A prefabricated cabin type substation active and passive collaborative temperature control method

Abstract

The application discloses a prefabricated cabin type substation active and passive collaborative temperature control method and relates to the technical field of intelligent environment control of power facilities. The method comprises the following steps: acquiring real-time operation data of the prefabricated cabin; constructing a thermal network model containing thermal physical parameters of a phase change material, and estimating a current phase change saturation degree of the phase change material and a predicted temperature in the cabin in real time; monitoring a deviation cumulative value of the predicted temperature and the actual temperature, and performing model parameter identification and correcting thermal resistance and heat capacity parameters in the model when the deviation cumulative value exceeds a mismatch threshold; and finally, based on the corrected model and the current phase change saturation degree, a rolling horizon optimization algorithm is used to generate and execute a collaborative control strategy. The application is used to solve the problems that existing temperature control technologies cannot perceive the energy storage state of the phase change material, the model mismatch is caused by the physical parameter drift caused by long-term operation, and the collaborative control precision and energy efficiency are reduced.

Description

Technical Field

[0001] This invention relates to the field of intelligent environmental control technology for power facilities, and more specifically, to a method for active and passive coordinated temperature control in prefabricated substations. Background Technology

[0002] Prefabricated substation modules are widely used due to their modular construction advantages. However, with the increasing integration of equipment, the modules exhibit high heat flux density and drastic load fluctuations. Continuous heat generation from the equipment can easily trigger localized heat island effects. Improper temperature control can lead to accelerated aging of electronic components or even system failure. This places stringent requirements on the stability and uniformity of environmental temperature control.

[0003] Existing temperature control technologies mainly include active mechanical refrigeration and passive phase change energy storage. Traditional air conditioners mostly use PID or simple temperature threshold logic to control cabin temperature; phase change materials (PCMs) utilize the latent heat of solid-liquid conversion to absorb heat, playing a certain role in peak shaving and valley filling. Currently, some technologies are attempting to combine the two in order to reduce system energy consumption.

[0004] However, existing technologies have significant drawbacks: First, traditional air conditioning control relies solely on air temperature feedback, resulting in strong lag and frequent equipment start-ups and shutdowns, as well as temperature oscillations. Second, in coordinated systems, the controller cannot perceive the "phase change saturation" of the PCM in real time, causing the PCM to often saturate before peak load and lose its thermal buffering capacity. Most critically, existing model predictive control schemes typically assume constant physical parameters, neglecting thermal resistance and thermal capacity drift caused by dust accumulation, seal aging, and material degradation during long-term operation. This "model mismatch" phenomenon can cause the control strategy to gradually deviate from the optimal solution or even fail completely, and existing technologies lack effective online parameter identification and adaptive correction mechanisms. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an active and passive coordinated temperature control method for prefabricated substations. This method estimates the phase change saturation in real time by constructing a thermal network model containing phase change characteristics, and identifies and corrects model parameters online based on temperature prediction deviations. This addresses the problems of the prior art being unable to sense the energy storage status of phase change materials and the decrease in coordinated control accuracy due to time-varying failure of model parameters.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for coordinated active and passive temperature control in a prefabricated substation includes the following steps: acquiring real-time operating data of the prefabricated substation compartment; constructing a prefabricated compartment thermal network model containing the thermophysical properties of the phase change material, using the real-time operating data as input, and calculating the current phase change saturation of the phase change material and the predicted temperature inside the compartment using a state estimation algorithm; calculating the deviation between the predicted temperature and the actual monitored temperature, and monitoring the cumulative value of the deviation within a set time window; when the cumulative value exceeds a preset mismatch threshold, performing model parameter identification, and correcting the thermal resistance and thermal capacity parameters in the prefabricated compartment thermal network model based on the identification results; based on the corrected prefabricated compartment thermal network model and the current phase change saturation, generating a coordinated control strategy using a rolling time-domain optimization algorithm, the coordinated control strategy including the switching sequence and power setting of the active temperature control system in the future time domain; and executing the coordinated control strategy to achieve coordinated operation of the active temperature control system and the passive temperature regulation of the phase change material.

[0008] In a preferred embodiment, the execution model parameter identification includes the following steps: controlling the active temperature control system to output a preset amplitude of cooling or heating power, and stopping the output after a preset test duration, thereby creating an active thermal disturbance inside the cabin; collecting cabin temperature response data during the active thermal disturbance and within a preset decay time after the disturbance ends, and constructing a step response curve of the system; based on the step response curve, using a system identification algorithm to solve the differential equation of the prefabricated cabin thermal network model in reverse, obtaining the equivalent heat transfer resistance and the effective heat storage capacity of the phase change material under the current operating conditions; and updating the prefabricated cabin thermal network model using the obtained equivalent heat transfer resistance and effective heat storage capacity.

