Multi-device temperature control cooperative regulation method and system

Abstract

The application discloses a kind of multi-device temperature control collaborative regulation method and system, it is related to intelligent control technical field, including S1, by the multidimensional variable data of current temperature control scene is collected, including temperature humidity device state and user demand, using cosine similarity method calculation and the matching value of case in historical experience database, obtain similar historical case set;S2, according to the case of the highest matching value in similar historical case set, extract its device collaborative configuration parameter as initial reference, fusion device power distribution and collaborative timing sequence and parameter threshold range and historical load mode, determine preliminary parameter adjustment basis;The multi-device temperature control collaborative regulation method and system, significantly improve the real-time decision efficiency of temperature control strategy and energy efficiency optimization level, adapt to complex dynamic environment, realize accurate collaborative control.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to a method and system for coordinated temperature control of multiple devices. Background Technology

[0002] In the field of industrial automation, multi-device collaborative temperature control technology has attracted much attention due to its crucial role in improving energy efficiency, optimizing user experience, and ensuring equipment operational stability. With the widespread adoption of IoT technology, multiple devices need to work collaboratively to achieve precise temperature control. For example, in smart homes, air conditioners, heaters, and ventilation equipment or semiconductor temperature control devices need to work together to maintain a comfortable indoor temperature. In industrial scenarios, multiple devices need to cooperate to ensure a stable production environment. The importance of this technology lies in its ability to meet complex and changing environmental demands through dynamic collaboration between devices, while simultaneously reducing energy consumption.

[0003] However, current multi-device temperature control methods have significant shortcomings when dealing with dynamic scenarios. Existing methods typically rely on preset control rules or static device collaboration models, making it difficult to adapt to rapid responses to environmental changes or new demands. For example, when the number of people indoors suddenly increases or the external weather changes abruptly, existing systems often require manual adjustment or reconfiguration, lacking adaptive capabilities. This limitation mainly stems from insufficient utilization of historical experience and weak ability to quickly match and optimize parameters for new scenarios. Existing solutions struggle to achieve efficient decision-making and precise control when handling complex and ever-changing scenarios. The core technical challenge lies in how to quickly identify historical cases similar to the current scenario in a dynamic environment and adjust parameters based on experience from similar scenarios. First, scenario similarity calculation is a critical issue. Temperature control scenarios involve multiple variables, such as temperature, humidity, device status, and user needs. How to comprehensively measure the similarity between scenarios directly affects the efficiency of subsequent decision-making. Therefore, designing a robust similarity calculation method based on multi-dimensional variables to ensure that the most relevant historical cases can still be quickly found in complex dynamic scenarios becomes a key issue for multi-device temperature control collaboration. Another technical challenge is parameter fine-tuning based on similar cases. Because historical scenarios differ from current scenarios, simply reusing historical solutions may lead to control deviations. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-device temperature control and coordination method and system to solve the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a multi-device temperature control collaborative adjustment method, comprising: S1, collecting multi-dimensional variable data of the current temperature control scenario, including temperature, humidity, device status, and user needs, and calculating the matching value with cases in a historical experience database using the cosine similarity method to obtain a set of similar historical cases; S2, extracting the device collaborative configuration parameters of the case with the highest matching value in the set of similar historical cases as an initial benchmark, and determining the preliminary parameter adjustment basis by integrating device power allocation and collaborative time sequence, parameter threshold range, and historical load mode; S3, if the difference between the preliminary parameter adjustment basis and the current dynamic environmental variables exceeds a preset threshold, optimizing the initial benchmark parameters through a genetic algorithm, incorporating configuration compatibility verification and benchmark weight factors, as well as adjustment base vectors and initial error calibration, to obtain an adjustment parameter set adapted to the current scenario; S4, obtaining... After selecting and adjusting the parameter set, the temperature control process is simulated for the equipment collaborative needs. The simulated heat flow distribution, control cycle iteration, demand response curve, and temperature gradient mapping are integrated to determine whether the simulation results meet the energy efficiency optimization standards, thus obtaining the verified parameter scheme. S5: The multi-equipment collaborative control model is updated using the verified parameter scheme. Combined with process energy consumption estimation, collaborative feedback loop, control stability verification, and simulated boundary conditions, the temperature control strategy after improving real-time decision-making efficiency is determined. S6: Environmental changes are monitored through the temperature control strategy. If control deviation occurs, similar cases are obtained from historical experience for supplementary adjustments. The equipment power allocation, collaborative timing sequence, simulated heat flow distribution, and control cycle iteration are integrated to obtain stable parameter output. S7: The stable parameter output is deployed to the actual equipment collaborative system to determine whether user needs are met, thus obtaining the final temperature control execution result.

