed based on experimental data and system operation experience. It lists the output weight coefficients and control weight coefficients that should be used under different exposure time periods. Its division principle divides imaging tasks into three categories: long exposure, medium exposure, and short exposure. During the query process, the system extracts two key coefficients from a lookup table, corresponding to the time interval within which the current exposure time falls. The output weight coefficient measures the controller's emphasis on chip temperature tracking accuracy during optimization calculations. A larger value indicates a higher requirement for temperature trajectory fitting accuracy, making it suitable for tasks that are extremely sensitive to image quality in long exposure modes. The control weight coefficient, on the other hand, measures the constraint on the smoothness of the controller's current regulation. A larger value indicates a desire to reduce control output volatility, making it suitable for tasks that require short exposures to avoid system instability caused by sudden regulation changes. These two weight coefficients are combined into a temperature tracking weight parameter and passed to the weight matrix construction module of the controller's objective function, guiding the model predictive controller to comprehensively consider the trade-off between temperature accuracy and control smoothness during optimization. The system jointly configures the prediction time domain and the temperature tracking weight parameter to form a control strategy object, which includes multiple elements such as the prediction length, control step size, and the cost weighting structure of the objective function.

[0034] In a specific embodiment, the process of executing step S3 may specifically include the following steps: Substituting the chip temperature values ​​at each prediction moment in the temperature prediction data into a thermal electron noise calculation model including a temperature coefficient and a reference noise parameter, and calculating the thermal electron noise intensity value at the corresponding moment; Based on the thermal electron noise intensity value, a noise response curve reflecting the noise variation with temperature is constructed to obtain the corresponding relationship data between temperature and noise level; The preset maximum allowable noise level threshold is queried according to the detector imaging quality standard, and a reverse lookup operation is performed in combination with the noise response curve to determine the chip temperature upper limit constraint value that meets the imaging quality requirements.

[0035] Specifically, the temperature prediction data consists of a set of chronologically ordered temperature values, each corresponding to a chip temperature state point in a future control cycle within the prediction time domain. This data covers the chip's thermal response trend under control from the current moment to several future moments. The system extracts the chip temperature value at each predicted moment and sequentially substitutes it into a thermal electron noise calculation model to estimate the thermal noise. This noise model is a nonlinear response function based on empirical modeling and physical laws. It includes key parameters such as reference temperature, reference noise intensity, and a temperature exponential factor. It can simulate and calculate electronic noise fluctuations caused by temperature changes based on different chip material properties and package thermal response characteristics. The model structure reflects a positive correlation between thermal electron noise intensity and chip temperature, with a certain accelerating growth trend. This trend indicates that as temperature increases, electron thermal excitation increases, the frequency of carriers' free motion within the semiconductor increases, and thus the noise floor amplitude in signal detection increases. Therefore, the control system needs to understand the noise value corresponding to each predicted temperature to determine whether the system is at risk of exceeding the noise standard. After the system completes the noise calculation for all chip temperature values ​​at predicted moments, it obtains a sequence of thermal electron noise intensity values ​​that is perfectly aligned with the time axis. This sequence serves as the basic data for constructing a mapping between temperature and noise level. This mapping uses temperature as the horizontal axis and noise intensity as the vertical axis, forming a noise response curve that describes how the chip noise response changes with temperature. This curve exhibits a nonlinear upward trend, and its shape and slope are determined by the temperature coefficient and reference parameters in the model. Through this response curve, the system can intuitively observe the amplification effect of small temperature changes on noise intensity. To apply this curve to actual control constraints, the system will call the maximum allowable noise level threshold preset in the detector imaging quality specification or mission standard. This threshold defines the maximum noise intensity boundary that the detection system can tolerate in a specific imaging application scenario. If the predicted noise exceeds this threshold during operation, it will result in blurred imaging, the target signal will be submerged in background noise, or the signal-to-noise ratio will be reduced to a level that cannot meet recognition requirements. Therefore, this threshold constitutes a rigid safety limit in the design of the temperature control system. This threshold is combined with the noise response curve constructed above. The specific process involves performing a reverse lookup on the curve, starting from the vertical axis and using the maximum allowable noise level as a reference. The corresponding horizontal axis value on the response curve is then searched downwards to find the maximum acceptable chip temperature value without exceeding this noise level. This temperature point serves as the required upper chip temperature constraint. This upper limit is dynamically adaptive and is calculated in real time based on the imaging requirements of different imaging tasks and the noise performance parameters of different chip components. Therefore, the accuracy of the noise boundary is adjusted according to each imaging configuration.