[0009] In a preferred embodiment, the cooperative control strategy generated using a rolling time-domain optimization algorithm is obtained by establishing a multi-objective optimization function and solving for its minimum value. Establishing the multi-objective optimization function includes the following steps: establishing a weighted summation function comprising an energy consumption cost term, a temperature deviation penalty term, and an equipment loss term; wherein the energy consumption cost term is positively correlated with the operating power and operating time of the active temperature control system; the temperature deviation penalty term is positively correlated with the degree to which the predicted cabin temperature deviates from the set target temperature; and the equipment loss term is positively correlated with the number of start-stop switching times or the power adjustment amplitude of the active temperature control system in the prediction time domain; and, under the constraint that the cabin temperature is within a preset safe range and the phase change saturation of the phase change material is within the physical limit range, solving for a control sequence that minimizes the weighted summation function.

[0010] In a preferred embodiment, before generating the cooperative control strategy using the rolling time-domain optimization algorithm based on the modified prefabricated cabin thermal network model and the current phase change saturation, a load prediction step is further included: collecting historical load data and historical meteorological data of the cabin electrical equipment; using a time series prediction model to predict the heat generation curve of the cabin equipment and the ambient temperature curve of the cabin outside the cabin within a future set time domain; wherein, the rolling time-domain optimization algorithm uses the heat generation curve of the cabin equipment and the ambient temperature curve of the cabin outside the cabin as disturbance variables to calculate the optimal control sequence in the future time domain.

[0011] In a preferred embodiment, the implementation of the collaborative control strategy includes hierarchical control: when the estimated current phase change saturation is in the unsaturated range and the predicted cabin temperature does not exceed the safety threshold in the first time domain in the future, an instruction is generated to control the active temperature control system to remain off or in low-power ventilation mode, utilizing the latent heat of the phase change material for passive heat absorption; when the estimated current phase change saturation reaches the critical saturation threshold, or the predicted cabin temperature will exceed the safety threshold in the first time domain in the future, an instruction is generated to control the active temperature control system to start in advance and supplement cooling capacity.

[0012] In a preferred embodiment, the collaborative control strategy further includes nighttime active recovery logic:

[0013] Determine whether the current time is during the off-peak electricity price period at night, and combine this with weather forecasts to determine whether the next day will be a high-temperature operating condition. If both the off-peak electricity price period at night and the high-temperature operating condition the next day are met, then under the premise of meeting the minimum temperature constraint of the prefabricated cabin, forcefully control the active temperature control system to operate until the estimated current phase change saturation is reduced to zero or the preset recovery bottom value, so as to complete the cold energy storage of the phase change material.

[0014] In a preferred embodiment, the differential equations of the prefabricated cabin thermal network model are solved in reverse using a system identification algorithm. Specifically, the recursive least squares method or particle swarm optimization algorithm is adopted, with the goal of minimizing the root mean square error between the model output temperature and the actual collected step response curve, and the optimal parameter combination is iteratively searched.

[0015] In a preferred embodiment, the calculation logic for the current phase change saturation of the phase change material is as follows: the instantaneous heat flow value flowing through the phase change material layer is calculated based on the prefabricated cabin thermal network model; the instantaneous heat flow value is integrated on the time axis to obtain the cumulative heat absorbed or released by the phase change material; the cumulative heat is compared with the total latent heat capacity of the phase change material to obtain a normalized phase change saturation value.

[0016] In a preferred embodiment, the real-time operating data includes external environmental parameters, multi-point temperatures inside the cabin, real-time power of electrical equipment inside the cabin, and the current operating status of the active temperature control system.

[0017] This invention provides a prefabricated substation active and passive coordinated temperature control system, comprising: a data acquisition module for acquiring real-time operating data of the prefabricated substation compartment; a thermal network modeling and state estimation module for constructing a prefabricated compartment thermal network model including the thermal property parameters of the phase change material, and using the real-time operating data as input, calculating the current phase change saturation of the phase change material and the predicted temperature inside the compartment in real time through a state estimation algorithm; a deviation accumulation and online identification and correction module for calculating the deviation between the predicted temperature and the actual monitored temperature, and monitoring the cumulative value of the deviation within a set time window; when the cumulative value exceeds a preset mismatch threshold, performing model parameter identification, and correcting the thermal resistance and thermal capacity parameters in the prefabricated compartment thermal network model online based on the identification results; a rolling optimization decision module for generating a coordinated control strategy based on the corrected prefabricated compartment thermal network model and the current phase change saturation using a rolling time-domain optimization algorithm; and a coordinated execution module for executing the coordinated control strategy to achieve coordinated operation of the active temperature control system and the passive temperature regulation of the phase change material.