[0006] Preferably, step S1 includes: collecting temperature data, humidity data, and equipment status of the temperature control scenario in real time through sensors; forming a multidimensional variable dataset by combining it with user needs; standardizing the multidimensional variable dataset using a preprocessing method to obtain a normalized dataset; calculating the matching degree value between the normalized dataset and each case in the historical experience database using a cosine similarity algorithm to obtain a matching degree value set; if the matching degree value exceeds a preset threshold, extracting the corresponding case from the historical experience database to generate a preliminary similar case set; sorting the preliminary similar case set according to user needs to obtain an optimized similar case set; generating recommended parameter configurations for the temperature control scenario by analyzing the equipment status and historical experience in the optimized similar case set; and adjusting the equipment in the current temperature control scenario using the recommended parameter configurations to obtain a real-time optimized equipment operating status.

[0007] Preferably, step S2 includes: obtaining the case with the highest matching value from the set of similar cases; extracting the device collaborative configuration parameters contained in the case with the highest matching value to determine the initial baseline configuration; analyzing the device power allocation and collaborative timing sequence based on the initial baseline configuration, calculating the load balancing coefficient, and obtaining an optimized device allocation scheme; using the optimized device allocation scheme and combining it with parameter threshold ranges to generate dynamic adjustment rules for device operating status; if the deviation between the dynamic adjustment rules and historical load patterns exceeds a preset threshold, extracting relevant load characteristics from historical experience data and updating the adjustment rules; generating real-time device collaborative control instructions based on the updated adjustment rules to determine device operating parameters; generating parameter optimization results through real-time device operating parameters and combining them with historical experience data to obtain the final device control configuration; extracting operating status data from the final device control configuration, storing it in the historical experience database, and updating the set of similar cases.

[0008] Preferably, step S3 includes: if the deviation between the initial parameters and the dynamic environment exceeds a preset threshold, obtaining the deviation value through deviation calculation, iteratively optimizing the initial benchmark using a genetic algorithm to obtain an optimized benchmark parameter set; analyzing the adaptability of the optimized benchmark parameter set to the current scene through configuration compatibility verification, and generating a compatibility verification result by adjusting the weight factors; calibrating the basic vectors using a linear regression model based on the compatibility verification result to obtain a calibrated vector set by fusing the weight factors; if the deviation between the calibrated vector set and the initial error exceeds a preset threshold, correcting the initial error using a gradient descent algorithm to generate error correction parameters; generating an adjustment parameter set adapted to the current scene by fusing the calibrated vector set based on the error correction parameters; generating equipment operation control commands by adjusting the parameter set and combining it with the dynamic environmental variables of the current scene using real-time data stream processing; and acquiring real-time operating status data based on the equipment operation control commands, storing it in the dynamic environment database, and updating the adaptability parameters of the current scene.

[0009] Preferably, step S4 includes acquiring an adjustment parameter set, combining it with equipment coordination requirements, and using real-time data stream processing to generate an initial temperature control instruction set; based on the initial temperature control instruction set, running a heat flux distribution simulation to calculate the heat flux distribution matrix during equipment operation; using the heat flux distribution matrix, employing time series analysis to iteratively optimize the control cycle and generate control cycle parameters; based on the control cycle parameters and real-time operating status data, generating a demand response curve; if the peak value of the demand response curve deviates from a preset threshold beyond a standard, then using a gradient descent algorithm to optimize the curve parameters to obtain an optimized response curve; using the optimized response curve and combining environmental variable analysis to generate a temperature gradient mapping; if the deviation between the temperature gradient mapping and the energy efficiency optimization standard exceeds a preset threshold, then adjusting the fine-tuning parameter set to generate a verified parameter scheme.

