Photovoltaic inverter heat dissipation regulation control system

Abstract

The present application relates to the field of power electronics, specifically to a photovoltaic inverter heat dissipation regulation control system. It includes a multimodal state prediction module for receiving real-time collected environmental data and inverter historical working state data to predict the predicted junction temperature in one or more time steps in the future; a collaborative optimization decision module for receiving the predicted junction temperature to obtain the optimal control parameter combination and the expected steady-state junction temperature; a feedforward control execution module for receiving the optimal control parameter combination and setting the optimal fan speed in the optimal control parameter combination as the reference set value of the heat dissipation system speed controller; a feedback correction module for obtaining the fan speed correction amount; and an instruction synthesis and sending module for sending the final target speed instruction to the underlying speed controller of the fan. The system avoids the temperature overshoot and frequent start-stop problem of the heat dissipation system caused by traditional passive response control, and improves the stability of system operation.

Description

Technical Field

[0001] This invention relates to the field of power electronics technology, specifically to a heat dissipation regulation and control system for a photovoltaic inverter. Background Technology

[0002] In current photovoltaic power generation systems, the inverter, as the core power conversion device, is crucial to the economic benefits of the entire system due to its operating efficiency, reliability, and service life. The junction temperature of power semiconductor devices is a key factor affecting the above performance indicators, so precise temperature regulation through a heat dissipation system is necessary. Traditional heat dissipation control strategies are mostly passive response-based, that is, adjusting heat dissipation measures such as fan speed based on real-time temperature feedback. This approach fails to systematically balance the complex coupling and conflicting relationships between power generation revenue, heat dissipation power consumption, and long-term device reliability, often leading to unreasonable resource allocation and difficulty in maximizing the overall net benefits throughout the system's entire life cycle.

[0003] In existing technologies, some heat dissipation control methods, although introducing simple predictive models or multi-objective control ideas, generally suffer from low accuracy in predicting future operating conditions and a single optimization dimension. For example, most control strategies separate the heat dissipation system from the inverter's operating point, such as switching frequency and MPPT strategy, and control them independently, ignoring the strong coupling between them. This results in optimization results being limited to local optima and failing to achieve overall performance improvement at the system level. In addition, when faced with drastic dynamic changes in environmental factors such as light and ambient temperature, traditional methods lack a forward-looking pre-adjustment mechanism, which can easily lead to problems such as temperature overshoot or frequent start-stop, not only increasing unnecessary energy consumption but also accelerating the thermal fatigue aging of power devices.

[0004] Therefore, how to provide a photovoltaic inverter heat dissipation regulation and control method that can proactively predict heat load and synergistically optimize multiple control variables is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] To solve the above-mentioned technical problems, this invention discloses a heat dissipation regulation and control system for photovoltaic inverters. Specifically, the technical solution of this invention is as follows:

[0006] A photovoltaic inverter heat dissipation regulation and control system includes:

[0007] The multimodal state prediction module is used to receive real-time environmental data and inverter historical operating state data. The environmental data includes light intensity and ambient temperature, and the inverter historical operating state data includes current junction temperature and output power. Based on a preset discrete-time dynamic thermal model, it predicts the junction temperature for one or more future time steps.

[0008] The collaborative optimization decision module receives the predicted junction temperature and, based on the cost function aimed at maximizing the overall net energy efficiency of the system, solves and analyzes the switching frequency, MPPT disturbance strategy, and fan speed to obtain the optimal combination of control parameters and the desired steady-state junction temperature.

[0009] The feedforward control execution module is used to receive the optimal control parameter combination, set the optimal fan speed in the optimal control parameter combination as the reference set value of the cooling system speed controller, send the optimal switching frequency to the main control unit of the inverter for adjusting the drive signal frequency of the power devices, and send the optimal MPPT disturbance strategy to the maximum power point tracking controller for updating its core parameters.

[0010] The feedback correction module is used to monitor the deviation between the actual junction temperature and the desired steady-state junction temperature in order to obtain the fan speed correction amount;

[0011] The instruction synthesis and transmission module is used to synthesize the reference setpoint and the fan speed correction value to obtain the final target speed instruction, and send the final target speed instruction to the fan's underlying speed controller;

[0012] Preferably, the multimodal state prediction module is further used for:

[0013] By using a system identification algorithm and offline collected historical operating data, the model coefficients of the discrete-time dynamic thermal model are fitted and calibrated.