[0036] In a specific embodiment, the process of executing step S4 may specifically include the following steps: A target function containing a temperature deviation term and a control increment term is constructed based on a temperature tracking weight parameter in the control strategy and a prediction time domain range; The chip temperature upper limit constraint value is converted into inequality constraint conditions of the chip temperature at each time in the prediction time domain, and a complete constraint condition group is formed by adding a thermoelectric cooler current amplitude limit and a current change rate limit; The target function and the constraint condition group are combined to form a constraint optimization equation, and the constraint optimization equation is numerically solved by an active set algorithm to obtain the driving current value of the current period.

[0037] Specifically, based on the running results of the prediction model and the parameters in the control strategy file, a mathematical expression system required for optimization solution is constructed. The control strategy clearly gives the temperature tracking weight parameters and prediction time domain range under the current imaging task. The former characterizes the evaluation criteria of the system's sensitivity to the error between the chip temperature and the target temperature, and the latter defines the number of future control cycles that the model prediction should cover. The two together determine the structure of the objective function and the depth of optimization. According to the set prediction time domain range, the system defines a time-weighted cumulative expression in the objective function to measure the overall deviation between the current predicted temperature trajectory and the expected target temperature, and uses the temperature tracking weight to weight the deviation amplitude of each prediction point, thereby giving priority to suppressing the deviation section in the temperature change trend during the optimization process to ensure that the entire trajectory fits the target temperature curve as closely as possible. To prevent system oscillation or drive circuit overload caused by drastic changes in the controller's output current between cycles, a control increment term reflecting the smoothness of the control output is added to the objective function. This term is the sum of the squares of the current changes between two consecutive cycles, and its optimization priority is adjusted by a control weight parameter. This imposes a smoothness penalty on the controller's output behavior, ensuring that the resulting current sequence maintains strong output coherence and execution stability while meeting temperature tracking requirements. These two components together constitute the optimization objective function, whose overall goal is to minimize the cost function caused by temperature deviation and control variation, thereby guiding the controller to find the optimal drive current command. The chip temperature upper limit constraint is converted into an inequality constraint for the chip temperature at each moment in the prediction time domain. This constraint forces the chip temperature at any moment in the prediction to remain within the noise-controllable region throughout the prediction trajectory, thereby preventing imaging performance degradation caused by abnormal temperature rise. To enhance the physical feasibility and practical controllability of the optimization process, the system incorporates hardware-driven constraints on the thermoelectric cooler itself into the constraint group. These constraints include current amplitude limits and current rate limits. The former stipulates that the controller's output current must fluctuate between the maximum drive capability of the thermoelectric element and the minimum sustaining current, preventing overload or burnout of the cooler due to excessive current, and heat accumulation due to insufficient cooling capacity due to insufficient current. The latter stipulates that the control current change between adjacent cycles should not exceed the driver's response range, avoiding abrupt adjustments that could cause a decrease in control accuracy or distortion in the drive response. The objective function is combined with the constraint group to form a constrained optimization equation, and the active set algorithm is used as the solution strategy. This algorithm is suitable for solving minimization problems with quadratic objective functions and linear inequality constraints. It constructs an active constraint set and dynamically adjusts the currently active constraint boundaries during the iteration process, thereby continuously approaching the optimal solution within the constrained feasible region.During execution, the algorithm determines the initial solution and its corresponding set of active constraints. It then verifies the current constraint system by solving a system of Lagrange multipliers. If any inequality bounds are violated, the algorithm incorporates them into the current active set and adjusts the search direction. If all conditions are met, the current solution is output as a candidate optimal solution. The system executes this optimization process only once per control cycle. After obtaining a complete set of control inputs for future cycles, the first current value in the sequence is taken as the actual control output for the current cycle.