[0010] Preferably, step S5 includes obtaining an initial collaborative control model from the verified parameter scheme, generating a multi-device collaborative instruction set through real-time data stream processing; calculating an energy consumption distribution matrix using the multi-device collaborative instruction set, and extracting energy consumption distribution characteristics using time series analysis; constructing a collaborative feedback loop based on the energy consumption distribution characteristics, and iteratively updating the control instruction set; if the deviation between the iterated control instruction set and the preset regulation stability threshold exceeds a standard, then using a gradient descent algorithm to optimize the instruction set parameters to obtain an optimized control instruction set; simulating boundary conditions using the optimized control instruction set to generate a boundary condition response matrix; analyzing real-time decision-making efficiency based on the boundary condition response matrix to determine an improved temperature control strategy; and updating the collaborative control model using the improved temperature control strategy to generate the final multi-device collaborative operation scheme.

[0011] Preferably, step S6 includes: acquiring real-time environmental temperature control data through environmental sensor data streams; generating a smoothed temperature control data stream using data smoothing processing to obtain a continuous environmental state sequence; if the continuous environmental state sequence deviates from a preset temperature control threshold, retrieving similar cases matching the environmental conditions from a historical case library to obtain a matching case set; extracting a device power allocation sequence from the matching case set, fusing the current device power allocation using a weighted average algorithm to obtain an optimized power allocation scheme; constructing a heat flow distribution model based on the optimized power allocation scheme and the collaborative time series to generate a heat flow distribution matrix; processing the heat flow distribution matrix using time series analysis to extract the heat flow change trend to obtain a heat flow trend sequence; iteratively adjusting the periodic parameters using the heat flow trend sequence; if the deviation between the iterative parameters and the preset stable threshold exceeds the range, optimizing the parameters using a gradient descent algorithm to obtain a stable control parameter set; updating the collaborative time series based on the stable control parameter set to generate the final temperature control operation scheme.

[0012] Preferably, step S7 includes real-time monitoring of the preliminary operation results through the equipment collaboration system, extracting the trend of operation results using time series analysis to obtain an operation trend sequence; if the deviation between the operation trend sequence and the user demand threshold exceeds a preset range, obtaining historical trend cases similar to the current operation trend sequence from the historical operation database to obtain a historical trend set; extracting equipment operation adjustment parameters based on the historical trend set, and fusing the current operation trend sequence using a support vector machine algorithm to obtain an adjustment parameter set.

[0013] Preferably, step S7 further includes updating the operating configuration of the device coordination system by adjusting the parameter set, generating an updated temperature control operating scheme, and obtaining optimized operating results; if the optimized operating results still deviate from the user's required threshold, the gradient descent algorithm is used to iteratively optimize the adjusted parameter set to obtain a stable operating parameter set; the device coordination system is reconfigured according to the stable operating parameter set to generate the final temperature control execution result.

[0014] A multi-device temperature control collaborative adjustment system is provided to implement the steps of the aforementioned multi-device temperature control collaborative adjustment method. The system includes a data acquisition module, which collects multi-dimensional variable data of the current temperature control scenario, including temperature, humidity, device status, and user needs. It then uses a cosine similarity method to calculate the matching value with cases in a historical experience database to obtain a set of similar historical cases. A parameter extraction module extracts the device collaborative configuration parameters of the case with the highest matching value in the similar historical case set as an initial benchmark. It integrates device power allocation, collaborative time sequence, parameter threshold range, and historical load patterns to determine the initial parameter adjustment basis. A parameter optimization module optimizes the initial benchmark parameters using a genetic algorithm if the difference between the initial parameter adjustment basis and the current dynamic environmental variables exceeds a preset threshold. This optimization incorporates configuration compatibility verification, benchmark weight factors, adjustment base vectors, and initial error calibration to obtain an adjustment parameter set adapted to the current scenario. Simulation... The verification module, after obtaining the adjusted parameter set, simulates the temperature control process for equipment collaboration needs, integrating simulated heat flow distribution, control cycle iteration, demand response curve, and temperature gradient mapping to determine whether the simulation results meet energy efficiency optimization standards, thus obtaining the verified parameter scheme. The model update module updates the multi-equipment collaborative control model using the verified parameter scheme, combining process energy consumption estimation, collaborative feedback loops, control stability checks, and simulated boundary conditions to determine the temperature control strategy after improving real-time decision-making efficiency. The strategy monitoring module monitors environmental changes through the temperature control strategy; if control deviations occur, it re-obtains similar cases from historical experience for supplementary adjustments, integrating equipment power allocation, collaborative timing sequences, simulated heat flow distribution, and control cycle iteration to obtain stable parameter output. The execution module deploys the stable parameter output to the actual equipment collaboration system, determines whether user needs are met, and obtains the final temperature control execution result.