[0014] The correction terms of the discrete-time dynamic thermal model are updated in real time using an online adaptive algorithm.

[0015] Preferably, the cost function unifies power generation revenue, heat dissipation costs, and operational risks under a single optimization objective. The cost function specifically includes:

[0016] The product of conversion efficiency and input power is used to characterize the revenue generated from power generation.

[0017] The heat dissipation and power consumption term is used to characterize the cost of heat dissipation;

[0018] The temperature penalty function term is used to quantify the risk of equivalent power loss due to excessive junction temperature;

[0019] Preferably, the temperature penalty function is used for:

[0020] The predicted junction temperature is compared with the preset safety threshold.

[0021] When the predicted junction temperature is below the safe threshold, the temperature penalty function is set to zero;

[0022] When the predicted junction temperature exceeds the safety threshold, the temperature penalty function is set to an exponential form.

[0023] Preferably, the collaborative optimization decision module is further used for:

[0024] The cost function is solved using a numerical optimization algorithm.

[0025] Preferably, the feedback correction module is specifically used for:

[0026] A proportional-integral controller is used;

[0027] Determine the deviation between the actual junction temperature and the desired steady-state junction temperature;

[0028] Based on the deviation, the fan speed correction amount is calculated;

[0029] Preferably, the calculation process for the final target speed command is as follows:

[0030] The optimal fan speed and the fan speed correction amount are combined to obtain the final target speed command;

[0031] The final target speed command is used to drive the fan.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. By establishing a dynamic thermal model, this system can proactively predict future junction temperature changes of key power devices. This prediction-based active adjustment method effectively avoids the temperature overshoot and frequent start-stop problems of the heat dissipation system caused by traditional passive response control, thus improving the stability of system operation.

[0034] 2. This system constructs a decision-making mechanism that unifies power generation revenue, heat dissipation costs, and operational risks under a single optimization objective; by coordinating the optimization of multiple interrelated control variables such as switching frequency, MPPT strategy, and fan speed, it breaks through the limitations of independent decision-making in each link of traditional control, and maximizes the comprehensive net power revenue throughout the system's entire life cycle.

[0035] 3. This system incorporates a temperature penalty function into its optimization objectives, which can quantify and mitigate the risk of power device aging caused by excessive junction temperature. When the predicted temperature exceeds the safety threshold, this mechanism will significantly increase the risk cost, thereby balancing the pursuit of immediate power generation efficiency with the long-term reliability of the equipment and effectively extending the service life of the inverter.

[0036] 4. This system adopts a feedforward-feedback composite control architecture, which combines the forward-looking nature of feedforward control with the precision of feedback correction. This architecture can not only actively respond to predictable changes in operating conditions, but also compensate for the influence of model errors and external unknown disturbances through real-time deviation correction, ensuring that the actual junction temperature can accurately and stably track the dynamic target, and significantly enhancing the control accuracy and robustness of the system. Attached Figure Description

[0037] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0038] Figure 1 This is a flowchart of the system of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0040] Example 1:

[0041] Please see Figure 1 A photovoltaic inverter heat dissipation regulation and control system includes:

[0042] The multimodal state prediction module is used to receive real-time environmental data and inverter historical operating state data. The environmental data includes light intensity and ambient temperature, and the inverter historical operating state data includes current junction temperature and output power. Based on a preset discrete-time dynamic thermal model, it predicts the junction temperature for one or more future time steps.

[0043] The collaborative optimization decision module receives the predicted junction temperature and, based on the cost function aimed at maximizing the overall net energy efficiency of the system, solves and analyzes the switching frequency, MPPT disturbance strategy, and fan speed to obtain the optimal combination of control parameters and the desired steady-state junction temperature.

[0044] The feedforward control execution module is used to receive the optimal control parameter combination and set the optimal fan speed in the optimal control parameter combination as the reference set value of the cooling system speed controller.