[0018] The technical effects and advantages of the active and passive coordinated temperature control method for prefabricated substations of this invention are as follows:

[0019] This invention achieves digital perception of passive energy storage status by constructing a thermal network model incorporating the characteristics of phase change materials and estimating the phase change saturation, which cannot be directly measured, in real time. Combined with a prediction deviation monitoring and online parameter identification mechanism, it can automatically correct thermal resistance and heat capacity parameters when the model is mismatched, overcoming the control failure problem caused by physical parameter drift due to long-term operation. Finally, based on the corrected high-precision model, a rolling time-domain optimization algorithm is used to achieve precise coordination between the active temperature control system and the passive temperature regulation of the phase change material, ensuring the real-time performance, accuracy, and stability of the temperature control strategy. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the active and passive coordinated temperature control method for prefabricated substations provided in an embodiment of the present invention;

[0021] Figure 2 The following is a simulation result curve of state estimation under typical summer operating conditions provided in the embodiments of the present invention;

[0022] Figure 3 This is a system identification step response diagram provided in an embodiment of the present invention;

[0023] Figure 4 This is a comparison chart of collaborative temperature control and traditional control provided in an embodiment of the present invention;

[0024] Figure 5 This is a block diagram of the active and passive coordinated temperature control system for a prefabricated substation provided in an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1, Figure 1 The present invention provides a method for active and passive coordinated temperature control in a prefabricated substation, comprising the following steps:

[0027] S1, acquires real-time operating data of the prefabricated substation compartment.

[0028] In this embodiment, the real-time operating data serves as the physical basis for constructing the subsequent thermal network model and performing state estimation. Specifically, it covers four dimensions: external environmental parameters, multi-point temperature inside the cabin, real-time power of electrical equipment inside the cabin, and the current operating status of the active temperature control system.

[0029] Specifically, to achieve comprehensive perception of the prefabricated cabin's thermal environment, data acquisition relies on a high-precision, multi-dimensional sensor array and communication interfaces. For acquiring temperatures at multiple points within the cabin, temperature sensors, such as high-precision PT100 platinum resistance thermometers or industrial-grade thermocouples, are deployed at various key thermal locations within the prefabricated cabin. These sensors are not randomly distributed but are strategically placed at the return air vents of the air conditioning units to monitor the average air temperature within the cabin, in cold aisle areas with dense electrical equipment to capture localized hotspot temperatures, and specifically on the surface and within the internal layers of the phase change material (PCM) energy storage modules to directly respond to the thermal state of the PCM. For the real-time power of the electrical equipment within the cabin, industrial fieldbus protocols such as Modbus-RTU are used, and the power register values ​​of smart meters or integrated protection devices within the prefabricated cabin are directly read via an RS485 physical interface. This allows for the real-time acquisition of the active power of major heat sources, including transformers, inverters, and switchgear. This data will be directly used to calculate the heat generation of the internal heat sources.

[0030] Furthermore, to determine the current operating status of the active temperature control system, the real-time operating frequency of the air conditioning compressor, the percentage speed of the condenser and evaporator fans, and the opening feedback of the electronic expansion valve are read through the communication interface (RS485 or CAN bus) of the air conditioning unit. This allows for the determination of the current cooling capacity and energy consumption level of the active temperature control system. For external environmental parameters, real-time data on outdoor dry-bulb temperature, relative humidity, and solar radiation intensity are obtained through a micro-weather station deployed on the top of the prefabricated cabin or by accessing local meteorological data services via a network interface. All the raw data collected above are typically preprocessed using low-pass filters or median filters before being input into the control model to remove noise signals caused by electromagnetic interference from the substation, ensuring the data's authenticity and reliability. Through this comprehensive state perception approach, this step achieves a complete digital mapping from external environmental disturbances to internal equipment heat sources and the state of the temperature control actuators, providing a solid data foundation for subsequently building a high-precision thermal network model and implementing precise collaborative control.

[0031] S2, construct a thermal network model of the prefabricated cabin, and calculate the phase change saturation and predict the temperature in real time.

[0032] After obtaining the real-time operating data, step S2 is executed to establish a mapping relationship between the physical entity and the digital space. First, a prefabricated cabin thermal network model containing the thermophysical properties of the phase change material is constructed, and the real-time operating data obtained in step S1 is used as input. Finally, a state estimation algorithm is used to calculate the current phase change saturation of the phase change material, which is physically difficult to measure directly, in real time, while predicting the future temperature trend inside the cabin.

[0033] Specifically, the constructed prefabricated cabin thermal network model typically employs the lumped parameter method, simplifying the prefabricated cabin into an RC equivalent thermal circuit composed of several thermal resistances and capacitors. This model not only considers the thermal response of the cabin enclosure structure to the external environment but also introduces a branch describing the heat absorption and release characteristics of the phase change material. The air thermal equilibrium process within the prefabricated cabin is abstracted and described by the following set of first-order linear differential equations:

[0034] (1)

[0035] (2)

[0036] in, Represents the equivalent total heat capacity of air and equipment inside the prefabricated cabin (unit: joules per Kelvin). represent Average air temperature inside the cabin at any given time (unit: °C); Represents the real-time heat generation power (unit: watts) of the electrical equipment inside the cabin. This represents the cooling power (in watts) provided by the active temperature control system. Represents the ambient temperature outside the cabin; Equivalent thermal resistance of the cabin enclosure structure (unit: Kelvin per watt); The instantaneous heat flux (in watts) represents the heat exchange between the phase change material layer and the cabin air. The core temperature of the phase change material is obtained by directly measuring it through a sensor embedded inside the material. The thermal resistance for convective heat transfer between the phase change material layer and the air.