[0015] As can be seen from the above technical solution, the present invention has the following beneficial effects:

[0016] This multi-device temperature control collaborative adjustment method and system addresses the challenges of collaborative control and energy efficiency optimization in complex temperature control scenarios. It collects multi-dimensional variables such as temperature, humidity, equipment status, and user needs, and uses cosine similarity to match similar cases from a historical experience database. The collaborative configuration parameters of highly matched cases are extracted as an initial benchmark, integrating power allocation, collaborative timing, and load patterns. If the difference from the current environment exceeds a threshold, a genetic algorithm optimizes the parameters, combining compatibility verification, weighting factors, and error calibration to generate a fine-tuning parameter set. This invention verifies parameter energy efficiency through simulated heat flow distribution, iterative control cycles, and demand response curves, updates the collaborative control model, and generates a real-time temperature control strategy by combining energy consumption estimation, feedback loops, and stability checks. Deviations are dynamically monitored and adjustments are made, ultimately deploying stable parameters to the equipment system to ensure user needs are met. This invention significantly improves the real-time decision-making efficiency and energy efficiency optimization level of temperature control strategies, adapts to complex dynamic environments, and achieves precise collaborative control. Attached Figure Description

[0017] Figure 1 This is a flowchart of the multi-device temperature control and coordinated adjustment method of the present invention;

[0018] Figure 2 This is a connection diagram of the multi-device temperature control and coordinated adjustment system of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1As shown, this invention provides a technical solution: a multi-device temperature control collaborative adjustment method, including S1, collecting multi-dimensional variable data of the current temperature control scenario, including temperature, humidity, device status, and user needs, and calculating the matching value with cases in a historical experience database using the cosine similarity method to obtain a set of similar historical cases; S2, extracting the device collaborative configuration parameters of the case with the highest matching value in the set of similar historical cases as an initial benchmark, and integrating device power allocation and collaborative time sequence, as well as parameter threshold range and historical load mode, to determine the preliminary parameter adjustment basis; S3, if the difference between the preliminary parameter adjustment basis and the current dynamic environmental variables exceeds a preset threshold, optimizing the initial benchmark parameters through a genetic algorithm, incorporating configuration compatibility verification and benchmark weight factors, as well as adjustment basis vectors and initial error calibration, to obtain an adjustment parameter set adapted to the current scenario; S4, obtaining the adjustment... After assembling the parameter set, the temperature control process is simulated based on the equipment coordination requirements. The simulated heat flow distribution, control cycle iteration, demand response curve, and temperature gradient mapping are integrated to determine whether the simulation results meet the energy efficiency optimization standards, thus obtaining the validated parameter scheme. S5: The multi-equipment coordinated control model is updated using the validated parameter scheme. Combined with process energy consumption estimation, coordinated feedback loop, control stability verification, and simulated boundary conditions, the temperature control strategy after improving real-time decision-making efficiency is determined. S6: Environmental changes are monitored through the temperature control strategy. If control deviation occurs, similar cases are obtained from historical experience for supplementary adjustments. The equipment power allocation, coordinated timing sequence, simulated heat flow distribution, and control cycle iteration are integrated to obtain stable parameter output. S7: Based on the stable parameter output, the system is deployed to the actual equipment coordination system to determine whether user needs are met, thus obtaining the final temperature control execution result.