[0045] The feedback correction module is used to monitor the deviation between the actual junction temperature and the desired steady-state junction temperature in order to obtain the fan speed correction amount;

[0046] The instruction synthesis and transmission module is used to synthesize the reference setpoint and the fan speed correction value to obtain the final target speed instruction, and send the final target speed instruction to the fan's underlying speed controller;

[0047] This invention discloses a photovoltaic inverter heat dissipation regulation and control system. Its technical objective is to maximize the comprehensive net power gain of the inverter system throughout its entire life cycle while ensuring the high reliability of power devices. The system includes a multi-mode state prediction module, a collaborative optimization decision module, a feedforward control execution module, and a feedback correction module. Through the collaborative work of each module, robust and precise collaborative control of the photovoltaic inverter heat dissipation and operating point is achieved.

[0048] This multimodal state prediction module receives real-time environmental data and historical inverter operating status data to predict the junction temperature of key power semiconductors in the inverter. The environmental data received includes light intensity and ambient temperature, while the historical inverter operating status data includes the current junction temperature and output power. Based on a pre-defined, simplified discrete-time dynamic thermal model for real-time calculation, the module predicts the junction temperature of key power semiconductors within one or more future time steps. The initial form of this discrete-time dynamic thermal model originates from a first-order lumped-parameter thermal model describing the heat transfer process of the power module, and is constructed after linearization and discretization as follows:

[0049]

[0050] in: : Predicted junction temperature, in Kelvin (K);

[0051] Current junction temperature, in Kelvin (K), is obtained in real time by the temperature sensor inside the inverter;

[0052] Output power, measured in watts (W), is acquired in real time by the power monitoring equipment inside the inverter.

[0053] Light intensity, measured in watts per square meter (W / m²), is collected in real time by sensors on the external environment of the photovoltaic array.

[0054] Ambient temperature, measured in Kelvin (K), is obtained in real time through external environmental sensors;

[0055] Prediction step size, in seconds (s), is preset according to the system control cycle;

[0056] Model coefficients are dimensionless parameters.

[0057] Model coefficients, in Kelvin per watt (K / W).

[0058] Model coefficients, in Kelvin per square meter per watt (K·m² / W).

[0059] Model coefficients are dimensionless parameters.

[0060] The values ​​of these coefficients are obtained by fitting and calibrating historical operating data collected offline through a system identification algorithm to reflect the thermal characteristics of a specific inverter.

[0061] Model coefficients, in Kelvin (K), represent the heat dissipation efficiency of the cooling system; This is a dimensionless heat dissipation function related to the current fan speed. It can be a simple linear or polynomial function, representing the contribution of fan speed to cooling, and its range is typically [0,1]. For example, it can be a normalized polynomial function, such as...

[0062]

[0063] in: This is the fan's maximum speed, coefficient. Calibration was achieved through experiments; : Correction term, in Kelvin (K); This correction term is updated in real time through an online adaptive algorithm to compensate for long-term drift of the model caused by changes in operating conditions;

[0064] The model calculates the predicted junction temperature. It is transmitted as core information to the collaborative optimization decision-making module;

[0065] The collaborative optimization decision-making module is connected to the output of the multimodal state prediction module. Its core purpose is to unify the three conflicting performance indicators of power generation revenue, heat dissipation cost, and operational risk under a single optimization objective, so as to maximize the overall net energy efficiency of the system. This module receives the predicted junction temperature. Based on a cost function aimed at maximizing the overall net energy efficiency of the system, the switching frequency, MPPT disturbance strategy, and fan speed are analyzed and solved to obtain a set of optimal control parameter combinations. And the expected steady-state junction temperature under these optimal parameters. ;

[0066] The expected steady-state junction temperature It is the expected equilibrium temperature calculated by substituting the optimal combination of control parameters into the system thermal model under the current environmental conditions. Specifically, this steady-state temperature can be obtained by solving the steady-state equation of the system thermal model, that is, by setting the predicted junction temperature in the dynamic thermal model equal to the current junction temperature. By substituting the current environmental data and the system state corresponding to the optimal control parameters, the desired steady-state junction temperature can be solved. Its solution expression is:

[0067]

[0068] The feedforward control execution module is connected to the output of the collaborative optimization decision module, aiming to achieve predictive and forward-looking control. This module will output the optimal fan speed from the optimal combination of control parameters from the collaborative optimization decision module. It is directly used as the reference setting value for the speed controller of the heat dissipation system;

[0069] Meanwhile, the optimal switching frequency in this optimal control parameter combination The signal is sent to the main control unit of the inverter to adjust the drive signal frequency of the power devices; optimal MPPT perturbation strategy. The data is then sent to the maximum power point tracking controller to update its core parameters, such as the perturbation step size. In this way, the system achieves comprehensive and coordinated control of power generation efficiency and heat dissipation strategy.