[0038] In a specific embodiment, the execution step combines the objective function with the constraint condition group to form a constrained optimization equation, and numerically solves the constrained optimization equation to obtain the driving current value of the current cycle. The process may specifically include the following steps: Based on the temperature deviation term and the control increment term in the objective function and the inequality constraints and equality constraints in the constraint condition group, a constrained optimization equation including decision variables, an objective coefficient matrix and a constraint coefficient matrix is ​​constructed; An initial feasible point is selected based on the current state of the thermoelectric cooler, and the active constraints of the initial feasible point are identified to establish an initial valid constraint set, and the corresponding Lagrange multiplier vector is calculated. By solving the linear equations, the search direction under the current valid constraint set is calculated to determine whether the search direction meets the optimality condition. If not, the decision variables are updated along the search direction and the valid constraint set is adjusted for the next round of iterative operations. When all Lagrange multiplier vectors are non-negative and the search direction is a zero vector during the iteration process, the iteration is terminated, and the corresponding thermoelectric cooler current component in the decision variable at this time is output as the driving current value of the current cycle.

[0039] Specifically, based on key inputs such as the preceding control strategy, temperature prediction data, and the chip temperature upper limit constraint generated by the thermal noise model, a mathematical optimization expression for predictive control of thermoelectric cooling systems was established. The core task was to construct a minimization problem with constrained boundaries. This problem uses multi-cycle control input variables as decision variables, and aims to minimize chip temperature error and control smoothness, while complying with various physical boundaries and noise safety constraints. The decision variable is the thermoelectric cooler current input sequence corresponding to each control cycle in the prediction time domain. The length of this sequence is equal to the prediction time domain step size, and it participates in the optimization solution as a column vector. Based on a temperature prediction model, the system models the chip temperature response to different input current sequences, expressing the temperature deviation as a function of the current sequence and the target temperature. The system then constructs the two main components of the objective function using a temperature tracking weight matrix and a control increment weight matrix. The first component measures the degree of deviation between the predicted temperature trajectory and the target temperature, while the second measures the magnitude of the control current variation between consecutive cycles. These two components are constructed as standard quadratic objective functions. The objective coefficient matrix is ​​generated based on the model parameters, time distribution, and weighting settings, forming a complete cost function. All hard constraints are integrated to construct a constraint set consisting of three types of constraints: a chip temperature upper limit constraint, derived from the thermal noise model and setting a temperature upper limit at each moment in the prediction time domain; a thermoelectric cooler current amplitude limit, which ensures that the current within each cycle must not exceed the minimum and maximum limits allowed by the driver circuit; and a control increment limit, which ensures that the current variation between consecutive cycles must not exceed preset limits. These three constraints are mostly linear inequalities. After coefficient normalization and matrixization, the constraint coefficient matrix and constraint vector are constructed. The system initiates an iterative solution process based on the active set algorithm. At the beginning of this process, the system selects an initial feasible point based on the actual drive current state of the thermoelectric cooler. This feasible point satisfies the boundary conditions of all inequality constraints and is as close as possible to the target minimization direction, thereby shortening the solution convergence time. At this point, the system identifies the boundary position of the selected initial point and determines which constraints are currently being precisely violated at this point. These constraints are then marked as active and the current active constraint set is established. A Lagrange multiplier system is constructed within this active constraint set and solved to obtain the multiplier vector at the current point, which is used to determine the direction and strength of each constraint's influence on the current target value. After solving the Lagrange multiplier system, the system further determines the search direction within the current active constraint set. This search direction guides the decision variables in the multidimensional solution space toward the direction of descent of the objective function. This search direction is determined by constructing and solving a set of linear equations related to the target coefficient matrix and the constraint matrix. The resulting solution vector serves as the direction for the next variable update.The optimality of the search direction is determined. If the direction is a zero vector and all Lagrange multipliers are non-negative, the optimality condition is met at the current point and no better feasible solution exists. Otherwise, the system will perform a linear step update on the decision variable along the current search direction, recalculate the active constraint set after the update, and reconstruct the linear equation system to enter the next round of iteration. During the continuous iteration process, the system continuously approaches the minimum value of the objective function, while dynamically adjusting the active constraint set to ensure that the boundary conditions are always met. Until one iteration finds that the search direction is zero and all Lagrange multipliers are non-negative values, the system determines that the current decision variable is the global optimal solution that meets all constraints. At this time, the current component corresponding to the current control cycle is extracted from the decision variable and output as the driving current value of this cycle to the thermoelectric cooling system, thereby driving the execution circuit to generate matching cooling power, completing the closed-loop regulation task of chip temperature control.