[0037] It should be noted that in actual high-precision control scenarios, the equivalent total heat capacity inside the prefabricated cabin is... It is not a constant. Because the density of the air inside the cabin changes with temperature, and more importantly, with the cabin temperature... As the temperature rises, the exhaust fans built into electrical equipment (such as inverters and transformers) automatically start or accelerate, causing a nonlinear change in the convective heat transfer coefficient between the equipment's metal casing and internal components and the cabin air. This alters the 'effective heat capacity' involved in the thermal dynamic equilibrium. Therefore, this embodiment will... Corrected to a nonlinear function that varies with cabin temperature:

[0038] (3)

[0039] in, The reference heat capacity at the reference temperature. For reference temperature (e.g., 25℃). The nonlinear coupling coefficient for heat capacity and temperature is typically 0.002 to 0.005. Substituting equation (3) into equation (2), we can see that the state variable... The nonlinear coupling with itself transforms the system state equation into a nonlinear form, rendering conventional linear Kalman filtering inapplicable.

[0040] To utilize computers for digital processing and prediction, the above formula (1) needs to be discretized. In this embodiment, the forward Euler method is used to transform the continuous-time model into a discrete-state-space model. Assume the sampling time interval is... ,but The formula for predicting the cabin temperature at a given time is as follows:

[0041] (4)

[0042] Given the nonlinear characteristics of the modified model, this step employs a strategy that combines prediction and correction:

[0043] First, the Extended Kalman Filter (EKF) algorithm is used to process the current time step. The EKF algorithm performs an optimal estimation of the state. It achieves local linearization by performing a Taylor series expansion at the current state estimation point (calculating the Jacobian matrix). The EKF algorithm uses the actual sensor temperature as the observation value, calculates the Kalman gain matrix, and corrects the theoretical temperature calculated by the above model, thereby eliminating sensor noise and model errors, and obtaining the closest approximation to the actual cabin temperature and heat flow state at the current moment. Figure 2 The figure shown is a simulation result curve of state estimation under typical summer conditions (24-hour cycle) in this embodiment.

[0044] Figure 2 It includes a dual Y-axis coordinate system. The left Y-axis represents temperature (°C), showing a high degree of overlap between the "sensor-measured average temperature curve" (solid line) and the "EKF-filtered estimated temperature curve" (dashed line), proving the convergence of the observer; the right Y-axis represents normalized phase transition saturation (0~1), where the "phase transition saturation curve" (dotted line) clearly shows the endothermic liquefaction process of PCM during the day (value increases from 0.2 to 0.9) and the exothermic solidification process at night (value decreases from 0.9 to 0.2), realizing the visual monitoring of unmeasurable physical quantities.

[0045] Furthermore, this embodiment strictly follows the principle of energy conservation in calculating the current phase change saturation of the phase change material. Since the energy storage state of the phase change material cannot be directly characterized by a single temperature point (because the temperature remains constant while the energy changes during the phase change process), an estimation logic based on heat flux integration is adopted. First, the instantaneous heat flux value flowing through the phase change material layer is calculated based on the aforementioned thermal network model. Then, this instantaneous heat flux value is integrated over the time axis to obtain the cumulative heat absorbed or released by the phase change material during the current phase change cycle. Finally, this cumulative heat is compared with the total latent heat capacity of the phase change material. The phase change saturation (… The specific calculation formula is as follows:

[0046] (5)

[0047] in, represent Normalized phase transition saturation of phase change material at any given time (value range 0~1, 0 represents completely solid state, 1 represents completely liquid state). The total latent heat capacity (unit: joules) of the phase change material laid in the prefabricated compartment is an inherent physical property of the material. This represents the starting moment of this phase transition process; and Representing the integral variables respectively The current cabin air temperature and the core temperature of the phase change material; This represents the convective heat transfer resistance between the phase change material layer and the air.

[0048] It should be noted that, in order to prevent integral drift, saturation correction needs to be set: when detected... If the temperature remains above the phase transition termination temperature (e.g., 28°C) for more than the set time, a forced reset will be performed. ;when Forced reset when the temperature remains below the phase transition initiation temperature (e.g., 18°C). .

[0049] This step, through the establishment of a refined mathematical model and integral algorithm, successfully transforms the "invisible and intangible" latent heat storage state of phase change materials into a visualized digital index. This solves the problem of control lag or failure caused by the inability to accurately know the remaining capacity of the material in traditional passive temperature control technology, and provides a quantitative basis for subsequent active and passive collaborative decision-making.

[0050] S3 monitors deviations and actively identifies and corrects model parameters when parameters fail.

[0051] During the continuous operation of the prefabricated cabin, in order to ensure that the thermal network model can accurately reflect the current physical characteristics of the cabin, step S3 is executed to monitor the model accuracy and make corrections when necessary.