[0070] The function of the feedback correction module is to compensate for the deviation of the feedforward control, thereby improving the control accuracy of the system; this module monitors the actual junction temperature collected by the temperature sensor inside the inverter. The dynamic expected steady-state junction temperature output by the collaborative optimization decision module Deviation between And thus obtain the fan speed correction amount. ;

[0071] The instruction synthesis and transmission module will use the reference setpoint provided by the feedforward control execution module. Fan speed correction amount obtained from feedback correction module The synthesis yields a final target rotational speed command that integrates feedforward prediction and feedback correction, and has clear physical meaning. And send the command to the fan's underlying speed controller to drive the fan;

[0072] This invention integrates four modules—multimodal state prediction, collaborative optimization decision-making, feedforward control execution, and feedback correction—to form a feedforward-feedback composite control architecture. This architecture enables robust and precise collaborative control of the photovoltaic inverter's heat dissipation and operating point. The system can proactively predict upcoming heat load changes and collaboratively optimize the switching frequency, MPPT strategy, and heat dissipation strategy at the system level. This avoids the temperature overshoot and frequent start-stop problems caused by passive response in existing technologies. While ensuring high reliability of power devices, it maximizes the overall net energy gain of the inverter system throughout its entire lifecycle.

[0073] Example 2:

[0074] The modal state prediction module is also used for:

[0075] By using a system identification algorithm and offline collected historical operating data, the model coefficients of the discrete-time dynamic thermal model are fitted and calibrated.

[0076] The correction terms of the discrete-time dynamic thermal model are updated in real time using an online adaptive algorithm.

[0077] Building upon Example 1, this embodiment further elaborates on the parameter determination method for the model in the multimodal state prediction module. This module utilizes a system identification algorithm to determine the model coefficients of the discrete-time dynamic thermal model by fitting and calibrating historical operational data collected offline. System identification algorithms refer to a method of building mathematical models using input and output data. For example, regression analysis can be used to utilize known input quantities, such as light intensity, from offline historical datasets. Ambient temperature Output power Current junction temperature and output, such as future junction temperature Solve the model to obtain the optimal model coefficients;

[0078] In addition, this module also uses an online adaptive algorithm to correct the terms of the discrete-time dynamic thermal model. Real-time updates are performed; the online adaptive algorithm, such as the Least Mean Square (LMS) algorithm, aims to dynamically adjust the correction term based on the deviation between the actual measured value and the model prediction value during system operation, in order to compensate for the long-term drift of the model caused by changes in operating conditions, thereby ensuring the accuracy of the prediction; this combination of offline calibration and online adaptation acknowledges and compensates for the partial physical fidelity sacrificed by using a linearly simplified model to ensure real-time calculation, and compensates for nonlinear effects and model drift through real-time correction while ensuring computational efficiency;

[0079] By combining offline fitting calibration with online adaptive updating, this embodiment not only ensures the initial accuracy of the model, but also enables it to adapt to changes in the thermal characteristics of the photovoltaic system under different seasons, weather conditions, and long-term operation. This greatly improves the robustness and accuracy of junction temperature prediction, providing a more reliable data foundation for subsequent collaborative optimization decisions.

[0080] Example 3:

[0081] The cost function unifies power generation revenue, heat dissipation costs, and operational risks under a single optimization objective. The cost function specifically includes:

[0082] The product of conversion efficiency and input power is used to characterize the revenue generated from power generation.

[0083] The heat dissipation and power consumption term is used to characterize the cost of heat dissipation;

[0084] The temperature penalty function term is used to quantify the risk of equivalent power loss due to excessive junction temperature;

[0085] The temperature penalty function is used for:

[0086] The predicted junction temperature is compared with the preset safety threshold.