[0040] In a specific embodiment, the process of executing step S5 may specifically include the following steps: Perform linear conversion from digital to analog on the driving current value to generate a corresponding reference voltage control signal; Inputting the reference voltage control signal into the thermoelectric cooler driving circuit to adjust the current and generate a stable thermoelectric cooler driving current; The stable thermoelectric cooler driving current is simultaneously delivered to the first layer of thermoelectric cooler close to the chip surface and the second layer of thermoelectric cooler close to the detector housing in the double-layer thermoelectric cooling structure according to a preset distribution ratio. The cold end of the first layer of thermoelectric cooler directly absorbs the heat generated by the chip and conducts it to the hot end. The second layer of thermoelectric cooler receives the heat conducted by the first layer and dissipates it to the external environment. The chip temperature is precisely adjusted to the target temperature range through a double-layer thermoelectric cooling structure. At the same time, the operating current and temperature status of each layer of thermoelectric cooler are monitored to ensure stable operation of the system and complete chip temperature control.

[0041] Specifically, the drive current output by the optimization controller is converted from its original digital signal form into an analog control signal that can be directly recognized and responded to by the drive circuit. This process is accomplished by a digital-to-analog conversion module configured in the main control processor. This module performs linear amplitude mapping based on the digital current output of the controller. Specifically, it sets the maximum and minimum current values ​​according to the current control range and establishes a one-to-one correspondence with the upper and lower limits of the analog output voltage. This allows each digital current value to be mapped in real time into a set of analog voltage signals, which serve as the reference control voltage required by the drive circuit. This reference voltage control signal is input into the thermoelectric cooler drive circuit for current regulation. This drive circuit utilizes a linear constant current source structure or a high-precision current regulation chip, capable of stably converting the input voltage signal into a corresponding control current output. This circuit internally achieves continuously adjustable output current through a precision voltage-to-current conversion stage. Its operating principle is to monitor the current output in real time in a feedback loop and compare it with a reference value set by the target control voltage. Through error amplification and power regulation, a closed-loop control circuit is formed to dynamically adjust the current output of the power path until the output current is fully consistent with the target value. The driving current of the stable thermoelectric cooler is simultaneously delivered according to a preset distribution ratio to a two-stage, series-connected, dual-layer thermoelectric cooling structure. In this structure, the first layer of thermoelectric coolers is physically bonded to the bottom of the detector chip, with its cold end facing the chip itself. Thermally coupled via a high-thermal-conductivity interface material, the first layer preferentially absorbs heat released during chip operation under the action of the cooling current and transfers this heat to its own hot end along the heat flow path. The second layer of thermoelectric coolers is mounted within the detector's metal casing. Its cold end connects to the hot end of the first layer via a highly thermally conductive intermediary (such as a copper sheet) to form a heat transfer interface, while its hot end faces an external heat sink or passive cooling structure, further discharging heat transferred from the first layer out of the system, completing a complete heat transfer path from the chip to the environment. By adjusting the current distribution ratio between the two layers of thermoelectric coolers, the system ensures that the first layer focuses on fast response and localized cooling, while the second layer handles long-term heat sinking and steady-state heat output, creating a layered cooling structure that is both fast inside and stable outside. To ensure system stability and continuous control during the operation of this cooling structure, the controller monitors the electrical operating status and thermal response of the two layers of thermoelectric coolers in real time. The current state monitoring module measures the actual instantaneous current flowing through each layer of thermoelectric cooler using current sampling resistors or Hall current sensors, and compares this current with the target current to determine whether there is any drive deviation or power supply anomaly. Simultaneously, temperature sensors located on the hot and cold ends of the thermoelectric cooler simultaneously measure the temperature of each layer's hot and cold ends, reflecting the heat transfer efficiency and cooling effect. If temperature anomalies or reduced cooling efficiency occur, the system will issue a warning and adjust the control strategy to prevent chip temperature runaway.At the same time, the temperature of the chip body is continuously collected through a high-precision platinum resistance sensor as feedback basis for the system's final adjustment target value. Under the guidance of the predictive control algorithm, the drive current command is continuously corrected to keep the system in closed-loop adaptive operation.