[0052] Specifically, definition Temperature prediction deviation at time ,in To predict temperature for the model, The measured temperature is from the sensor. A sliding integration window is used to evaluate long-term cumulative error, for example, by setting the time window length. The time frame is 24 hours. When the calculated cumulative deviation value... When the preset mismatch threshold is exceeded (for example, when the 24-hour moving average error exceeds 1.5℃), it is determined that the current prefabricated cabin thermal network model parameters no longer match the actual physical conditions (such as phase change material aging or filter dust accumulation causing changes in thermal resistance), thereby triggering the active identification and correction process of model parameters.

[0053] Once the aforementioned active identification and correction process is triggered, the current conventional temperature control strategy based on Model Predictive Control (MPC) is first suspended, and the system switches to a "step excitation mode." In this mode, the goal is no longer to maintain temperature stability, but rather to acquire dynamic characteristic data of the system. Specifically, the active temperature control system (such as an air conditioning unit) is controlled to output a preset amplitude of cooling or heating power; for example, the compressor is forced to operate at 100% rated power for a preset test duration (e.g., 15 minutes), after which the output is immediately stopped and the fan is shut down. This process creates a standardized active thermal disturbance signal within the cabin. Subsequently, cabin temperature response data is collected during the active thermal disturbance and for a preset decay time (e.g., 30 minutes) after the disturbance ends, thereby constructing a system step response curve including both rising and falling edges. Figure 3 As shown, the dynamic response characteristics of the prefabricated cabin thermal system to a step excitation signal are demonstrated during a typical active identification process.

[0054] Figure 3 The horizontal axis represents the time axis (time / minute), and the vertical axis represents the change in cabin temperature. / ℃). The figure shows at After the active temperature control system outputs 100% cooling power, the cabin temperature exhibits an exponential decay trajectory (measured scatter plot data). The superimposed smooth curve in the figure is the theoretical response curve obtained by fitting using the recursive least squares (RLS) method. By comparing the fitting residuals (Residual Error) between the measured scatter plots and the theoretical curve, the identification algorithm converges to obtain the corrected equivalent thermal resistance value. and the effective heat capacity of phase change materials This reduced the root mean square error (RMSE) of the model from 1.5℃ before the correction to less than 0.2℃.

[0055] It should be noted that the triggering of the step excitation mode requires the fulfillment of a safety prerequisite: the current cabin temperature is within a safe buffer zone (e.g., ...). Furthermore, the current time is not a period of high grid load, in order to avoid affecting the safe operation of the equipment during the testing process.

[0056] Based on the collected step response curves, the differential equations of the prefabricated cabin thermal network model are solved in reverse using a system identification algorithm. Essentially, this is the process of solving the inverse problem of the heat conduction equation.

[0057] Specifically, this embodiment transforms the reverse solution process into a parameter optimization problem. To accurately obtain the physical parameters under the current operating conditions, this embodiment employs Recursive Least Squares (RLS) or Particle Swarm Optimization (PSO) as parameter optimization tools. Its core logic is to construct an objective function with model parameters as variables, aiming to minimize the root mean square error between the model output temperature and the actually acquired step response curve. The objective function... The expression is as follows:

[0058] (6)

[0059] in, This represents the parameter vector to be identified, containing the current equivalent thermal resistance. and the effective heat storage capacity of phase change materials ; The total number of sampling points representing the step response curve; Representing the Measured temperature at each sampling point; Representative will use parameters The first result calculated after substituting into the heat network model The theoretical temperature at each moment. Through iterative search, the theoretical temperature at each moment is obtained. To reach the minimum value, thus calculating the current optimal value. and .

[0060] Finally, using the equivalent heat transfer resistance and effective heat storage capacity identified above, the original parameters in the prefabricated cabin thermal network model are replaced, completing the online update of the model. This step, by introducing this unique active thermal excitation technology, effectively solves the model inaccuracy problem caused by the physical decay of phase change materials (such as the decrease in latent heat value due to phase separation) and changes in environmental factors (such as the increase in thermal resistance due to ash accumulation), ensuring that subsequent collaborative control strategies are always based on the high-precision physical model generation.

[0061] S4, based on the modified model and load forecast, uses a rolling time-domain optimization algorithm to generate a collaborative control strategy.

[0062] After completing the online correction of the model parameters, step S4 addresses the core issue of "how to optimally operate the temperature control equipment over a future period." To achieve forward-looking control, a load forecasting step is first performed before generating a specific control strategy. Specifically, historical load data of the cabin's electrical equipment (such as power curves over the past 24 hours) and local historical meteorological data need to be collected. Using a time-series forecasting model, such as a Long Short-Term Memory (LSTM) network or the XGBoost algorithm, this data is trained to predict the heat generation curves of the cabin equipment and the ambient temperature curves of the external environment within a future set time domain (such as the next 4 hours). These two sets of predicted curves are not merely static references but are used as key "disturbance variables" input into the rolling time-domain optimization (MPC) algorithm to assess in advance the impact of the external environment and internal load on the cabin temperature.