[0021] This implementation method is based on a data-driven intelligent matching and optimization algorithm. First, it collects multi-dimensional environmental variables (such as temperature and humidity), equipment operating status, and user-defined requirements. Using cosine similarity, it selects the control case most similar to the current scenario from a historical database and extracts its equipment configuration parameters as the initial adjustment basis. Then, it compares this initial configuration with the current environmental dynamic variables. If the difference is significant, a genetic algorithm is used to optimize the parameters, considering parameter compatibility, benchmark weighting factors, and error adjustments to generate a new set of suitable parameters. Next, these parameters are simulated in a simulated environment for heat flow and control cycles. The response curves and temperature distribution are used to determine whether the energy-saving and temperature control targets are met, resulting in a validated configuration scheme. Finally, this scheme is applied to a real-world multi-device control system. Based on feedback loops and energy consumption assessment, the control strategy is further optimized to achieve real-time, intelligent temperature control. During operation, the system continuously monitors environmental changes and automatically calls historical cases to compensate and correct parameters when deviations occur, ensuring long-term stable operation of the control system.

[0022] This invention combines historical experience with real-time data analysis to achieve intelligent collaborative optimization of temperature control systems in complex multi-device environments, effectively improving energy efficiency and response accuracy. Similarity calculation ensures high matching of configuration schemes and reduces parameter initialization errors; genetic algorithms enable efficient parameter search and optimization, enhancing adaptability and control sensitivity; temperature simulation and response mapping processes ensure high reliability of the solution before actual deployment; real-time monitoring and adaptive adjustment capabilities enhance the system's robustness and long-term operational stability; the overall system structure supports flexible deployment and expansion across devices and scenarios, demonstrating good practicality and market prospects.

[0023] S1 includes: real-time acquisition of temperature, humidity, and equipment status data from sensors in the temperature control scenario; combining this with user needs to form a multidimensional variable dataset; standardizing the multidimensional variable dataset using preprocessing methods to obtain a normalized dataset; calculating the matching degree between the normalized dataset and each case in the historical experience database using a cosine similarity algorithm to obtain a matching degree value set; if the matching degree value exceeds a preset threshold, extracting the corresponding case from the historical experience database to generate a preliminary similar case set; sorting the preliminary similar case set according to user needs to obtain an optimized similar case set; generating recommended parameter configurations for the temperature control scenario by analyzing the equipment status and historical experience in the optimized similar case set; and adjusting the equipment in the current temperature control scenario using the recommended parameter configurations to obtain a real-time optimized equipment operating status.

[0024] First, temperature and humidity sensors installed at key points within the temperature-controlled environment continuously collect current temperature and humidity values ​​within the controlled area at 1-second intervals. Temperature is recorded in degrees Celsius, and humidity is recorded as a percentage. Simultaneously, the system collects current status data from all on-site temperature-controlled devices (such as air conditioners, fresh air systems, and geothermal systems) via control interfaces. This includes whether each device is on, its current operating mode (specific values: 0 for off, 1 for cooling, 2 for heating, 3 for ventilation), set temperature value (in degrees Celsius), fan speed level (integer values: 1 for low, 2 for medium, 3 for high), and current cumulative operating time (in minutes). In parallel, the system records the user's currently set target temperature lower limit, target temperature upper limit, desired humidity range, adjustment response time, and control target preference (e.g., energy saving priority or comfort priority; specific values: 1 for comfort priority, 2 for energy saving priority). All the above data items are aggregated to form a multidimensional variable dataset for the current temperature control scenario. This dataset is a vector containing several real and integer values ​​in a fixed order, with each dimension corresponding to a variable with a specific meaning. After data acquisition, the system standardizes the original dataset to eliminate the influence of dimensional differences between different physical quantities on subsequent calculations. The standardization method uses min-max normalization, specifically: first, the system looks up the minimum and maximum values ​​of the variable in the historical database; then, it subtracts the minimum value from the current variable value, and divides the result by the difference between the maximum and minimum values ​​to obtain the standardized normalized value. This process is performed on each dimension of the multidimensional variable dataset one by one, ultimately forming a normalized dataset where all variable values ​​are uniformly mapped to the range of 0 to 1. Next, the system retrieves all stored standardized case data from the historical experience database and compares the similarity between the feature vector of each case and the current normalized dataset. The specific calculation method is as follows: First, the system multiplies the values ​​of each corresponding dimension in the current normalized data and the data of a certain historical case, and sums the products of all dimensions. Second, the system calculates the sum of squares of the values ​​of all dimensions in the current data vector and the historical case vector, and takes the square root. Third, the system uses the sum of the aforementioned products as the numerator and the product of the square roots as the denominator, and performs a division calculation. The result is the cosine similarity value between the historical case and the current scene. This value is between 0 and 1, with the closer to 1 indicating greater similarity. The system sequentially calculates the similarity between the current normalized data and all historical cases in the database, forming a complete set of matching scores. Each value in this set corresponds to a specific historical case and is used for subsequent similar case filtering and sorting operations.