[0087] When the predicted junction temperature is below the safe threshold, the temperature penalty function is set to zero;

[0088] When the predicted junction temperature exceeds the safety threshold, the temperature penalty function is set to an exponential form.

[0089] This embodiment further discloses the composition details and specific implementation of the cost function adopted by the collaborative optimization decision module in the system of Embodiment 1; the cost function aims to unify power generation revenue, heat dissipation costs, and operational risks under a single optimization objective, and its expression is:

[0090]

[0091] in:

[0092] Cost function, in watts (W);

[0093] Switching frequency, measured in Hertz (Hz), is calculated by the collaborative optimization decision module;

[0094] MPPT perturbation strategy, dimensionless, represents control parameters such as perturbation step size and perturbation direction, which are calculated by the collaborative optimization decision module; the optimization here does not replace the real-time microsecond-level perturbation of the MPPT algorithm, but rather performs minute-level dynamic optimization of its core hyperparameters to balance tracking speed and steady-state efficiency over long-term scales.

[0095] Fan speed, measured in revolutions per minute (RPM), is calculated by the collaborative optimization decision module;

[0096] Weighting coefficient, dimensionless, whose value is preset based on economic and engineering factors such as electricity price and equipment life cycle cost;

[0097] Conversion efficiency is dimensionless and its value is related to the switching frequency and the MPPT strategy. For example, conversion efficiency can be modeled as a function of switching losses related to the switching frequency and steady-state tracking error related to the MPPT strategy. Its specific form can be obtained through offline calibration or theoretical modeling. For example, conversion efficiency can be modeled as a function including switching losses and conduction losses, where the switching losses are related to the switching frequency. Positive correlation, while MPPT strategy This affects the system's steady-state tracking error, and consequently, the total output power; a simplified model could be...

[0098]

[0099] in, Based on basic efficiency, This is the switching loss factor, expressed in seconds (s). Represented by MPPT strategy The dimensionless equivalent efficiency loss is determined; these coefficients can all be calibrated through offline experiments, for example, when the MPPT policy... This mainly refers to the power loss corresponding to the perturbation step size s. It can be modeled as a function related to steady-state oscillations and dynamic tracking velocity, the specific form of which is obtained through offline simulation or fitting of experimental data; for example, when the MPPT strategy... Mainly refers to the perturbation step size At that time, its corresponding dimensionless efficiency loss factor It can be modeled as a function related to steady-state oscillations and dynamic tracking speed;

[0100] It should be noted that this model is a simplified form; a more accurate conversion efficiency model would also include the current junction temperature. and output power In this embodiment, the impact of these factors will be primarily compensated by the feedback correction mechanism.

[0101] Input power, measured in watts (W), is obtained in real time through the power monitoring equipment at the input end of the photovoltaic array;

[0102] Heat dissipation power, measured in watts (W), is related to fan speed. For example, based on the aerodynamic principles of fans, heat dissipation power can usually be approximated as a function proportional to the cube of the fan speed, i.e. ,in This is the fan power consumption factor, which can be determined by consulting the fan's datasheet or by conducting actual power tests on the fan. The unit is watts per cubic revolution per minute (W / min). );

[0103] Temperature penalty function, in watts (W);

[0104] The cost function specifically includes the following three items:

[0105] The product of conversion efficiency and input power Used to characterize power generation revenue, its purpose is to maximize the system's output power by optimizing the switching frequency and MPPT strategy;

[0106] Heat dissipation power consumption Used to characterize heat dissipation costs, its purpose is to minimize the energy consumption of the heat dissipation system itself by optimizing fan speed;

[0107] Temperature penalty function term It is used to quantify the risk of equivalent power loss due to excessive junction temperature. Its purpose is to force the system to maintain the junction temperature within a safe range by introducing a penalty for high junction temperature into the optimization target, thereby ensuring the reliability and service life of power devices.