[0063] It should be noted that this embodiment uses LSTM to predict the heat generation of the equipment inside the cabin and the ambient temperature outside the cabin. The specific steps are as follows:

[0064] 1) Input Data Construction: Historical power data of the cabin's electrical equipment and historical ambient temperature data from the local weather station were collected. A sliding window technique was used to construct training samples, with the historical observation window length set to H=24 (the past 24 hours) and the prediction step size to F=4 (the next 4 hours). The dimensions of the input feature tensor were set to (Batch_Size, H, 2), where Batch_Size is the batch size and 2 represents two feature dimensions. Before being input into the network, the data was mapped to the [0,1] interval using the Min-Max normalization method.

[0065] 2) Model network architecture: The prediction model consists of: an input layer; two stacked LSTM layers, each containing 64 hidden units, with a dropout rate of 0.2 between layers to prevent overfitting; and a fully connected layer with an output dimension of 4, corresponding to the predicted values ​​at the next 4 time points.

[0066] 3) Training configuration: Mean squared error (MSE) is used as the loss function, and the Adam optimizer is used for parameter iteration. The initial learning rate is set to 0.001.

[0067] Subsequently, based on the modified prefabricated cabin thermal network model in step S3, the current phase change saturation estimated in step S2, and the predicted disturbance variables, a cooperative control strategy is generated using a rolling time-domain optimization algorithm. In this embodiment, the generation process of this strategy is essentially a process of solving for the minimum value of a multi-objective optimization function. This multi-objective optimization function is not a single-dimensional consideration, but rather a weighted summation function that includes an energy consumption cost term, a temperature deviation penalty term, and an equipment loss term. Among them, the energy consumption cost term aims to reduce operating costs, and its value is positively correlated with the operating power and operating time of the active temperature control system; the temperature deviation penalty term aims to ensure environmental indicators, and its value is positively correlated with the degree to which the predicted cabin temperature deviates from the set target temperature; the equipment loss term aims to extend the life of the actuators, and its value is positively correlated with the number of start-stop switching times or the power adjustment amplitude (i.e., the intensity of the control action) of the active temperature control system in the prediction time domain.

[0068] In mathematical expression, the multi-objective optimization function The constraints are constructed as follows:

[0069] (7)

[0070] (8)

[0071] in, This represents the total cost of optimizing the objective function; Represents the number of steps in the prediction time domain; Represents the discrete time step at the current moment; This represents the control sequence to be solved; Represents active temperature control system Electric power at any given moment; The sampling time interval; The cabin temperature predicted by the model at the next moment; To achieve the optimal target temperature; This represents the amount of change in control action (used to suppress frequent starts and stops). Let k be the control variable of the active temperature control system at time k; These represent the weighting coefficients for energy consumption, comfort, and equipment wear, respectively, and are adjusted according to the order of magnitude of each variable to achieve dimensional balance. and The prefabricated cabin's permissible temperature safety range; To constrain the saturation of phase change materials and ensure that they operate within physical limits; and This represents the output capability boundary of the active temperature control system.

[0072] In order to achieve the best balance between energy saving, comfort and equipment lifespan, this embodiment sets the weight coefficients and constraint boundaries of each item in the optimization function. The specific parameter settings are shown in Table 1.

[0073] Table 1

[0074]

[0075] By solving this function under the constraints of the aforementioned temperature safety range and phase transition saturation limit, a set of conditions for maximizing the total cost in the future time domain can be obtained. The minimum optimal control sequence. This sequence includes the on / off state and power setpoint of the active temperature control system at each future time step (e.g., power on to 50% at minute 10, power off at minute 20). The system executes only the first instruction in this sequence and repeats the prediction and optimization process at the next time step. This step, through this MPC-based feedforward control mechanism, effectively transforms future uncertainties into optimal decisions in the present, achieving a globally optimal balance between energy saving, stable temperature control, and extended equipment lifespan.

[0076] S5 executes a collaborative control strategy to achieve coordinated operation of the active temperature control system and the passive temperature regulation of the phase change material.

[0077] After generating the optimal collaborative control strategy that includes future time-domain switching sequences and power settings, the strategy is finally executed in step S5 to achieve seamless collaboration between the active temperature control system and the passive temperature regulation of the phase change material. This step is based on hierarchical control logic to issue instructions, aiming to dynamically adjust the control weights according to the real-time thermal state of the phase change material.

[0078] Specifically, the hierarchical control logic is mainly divided into passive priority mode and active compensation mode. When the current phase transition saturation estimated in step S2 is within the preset unsaturated range (e.g., If, based on model predictions, the cabin temperature does not exceed a safe threshold (e.g., 28°C) within the first time domain (e.g., the next hour), the control logic will determine that the prefabricated cabin has sufficient thermal buffering capacity. Under this condition, it generates instructions to keep the active temperature control system off or only operate in a low-power ventilation mode, fully utilizing the physical property of the phase change material to absorb heat at a constant temperature within the phase change range to dissipate the heat generated by the equipment, thereby minimizing the compressor's operating time and achieving passive energy saving. Conversely, when the estimated current phase change saturation reaches the critical saturation threshold (e.g., ... This means that the latent heat capacity is about to be exhausted, or that although the current temperature is normal, the prediction results show that it will exceed the safety threshold in the first time domain. In this case, the control logic will determine that the risk of thermal buffer failure is high. Under this condition, an instruction is generated to start the active temperature control system in advance and replenish the cooling capacity. This "advance" intervention avoids the energy waste and temperature fluctuation caused by temperature overshoot and subsequent cooling in traditional temperature control.