[0042] The above describes the control method of the low-noise thermoelectric cooling type detector in the embodiment of the present invention. The following describes the control device of the low-noise thermoelectric cooling type detector in the embodiment of the present invention. Figure 2 An embodiment of a control device for a low-noise thermoelectric cooling detector according to an embodiment of the present invention includes: The acquisition module is used to collect the real-time temperature parameters of the detector and calculate the temperature prediction data; The adjustment module is used to detect the current exposure time length of the detector and form a corresponding control strategy; A calculation module is used to calculate thermal electron noise based on temperature prediction data and determine an upper limit constraint value of chip temperature that meets imaging quality requirements; A solution module, used for solving the driving current value of the current cycle based on the control strategy and the chip temperature upper limit constraint value; The output module is used for the thermoelectric cooler driving circuit to adjust the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control.

[0043] Through the collaborative efforts of these components, a rolling-horizon model predictive control algorithm is employed to predict chip temperature trends over multiple future control cycles. This proactive, preventative control capability, compared to the passive feedback regulation of traditional PID control, significantly improves the dynamic response and steady-state accuracy of temperature control. By comprehensively considering the physical coupling of the Seebeck, Peltier, and Joule heating effects, a precise dynamic heat transfer model is established. Compared to the simplified heat transfer modeling used in existing techniques, this model more accurately describes the actual operating characteristics of the thermoelectric cooler. Thermal electron noise levels are incorporated as hard constraints into the control optimization process, enabling coordinated optimization of temperature control and noise suppression, resolving the technical challenge of conflicting temperature accuracy and noise performance in traditional control methods. Dynamic adjustment of the prediction horizon length and control weight parameters based on the detector's exposure mode and operating conditions allows for adaptability to diverse imaging tasks, enhancing the system's flexibility and applicability compared to fixed-parameter control strategies. A cascaded dual-layer thermoelectric cooler structure is employed, with the first layer directly bonded to the chip for rapid heat absorption and the second layer bonded to the housing for efficient heat dissipation. This provides enhanced cooling capacity and better temperature uniformity than a single-layer structure. Through high-precision digital-to-analog conversion and precision current source drive, precise regulation of the thermoelectric cooler current is achieved, ensuring that the control algorithm output can be accurately converted into actual cooling effect, thereby improving the control accuracy and reliability of the overall system.

[0044] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0045] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0046] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control method for a low-noise thermoelectric cooling detector, characterized in that: include: Collect the real-time temperature parameters of the detector and calculate the temperature prediction data; Detect the current exposure time length of the detector and form a corresponding control strategy; Performing thermal electron noise calculation based on the temperature prediction data to determine an upper limit constraint value of the chip temperature that meets imaging quality requirements; Calculating a driving current value of a current cycle based on the control strategy and the chip temperature upper limit constraint value; The thermoelectric cooler driving circuit adjusts the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control.