[0108] The temperature penalty function The specific rules are as follows: This function will predict the junction temperature. With preset safety threshold Compare; this safety threshold It is a fixed value preset based on the maximum safe operating temperature of the power device and a certain safety margin; when the junction temperature is predicted... Below this safety threshold At that time, the value of the temperature penalty function Setting it to zero means that within a safe temperature range, the system is not penalized by temperature factors; when predicting the junction temperature... Exceed At that time, the value of the temperature penalty function Set in exponential form:

[0109]

[0110] in:

[0111] The base penalty factor, measured in watts (W), is a preset constant with power dimensions. This value is set to quantify the basic risk cost when the temperature just begins to exceed the safety threshold. Its magnitude can be estimated based on economic factors such as the cost of replacing components and the power generation loss caused by derating operation. For example, it can be set as a multiple of the maximum power consumption of the heat dissipation system.

[0112] Characteristic temperature, in Kelvin (K), is set based on the device's reliability curve, such as the Arrhenius model, and is used to characterize the severity of penalty growth. This value can be obtained by fitting the device's lifetime data at different temperatures, for example, by determining the coefficient related to the device's lifetime decay rate through regression analysis.

[0113] This embodiment achieves coordinated optimization of multiple control dimensions of photovoltaic inverters by constructing a cost function that integrates power generation revenue, heat dissipation costs, and operational risks. In particular, the design of the temperature penalty function, by quantifying and penalizing the risk of over-temperature, ensures that the optimization objective not only pursues immediate power generation efficiency but also takes into account the long-term reliability of the equipment. This effectively avoids the risk of sacrificing equipment lifespan in pursuit of short-term high returns, thereby maximizing the comprehensive net electrical energy revenue of the photovoltaic system throughout its entire life cycle.

[0114] It should be noted that the cost function in this embodiment mainly focuses on the balance between electrical energy and thermal energy; in more complex implementations, the cost function can be further extended to include EMI filtering costs related to switching frequency, grid voltage fluctuation penalties related to MPPT disturbances, etc., thereby achieving more comprehensive system-level optimization.

[0115] Example 4:

[0116] The collaborative optimization decision-making module is also used for:

[0117] The cost function is solved using a numerical optimization algorithm.

[0118] This embodiment further discloses the specific method for the collaborative optimization decision module to solve the cost function in the system of Embodiment 1; this module uses a numerical optimization algorithm to solve the cost function. In order to obtain the optimal combination of control parameters The numerical optimization algorithm refers to algorithms such as genetic algorithms, particle swarm optimization, or the simpler gradient descent method; this algorithm searches iteratively within a preset control parameter space to find the algorithm that optimizes the cost function. The parameter combination that achieves the maximum value; for example, the search space can be set as: switching frequency. MPPT perturbation step size Fan speed These boundaries are determined by the specifications of the power devices and the physical capabilities of the heat dissipation system; the optimization process is repeated in each decision cycle to adapt to the constantly changing operating conditions.

[0119] By employing numerical optimization algorithms, this embodiment can efficiently and accurately find the optimal solution among multiple interrelated control variables, avoiding the local optimum problem caused by independent decision-making of each control link in traditional methods, and ensuring the optimization of the overall system performance.

[0120] Example 5:

[0121] The feedback correction module is specifically used for:

[0122] A proportional-integral (PI) controller is used;

[0123] Determine the deviation between the actual junction temperature and the desired steady-state junction temperature;

[0124] Based on the deviation, the fan speed correction amount is calculated;

[0125] Based on Embodiment 1, this embodiment further discloses the specific implementation of the feedback correction module; this module uses a proportional-integral (PI) controller, the purpose of which is to calculate a real-time correction amount based on the deviation between the actual and expected values; the controller determines the actual junction temperature. With the expected steady-state junction temperature Deviation between ,in Based on this deviation, the fan speed correction amount is calculated. The formula for calculating this correction amount is:

[0126]

[0127] in:

[0128] Fan speed correction, in revolutions per minute (RPM).

[0129] : Proportional coefficient, in units of RPM / K, obtained through experimental tuning of a system with deployed feedforward control;

[0130] : Integral coefficient, in units of RPM / (K·s), obtained through experimental tuning of a system with deployed feedforward control;

[0131] Temperature error, in Kelvin (K);

[0132] : Integral variable;

[0133] This embodiment introduces a PI controller, enabling the system to compensate for deviations in feedforward control caused by inaccurate models or unknown external disturbances in real time. The proportional term can quickly respond to the current deviation, while the integral term can eliminate long-term steady-state errors, thereby ensuring that the actual junction temperature accurately and stably tracks the dynamically desired steady-state junction temperature, significantly improving the robustness and control accuracy of the system.