[0025] After calculating the similarity between the current temperature control scenario and historical cases, a matching score set containing all historical cases' matching scores is obtained. The system sets a fixed similarity filtering threshold of 0.75, determined after statistical analysis of 1,000 real-world environmental samples, striking a balance between representativeness and quantity in the selected cases. When the system iterates through the matching score set, for each matching score value, it checks if it is greater than or equal to 0.75. If this condition is met, all records of the historical case corresponding to that matching score value are extracted from the database, including the case's environmental variables, equipment status configuration, response process, energy consumption records, and final temperature control effect. These historical cases that meet the similarity threshold together constitute a preliminary similar case set. Each case in the set is accompanied by its original matching score information and a case number for subsequent sorting and referencing.

[0026] After obtaining a preliminary set of similar cases, the system sorts the set according to the user's current preferences set in the control terminal. User preferences include the priority selection of control objectives, i.e., whether comfort or energy saving is prioritized. If the user sets comfort as the priority, the system calculates the environmental stability indicators for each case within the control period, mainly including temperature fluctuation amplitude, humidity fluctuation amplitude, and response time. Temperature fluctuation amplitude is the difference between the highest and lowest temperatures within the control period, humidity fluctuation amplitude is the difference between the maximum and minimum humidity values, and response time is the time (in minutes) from the start of control until the indoor temperature and humidity reach the user-set target range. The system scores all preliminary cases based on these three indicators, with lower scores indicating higher comfort. The system calculates a weighted average of the three indicators, using a weight of 0.5 for temperature fluctuation, 0.3 for humidity fluctuation, and 0.2 for response time, and then sorts them from lowest to highest based on the overall score. If the user's preference is set to prioritize energy efficiency, the system calculates the total energy consumption (in kilowatt-hours), average equipment operating time (in minutes), and temperature change rate per unit time (i.e., the average speed at which the temperature approaches the set value per minute) for each case within the control cycle. A weighted comprehensive score is calculated, with total energy consumption weighted at 0.6, equipment operating time at 0.3, and temperature change rate at 0.1, and the cases are sorted from lowest to highest score. After sorting, the system generates an optimized set of similar cases, where each case is sorted based on both matching degree and user priority objective, with higher priority cases listed first. This allows for the extraction of recommended configuration parameters for equipment control in the current scenario.