[0079] Furthermore, to further tap into energy-saving potential, this collaborative control strategy also incorporates an economically-based active nighttime recovery logic. This logic first determines whether the current time falls within a nighttime off-peak electricity price period (typically from 22:00 to 06:00 the next day) by analyzing grid electricity price data, and then combines this with the next day's weather forecast from meteorological services to determine if the next day will be a high-temperature, high-load operating condition (e.g., a forecast maximum temperature exceeding 30°C). If the system detects that both conditions are met—being in a nighttime off-peak electricity price period and the next day being confirmed as a high-temperature operating condition—it will trigger a cold storage mode. In this mode, even if the current cabin temperature is within a comfortable range, as long as it is not lower than the minimum temperature constraint allowed by the prefabricated cabin (e.g., 18°C), the active temperature control system will be forcibly activated to start cooling operation. This forced operation will continue until the current phase change saturation calculated by the state estimation module decreases to zero or the preset recovery threshold (i.e., the phase change material completely solidifies and returns to its maximum heat absorption potential state), thus completing the "cold energy storage" of the phase change material.

[0080] This step, through the aforementioned hierarchical coordination and active recovery mechanism, fully utilizes the economic characteristics of peak-valley electricity price differences and the thermal characteristics of phase change materials to cleverly transfer daytime electricity load to nighttime. Under the premise of ensuring equipment thermal safety, it achieves the optimal operating mode in terms of economy and energy efficiency throughout the entire life cycle.

[0081] To verify the beneficial effects of the present invention, a comparative experiment was conducted between the collaborative temperature control method described in this embodiment and the traditional "on-off control" in the same environmental chamber. The experimental results are as follows: Figure 4 As shown.

[0082] Figure 4 (a) and Figure 4 (b) shows a comparison of cabin temperature fluctuations and energy consumption under two control strategies.

[0083] Traditional control method curve: The cabin temperature shows a sawtooth-like, violent fluctuation around the set value (25℃) (fluctuation range ±2℃), and the power curve below shows that the air conditioning compressor starts and stops frequently during the day (about 4-5 times per hour).

[0084] The control method curve of this invention shows that the cabin temperature is smoothly maintained within a safe range of 22°C throughout the day. Especially during the high-temperature period at 14:00, the power curve shows that the air conditioner is basically off or operating at low frequency, mainly relying on the latent heat absorption of the PCM to maintain the temperature; while during the nighttime period from 02:00 to 05:00, the power curve shows an active cold storage peak.

[0085] A comprehensive comparison of the curves from traditional control methods and the control method of this invention shows that the synergistic temperature control method described in this embodiment exhibits significant advantages in both energy consumption and equipment operational stability.

[0086] 1) In terms of energy consumption: By making full use of the latent heat absorption capacity of phase change materials during the high-temperature period of the day and the active cold storage strategy during the low-price period at night, the control method curve of this invention shows that the operation of the active temperature control system is mainly concentrated at night, and the operating time is significantly reduced during the high-load period of the day, realizing the "peak shaving and valley filling" of electricity, thereby greatly reducing the overall operating cost of the whole day.

[0087] 2) Regarding equipment lifespan: The control method curve of this invention clearly shows that the start-stop frequency of the air conditioning compressor is much lower than that of the traditional control method curve. Thanks to the global optimization strategy of model predictive control (MPC), the system avoids the frequent "oscillating" start-stop phenomenon caused by temperature lag in traditional control, thereby effectively reducing the mechanical wear of the compressor and significantly extending the service life of key equipment.

[0088] Example 2, Figure 5 A coordinated active and passive temperature control system for prefabricated substations is presented, including:

[0089] The data acquisition module is used to acquire real-time operating data of the prefabricated substation modules;

[0090] The thermal network modeling and state estimation module is used to construct a prefabricated cabin thermal network model that includes the thermal property parameters of the phase change material, and takes the real-time operating data as input to calculate the current phase change saturation of the phase change material and the predicted temperature inside the cabin in real time through the state estimation algorithm.

[0091] The deviation accumulation and online identification and correction module is used to calculate the deviation between the predicted temperature and the actual monitored temperature, and monitor the cumulative value of the deviation within a set time window; when the cumulative value exceeds a preset mismatch threshold, model parameter identification is performed, and the thermal resistance parameters and thermal capacity parameters in the prefabricated cabin thermal network model are corrected online according to the identification results.

[0092] The rolling optimization decision module is used to generate a collaborative control strategy based on the modified prefabricated cabin thermal network model and the current phase change saturation using a rolling time-domain optimization algorithm.

[0093] The collaborative execution module is used to execute the collaborative control strategy to achieve collaborative operation of the active temperature control system and the passive temperature regulation of the phase change material.