2. The control method of the low-noise thermoelectric cooling type detector according to claim 1, characterized in that: The collecting of real-time temperature parameters of the detector and calculation of temperature prediction data include: Platinum resistance temperature sensors arranged on the surface of each chip and the hot and cold ends of the thermoelectric cooler sense temperature changes and output resistance value change signals; Based on the resistance value change signal, the platinum resistance temperature sensor is driven by a constant current source circuit to generate a voltage signal proportional to the temperature, and the voltage signal is amplified by a signal conditioning circuit and converted into an analog temperature signal within a standard voltage range; The analog temperature signals are synchronously collected by a multi-channel analog-to-digital converter and converted into corresponding digital temperature values, and real-time temperature parameters containing temperature information of each measuring point are generated according to the digital temperature values; The chip temperature variation trend in multiple future control cycles is calculated based on the real-time temperature parameters and the thermoelectric coupling heat transfer equation to generate temperature prediction data.

3. The control method of the low-noise thermoelectric cooling type detector according to claim 2, characterized in that: The step of calculating the chip temperature variation trend in multiple future control cycles based on the real-time temperature parameters and the thermoelectric coupling heat transfer equation to generate temperature prediction data includes: Substituting the chip temperature value and the hot and cold end temperature values ​​in the real-time temperature parameters into the Seebeck effect electromotive force equation to calculate the reverse electromotive force value generated by the temperature difference; Substituting the chip temperature value and the current thermoelectric cooler current into the Peltier effect cooling power equation and the Joule heating effect heating power equation to calculate the cooling power value and the heating power value respectively; Combining the back electromotive force value, the cooling power value, and the heating power value to establish a dynamic heat transfer model with the chip temperature as a state variable; creating a state space model including a system matrix and a control matrix based on the dynamic heat transfer model; The chip temperature response value at each moment in a plurality of future control cycles is calculated by recursive operation using the state space model, and the chip temperature response values ​​are combined into temperature prediction data in chronological order.

4. The control method of the low-noise thermoelectric cooling type detector according to claim 3, characterized in that: The recursive calculation of the chip temperature response value at each moment in a plurality of future control cycles by the state space model and the formation of temperature prediction data by the chip temperature response values ​​in chronological order include: Setting the chip temperature value in the real-time temperature parameter as the initial state vector of the state space model, and setting the prediction time domain length to a preset number of control cycles; Based on the system matrix and the control matrix in the state space model, recursively calculate the state vector evolution value of each future control cycle starting from the current moment, wherein each recursive operation uses the state vector at the previous moment and the current control input; The temperature component in the state vector evolution value is extracted to obtain a chip temperature value sequence corresponding to each prediction moment, and the chip temperature value sequence is paired and combined with the corresponding time tag to generate temperature prediction data.

5. The control method of the low-noise thermoelectric cooling type detector according to claim 1, characterized in that: The detecting the current exposure time length of the detector and forming a corresponding control strategy includes: The detector working status monitoring module reads the exposure time setting value of the current imaging task in real time to obtain the current exposure time length; Substituting the current exposure time length into the prediction time domain adjustment algorithm, and calculating the prediction time domain range by a linear combination of the basic prediction time domain length and the exposure time increment; Querying a preset weight adjustment lookup table according to the current exposure time length, obtaining an output weight coefficient corresponding to the long exposure mode and a control weight coefficient corresponding to the short exposure mode, and generating a temperature tracking weight parameter; Based on the combination of the prediction time domain range and the temperature tracking weight parameter, a control strategy including time domain configuration and weight configuration is formed.

6. The control method of the low-noise thermoelectric cooling type detector according to claim 1, characterized in that: The performing thermal electron noise calculation based on the temperature prediction data to determine the chip temperature upper limit constraint value that meets the imaging quality requirements includes: Substituting the chip temperature values ​​at each prediction moment in the temperature prediction data into a thermal electron noise calculation model including a temperature coefficient and a reference noise parameter, and calculating the thermal electron noise intensity value at the corresponding moment; Based on the thermal electron noise intensity value, a noise response curve reflecting the noise variation with temperature is constructed to obtain corresponding relationship data between temperature and noise level; The preset maximum allowable noise level threshold is queried according to the detector imaging quality standard, and a reverse lookup operation is performed in combination with the noise response curve to determine the chip temperature upper limit constraint value that meets the imaging quality requirements.