[0027] After obtaining the optimized set of similar cases, the system iterates through the equipment status information and control process data of each case in the set, extracting control parameters directly relevant to the current temperature control scenario. Specifically, the extracted parameters include: the operating mode of each device, the set temperature value, the wind speed level, the average start-stop frequency, the continuous running time, the device response time, and the final stable temperature and humidity values. To ensure the representativeness and adaptability of the parameters, the system uses statistical analysis methods to process similar parameters in each case in the set. The mode, mean, and standard deviation are calculated for each device's set temperature, wind speed level, operating mode, and other parameters. If the mode of a parameter occurs more than 60% of the time in the set, the system prioritizes using that mode as the recommended value; if there is no obvious mode or the parameter value has large dispersion, the mean is used as the recommended value; parameters with a standard deviation greater than 0.2 are marked as unstable terms, and upper and lower tolerance limits are added to the recommended values ​​as adjustment ranges. In addition, the system references historical operational records from similar cases, including the average time required to achieve the target temperature and humidity under set parameters, the duration of maintenance within the target range, and the resulting energy consumption trends. This data helps optimize recommended configurations, ensuring both comfort and energy efficiency goals are met while maintaining good practical controllability. The system integrates these analysis results into a set of recommended parameter configurations, which include: the set temperature value (in degrees Celsius) for each device, the fan speed level (integer value), the operating mode number (integer value, e.g., 1 for cooling, 2 for heating), the estimated operating time (in minutes), and whether to enable the real-time temperature and humidity feedback correction function (Boolean value). The system then uploads the recommended parameter configurations to the device control module and issues control commands to each device. Specifically, this includes: writing the set temperature value to the thermostat via the communication bus or wireless interface, synchronously writing the set fan speed level parameter to the fan speed adjustment circuit, sending the operating mode signal to the device's main control chip, and setting a minimum operating time of no less than 10 minutes to prevent frequent device start-ups and shutdowns. During execution, the control module continuously collects the current temperature and humidity trends and compares them with recommended target values. If the device's operating status is detected to be stable within the target range, the current control strategy is maintained. If a temperature deviation exceeds 1 degree Celsius or a humidity deviation exceeds 5 percent, the system activates the dynamic correction function, automatically fine-tuning the set parameters within a range of no more than 1 unit above or below the original recommended value. After the device is adjusted, a real-time optimized device operating status based on the recommended parameters is formed. This status ensures user needs while possessing a higher energy efficiency ratio and faster adjustment response speed. The system records this status in the control log and uses it for subsequent strategy learning and iterative optimization.

[0028] S2 includes: obtaining the case with the highest matching value from the set of similar cases; extracting the device collaborative configuration parameters contained in the case with the highest matching value to determine the initial baseline configuration; analyzing the device power distribution and collaborative timing sequence based on the initial baseline configuration, calculating the load balancing coefficient, and obtaining the optimized device allocation scheme; using the optimized device allocation scheme and combining it with the parameter threshold range, generating dynamic adjustment rules for the device operating status; if the deviation between the dynamic adjustment rules and the historical load pattern exceeds a preset threshold, extracting relevant load characteristics from historical experience data and updating the adjustment rules; generating real-time device collaborative control instructions based on the updated adjustment rules to determine the device operating parameters; generating parameter optimization results by combining the real-time device operating parameters with historical experience data to obtain the final device control configuration; extracting operating status data from the final device control configuration, storing it in the historical experience database, and updating the set of similar cases.