[0094] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0095] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0096] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0097] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0098] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0099] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for coordinated active and passive temperature control in a prefabricated substation, characterized in that, Includes the following steps: Obtain real-time operational data of the prefabricated cabin; A thermal network model containing the thermophysical properties of the phase change material is constructed. Real-time operating data is used as input, and the current phase change saturation of the phase change material and the predicted temperature inside the chamber are obtained through state estimation. Calculate the deviation between the predicted temperature and the actual monitored temperature, and determine whether the cumulative value of the deviation within the set time window exceeds the preset mismatch threshold. If so, perform model parameter identification to correct the thermal resistance and heat capacity parameters of the thermal network model; Based on the modified thermal network model and the current phase change saturation, a collaborative control strategy for active temperature control and passive temperature regulation of phase change material is generated using a rolling time-domain optimization method, and the collaborative control strategy is executed.

2. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, The execution model parameter identification includes the following steps: The active temperature control system is controlled to run at a preset power for a preset test duration to create active thermal disturbances inside the chamber. Collect cabin temperature response data during active thermal disturbance and within a preset decay time after the disturbance ends, and construct the system step response curve; Based on the step response curve, the differential equation of the thermal network model is solved in reverse by using the system identification algorithm to obtain the equivalent heat transfer resistance and the effective heat storage capacity of the phase change material. The thermal network model is updated using the equivalent heat transfer resistance and effective heat storage capacity.

3. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, The method of generating a collaborative control strategy for the active temperature control system and the passive temperature regulation of the phase change material using a rolling time-domain optimization method is obtained by establishing a multi-objective optimization function and solving for its minimum value. The establishment of the multi-objective optimization function includes the following steps: A target optimization function is established, which includes a weighted summation function of energy consumption cost, temperature deviation penalty, and equipment loss. The energy consumption cost is determined based on the predicted operating power and duration of the active temperature control system, the temperature deviation penalty is determined based on the degree to which the predicted cabin temperature deviates from the target temperature, and the equipment loss is determined based on the number of start-stop switching times or power adjustment range of the active temperature control system in the predicted time domain. Under the constraints that the cabin temperature is within a preset safe range and the phase change saturation of the phase change material is within the physical limit, a control sequence that minimizes the weighted summation function is solved as a cooperative control strategy.

4. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, Before generating the collaborative control strategy of active temperature control system and passive temperature regulation of phase change material using rolling time-domain optimization method based on the modified thermal network model and the current phase change saturation, a load prediction step is also included: Collect historical load data and historical meteorological data for the electrical equipment inside the cabin; Using a time series forecasting model, predict the heat generation curve of the cabin equipment and the ambient temperature curve of the cabin environment within a future set time domain; The heat generation curve of the device and the ambient temperature curve are used as disturbance variables and input into the rolling time-domain optimization algorithm to calculate the optimal control sequence in the future time domain.

5. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, The implementation of the aforementioned collaborative control strategy includes hierarchical control: When the estimated current phase change saturation is in the unsaturated range, and the predicted temperature inside the cabin does not exceed the safety threshold in the first time domain in the future, an instruction is generated to control the active temperature control system to remain off or operate at low power, so as to make priority use of the latent heat of the phase change material for passive temperature regulation. When the current phase change saturation reaches the critical saturation threshold, or when the predicted cabin temperature will exceed the safety threshold in the first time domain in the future, an instruction is generated to control the active temperature control system to start and replenish the cooling capacity.

6. The active and passive coordinated temperature control method for prefabricated substations according to claim 5, characterized in that, The collaborative control strategy also includes nighttime active recovery logic: If the current time is during the preset off-peak electricity price period and the weather forecast indicates that the next day will be a high-temperature operating condition, then under the premise of meeting the minimum temperature constraint of the prefabricated cabin, an instruction is generated to control the operation of the active temperature control system until the current phase change saturation of the phase change material is reduced to a preset minimum value, so as to utilize the off-peak electricity price to complete the pre-cooling and energy storage of the phase change material.

7. The active and passive coordinated temperature control method for prefabricated substations according to claim 2, characterized in that, The system identification algorithm is used to solve the differential equation of the thermal network model in reverse. Specifically, the recursive least squares method or particle swarm optimization algorithm is used to minimize the root mean square error between the model output temperature and the actual collected step response curve, and the optimal parameter combination is searched iteratively.

8. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, The calculation logic for the current phase change saturation of the phase change material is as follows: The instantaneous heat flux through the phase change material layer was calculated based on a thermal network model. Integrating the instantaneous heat flux value over the time axis yields the cumulative heat absorbed or released by the phase change material. The accumulated heat is compared with the total latent heat capacity of the phase change material to obtain a normalized phase change saturation value.

9. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, The real-time operating data includes external environmental parameters, multi-point temperatures inside the cabin, real-time power of electrical equipment inside the cabin, and the current operating status of the active temperature control system.

10. The active and passive coordinated temperature control method for prefabricated substations according to claim 1, characterized in that, The collaborative control strategy includes the switching sequence and power setting of the active temperature control system in the future time domain.

Images