[0134] Example 6:

[0135] The calculation process for the final target speed command is as follows:

[0136] The optimal fan speed and the fan speed correction amount are combined to obtain the final target speed command;

[0137] The final target speed command is used to drive the fan;

[0138] Based on Example 1, this embodiment further elaborates on the calculation process of the final target speed command. The control command ultimately applied to the cooling system actuators, such as the fan driver, is a final target speed command that integrates feedforward prediction and feedback correction and has clear physical meaning. This instruction is based on the optimal fan speed provided by the feedforward control execution module. Fan speed correction amount calculated by the feedback correction module It is obtained through synthesis, and its synthesis formula is as follows:

[0139]

[0140] Final target speed command The signal is sent to the fan's underlying speed controller, which is responsible for generating a specific PWM signal or drive voltage to ensure that the actual fan speed accurately tracks this dynamic command, thereby driving the fan to work according to the system's optimized command.

[0141] This embodiment constructs a logically clear and easily implemented composite control structure by using the optimal speed predicted by feedforward as the basis for active adjustment and by using feedback correction to compensate for deviations in real time. This structure enables the system to maintain its forward-looking and proactive advantages while using the accuracy of feedback to eliminate the influence of model errors and external disturbances, thus ensuring the accuracy and stability of control.

[0142] This system discloses a photovoltaic inverter heat dissipation regulation and control system. Its technical advantage lies in the construction of a feedforward-feedback composite control architecture, which is fundamentally different from the passive response heat dissipation control method in the existing technology. By integrating four core modules, namely multimodal state prediction, collaborative optimization decision-making, feedforward control execution and feedback correction, the system achieves robust and precise collaborative control of photovoltaic inverter heat dissipation and operating point.

[0143] The multimodal state prediction module uses real-time environmental data and historical operating state data to make forward-looking predictions of future junction temperatures based on a discrete-time dynamic thermal model. The prediction model is initially calibrated through an offline system identification algorithm and updated in real time with an online adaptive algorithm, which effectively compensates for the long-term drift of the model caused by changes in operating conditions, greatly improves the robustness and accuracy of junction temperature prediction, and provides a reliable data foundation for subsequent optimization decisions.

[0144] The collaborative optimization decision-making module unifies power generation revenue, heat dissipation costs, and operational risks under a single optimization objective, avoiding the local optima problem caused by independent decision-making in traditional control. This module obtains the optimal combination of control parameters by solving the cost function and collaboratively analyzing the switching frequency, MPPT disturbance strategy, and fan speed. The cost function specifically includes a temperature penalty function term. When the predicted junction temperature exceeds the preset safety threshold, this penalty function will grow exponentially. This allows the system to effectively quantify and avoid the risk of power device losses caused by excessive junction temperature while pursuing real-time power generation efficiency, thus taking into account the long-term reliability of the equipment and maximizing the comprehensive net power revenue of the inverter system throughout its entire life cycle.

[0145] The feedforward control execution module and the feedback correction module work together to form a control closed loop that combines foresight and precision. The feedforward control module directly uses the optimized fan speed as the reference setpoint to achieve proactive adjustment based on prediction, effectively avoiding the temperature overshoot and frequent start-stop problems caused by passive response in existing technologies. The feedback correction module uses a proportional-integral (PI) controller to monitor the deviation between the actual junction temperature and the desired steady-state junction temperature in real time and calculates the fan speed correction amount. Finally, the system combines the optimal fan speed and the correction amount to generate a final target speed command with clear physical meaning. This composite control structure ensures that the actual junction temperature can accurately and stably track the dynamically desired steady-state junction temperature, significantly enhancing the control accuracy and robustness of the system and effectively eliminating the influence of model errors and external unknown disturbances.

[0146] In addition, this control system has good robustness; in response to abnormal situations such as the sensor measuring a value of 0 or saturation, the system has built-in input data validity verification logic. When the sensor measures a value of 0 or saturation, it can switch to a conservative operation mode. Specifically, when the temperature sensor fails, the system will find the worst-case fan speed from the pre-stored safe operation curve based on the current light intensity and output power, and run it at full speed, while issuing an alarm to the monitoring system.