7. The control method of the low-noise thermoelectric cooling type detector according to claim 1, characterized in that: The step of solving the driving current value of the current cycle based on the control strategy and the chip temperature upper limit constraint value includes: Constructing an objective function including a temperature deviation term and a control increment term based on the temperature tracking weight parameter and the prediction time domain range in the control strategy; Converting the chip temperature upper limit constraint value into an inequality constraint condition for the chip temperature at each moment in the prediction time domain, and adding the thermoelectric cooler current amplitude limit and current change rate limit to form a complete constraint condition group; The objective function is combined with the constraint condition group to form a constrained optimization equation, and the constrained optimization equation is numerically solved to obtain the driving current value of the current cycle.

8. The control method of the low-noise thermoelectric cooling type detector according to claim 7, characterized in that: The step of combining the objective function with the constraint condition group to form a constrained optimization equation and numerically solving the constrained optimization equation to obtain the driving current value of the current cycle includes: Based on the temperature deviation term and the control increment term in the objective function and the inequality constraints and equality constraints in the constraint condition group, constructing a constrained optimization equation including decision variables, an objective coefficient matrix, and a constraint coefficient matrix; Selecting an initial feasible point based on the current state of the thermoelectric cooler, identifying active constraints of the initial feasible point to establish an initial valid constraint set, and calculating a corresponding Lagrange multiplier vector; By solving the linear equations, the search direction under the current valid constraint set is calculated to determine whether the search direction meets the optimality condition. If not, the decision variables are updated along the search direction and the valid constraint set is adjusted for the next round of iterative operations. When all Lagrange multiplier vectors are non-negative and the search direction is a zero vector during the iteration process, the iteration is terminated, and the corresponding thermoelectric cooler current component in the decision variable at this time is output as the driving current value of the current cycle.

9. The control method of the low-noise thermoelectric cooling type detector according to claim 1, characterized in that: The thermoelectric cooler driving circuit adjusts the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control, including: Performing linear conversion from digital to analog on the driving current value to generate a corresponding reference voltage control signal; Inputting the reference voltage control signal into the thermoelectric cooler drive circuit for current regulation to generate a stable thermoelectric cooler drive current; The stable thermoelectric cooler driving current is simultaneously delivered to the first layer of thermoelectric cooler in close contact with the chip surface and the second layer of thermoelectric cooler in close contact with the detector housing in the double-layer thermoelectric cooling structure according to a preset distribution ratio, so that the cold end of the first layer of thermoelectric cooler directly absorbs the heat generated by the chip and conducts it to the hot end, and the second layer of thermoelectric cooler receives the heat conducted by the first layer and dissipates it to the external environment; The chip temperature is precisely adjusted to the target temperature range through the double-layer thermoelectric cooling structure, and the working current and temperature status of each layer of thermoelectric cooler are monitored to ensure stable operation of the system and complete chip temperature control.

10. A control device for a low-noise thermoelectric cooling detector, characterized in that: A control method for executing a low-noise thermoelectric cooling detector according to any one of claims 1 to 9, comprising: The acquisition module is used to collect the real-time temperature parameters of the detector and calculate the temperature prediction data; The adjustment module is used to detect the current exposure time length of the detector and form a corresponding control strategy; a calculation module, configured to calculate thermal electron noise based on the temperature prediction data and determine an upper limit constraint value of the chip temperature that meets imaging quality requirements; A solution module, configured to solve a driving current value of a current cycle based on the control strategy and the chip temperature upper limit constraint value; The output module is used for the thermoelectric cooler driving circuit to adjust the cooling power of the double-layer thermoelectric cooling structure according to the driving current value to complete the chip temperature control.

Images