[0147] When the light or power sensor data is abnormal, the system will lock at a fixed, low switching frequency and use a large MPPT perturbation step size to ensure tracking stability. At the same time, the fan speed will be set to a higher value, sacrificing some energy efficiency for absolute safety. It can switch to a conservative operating mode based on historical data and safety margin. When faced with drastic changes in operating conditions such as sudden changes in light intensity, the feedforward control module can respond quickly, while the PI parameters of the feedback correction module are carefully tuned to effectively suppress overshoot and oscillation, ensuring a smooth transition of the system. This ensures safe and stable operation under various complex and non-ideal operating conditions.

[0148] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention; any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0149] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A heat dissipation regulation and control system for a photovoltaic inverter, characterized in that, include: The multimodal state prediction module is used to receive real-time environmental data and inverter historical operating state data. The environmental data includes light intensity and ambient temperature, and the inverter historical operating state data includes current junction temperature and output power. Based on a preset discrete-time dynamic thermal model, it predicts the junction temperature for one or more future time steps. The collaborative optimization decision module receives the predicted junction temperature and, based on the cost function aimed at maximizing the overall net energy efficiency of the system, solves and analyzes the switching frequency, MPPT disturbance strategy, and fan speed to obtain the optimal combination of control parameters and the desired steady-state junction temperature. The feedforward control execution module is used to receive the optimal control parameter combination, set the optimal fan speed in the optimal control parameter combination as the reference set value of the cooling system speed controller, send the optimal switching frequency to the main control unit of the inverter for adjusting the drive signal frequency of the power devices, and send the optimal MPPT disturbance strategy to the maximum power point tracking controller for updating its core parameters. The feedback correction module is used to monitor the deviation between the actual junction temperature and the desired steady-state junction temperature in order to obtain the fan speed correction amount; The instruction synthesis and transmission module is used to synthesize the reference setpoint and the fan speed correction value to obtain the final target speed instruction, and send the final target speed instruction to the fan's underlying speed controller; The cost function unifies power generation revenue, heat dissipation costs, and operational risks under a single optimization objective. The cost function specifically includes: The product of conversion efficiency and input power is used to characterize the revenue generated from power generation. The heat dissipation and power consumption term is used to characterize the cost of heat dissipation; The temperature penalty function term is used to quantify the risk of equivalent power loss due to excessive junction temperature; The temperature penalty function is used for: The predicted junction temperature is compared with the preset safety threshold. When the predicted junction temperature is below the safe threshold, the temperature penalty function is set to zero; When the predicted junction temperature exceeds the safety threshold, the temperature penalty function is set to an exponential form. The expression for the cost function is: in: Cost function; Switching frequency; MPPT perturbation strategy, dimensionless; Fan speed; Weighting coefficient; Conversion efficiency is a dimensionless quantity whose value is related to the switching frequency and the MPPT strategy. Input power; Heat dissipation and power consumption; : Temperature penalty function; the temperature penalty function The expression is: in: Base penalty coefficient; Characteristic temperature; Predicting the junction temperature; : Preset safety threshold.

2. The photovoltaic inverter heat dissipation regulation and control system according to claim 1, characterized in that, The multimodal state prediction module is also used for: By using a system identification algorithm and offline collected historical operating data, the model coefficients of the discrete-time dynamic thermal model are fitted and calibrated. The correction terms of the discrete-time dynamic thermal model are updated in real time using an online adaptive algorithm.

3. The photovoltaic inverter heat dissipation regulation and control system according to claim 1, characterized in that, The collaborative optimization decision-making module is also used for: The cost function is solved using a numerical optimization algorithm.

4. The photovoltaic inverter heat dissipation regulation and control system according to claim 1, characterized in that, The feedback correction module is specifically used for: A proportional-integral controller is used; Determine the deviation between the actual junction temperature and the desired steady-state junction temperature; Based on the deviation, the fan speed correction amount is calculated.

5. The photovoltaic inverter heat dissipation regulation and control system according to claim 1, characterized in that, The calculation process for the final target speed command is as follows: The optimal fan speed and the fan speed correction amount are combined to obtain the final target speed command; The final target speed command is used to drive the fan.

Images