[0029] In step S2 of this embodiment, the similar cases are first sorted from high to low according to their matching values, and the single case with the highest matching value is selected. The device collaborative configuration parameters recorded in this case are read. These parameters include the operating mode, set temperature, operating power, start and stop time, start and stop interval between adjacent devices, minimum interval between two consecutive start and stop of the same device, allowable upper and lower limits of set temperature, allowable upper and lower limits of power, allowable range of fan speeds, allowable upper limit of heating and cooling rate, allowable upper limit of power ramp-up and fall rate, allowable time offset range, and the historical energy consumption and time required to reach the target temperature and humidity under this configuration. The system directly uses the above parameters as the initial baseline configuration. Subsequently, the system calculates power allocation and collaborative timing based on the initial baseline configuration. The calculation process of power allocation is to first calculate the initial... The total power is calculated by summing the operating power of all devices under the baseline configuration. Then, the operating power of each device is divided by the total power to obtain its power percentage. The device with the largest and smallest power percentage is recorded. The load balancing coefficient is obtained by dividing the larger power percentage by the smaller power percentage. The system aims for a load balancing coefficient of no more than 1.20, determined through comparative testing of 1,000 historical samples, ensuring optimal balance between comfort and energy consumption. If the load balancing coefficient exceeds 1.20, the system adjusts the power percentage in fixed increments: each round, the power percentage of the device with the largest power percentage is reduced by 0.05, while the power percentage of the device with the smallest power percentage is increased by 0.05. If a device reaches its upper or lower power limit, the adjustment ceases until the load balancing coefficient is no greater than 1.The optimized equipment allocation scheme is obtained when 20 or all devices reach the boundary. After obtaining the optimized equipment allocation scheme, the system generates dynamic adjustment rules based on parameter threshold ranges. The rule generation process involves defining three types of control boundaries and two types of timing constraints for each device. The three types of control boundaries include the set allowable temperature range, the allowable operating power range, and the allowable fan speed range. The two types of timing constraints include the minimum start-stop interval for the same device and the minimum start-stop interval between different devices. The system writes the above boundaries and constraints into the rules one by one, and sets specific values ​​for the upper limit of the heating / cooling rate and the upper limit of the power ramp-up and fall-down rate for each device. These values ​​are obtained by statistically analyzing similar cases. The historical median value of the device type within the stable period is determined, and a tolerance of no more than 10% above or below the median value is used as the upper and lower limits. When generating rules, the system simultaneously defines a timing offset strategy: when multiple devices need to start or stop at the same time, a start or stop time offset of no less than 1 minute is added sequentially according to the device power proportion from largest to smallest to avoid instantaneous power peaks. Subsequently, the system compares the generated dynamic adjustment rules with historical load patterns point by point. Specifically, on a unified time scale, the percentage difference between the power predicted by the rule and the average power of similar historical scenarios is calculated as a percentage of the historical value, and the average of all points is taken as the deviation value. If the deviation value is greater than 10%, it is considered to exceed a preset threshold. The threshold value is determined through a grid search of 1,000 samples within an integer range of 5 to 20, with 10 being the optimal value for a combined energy consumption and response time score. When the deviation exceeds 10, the system extracts load characteristics from historical data that match the current target temperature, target humidity, number of devices, and space volume. These load characteristics include power ramp-up curves, fall-down curves, steady-state power levels, and typical start-stop interval distributions under different operating conditions. The system updates the dynamic adjustment rules based on these characteristics. The update rules are to bring the upper limits of the heating / cooling rate and the upper limits of the power ramp-up / fall-down rate closer to the historical median value, with each adjustment not exceeding 10% of the original value, and to increase the minimum start-stop interval. The updated rules are obtained by adjusting the values ​​to at least the 90th percentile of the historical load pattern. After obtaining the updated rules, the system generates real-time equipment collaborative control instructions and determines the equipment operating parameters. The instructions include the operating mode, set temperature, operating power or power level, fan speed level, start time, stop time, and necessary time offset for each device. The system issues the instructions in sequence and compares the real-time temperature and humidity with the target range and the real-time power with the rule upper limit every 1 minute during execution. If any indicator touches the boundary, a single-step fine adjustment is made according to the maximum adjustment range allowed by the rules. The single-step fine adjustment range is such that the set temperature does not exceed 1 degree Celsius and the power ratio does not exceed 0.5. Wind speed settings should not be changed by more than one level, and the minimum continuous operating time should be no less than 10 minutes. The system generates parameter optimization results with a 15-minute evaluation window. The evaluation indicators are the percentage of time within the target temperature and humidity range, total energy consumption, and number of start-stop cycles. The system uses the optimal range given by historical experience data as a reference. If the percentage of time is not less than 90%, the total energy consumption is not higher than the baseline by 5%, and the number of start-stop cycles is not more than the baseline, then the current parameter set is determined as the final equipment control configuration. Otherwise, the single-step fine-tuning strategy in the previous section continues to iterate until the conditions are met or the maximum number of iterations (10) is reached. At the end of the iteration, the parameter set with the best comprehensive indicators is used as the final equipment control configuration. Finally, the system extracts the operating status data from the final equipment control configuration and writes it into the historical experience database. The written content includes the final operating mode, set temperature, operating power, wind speed setting, start-stop time, minimum start-stop interval, time offset, percentage of time within the target range, total energy consumption, and number of start-stop cycles. At the same time, feature entries for similarity matching are generated and the similar case set is updated.

[0030] S3 includes the following steps: If the deviation between the initial parameters and the dynamic environment exceeds a preset threshold, the deviation value is obtained through deviation calculation, and the initial benchmark is iteratively optimized using a genetic algorithm to obtain an optimized benchmark parameter set; through configuration compatibility verification, the adaptability of the optimized benchmark parameter set to the current scene is analyzed, and a compatibility verification result is generated by adjusting the weight factors; based on the compatibility verification result, the weight factors are fused, and a linear regression model is used to calibrate the basic vectors to obtain a calibrated vector set; if the deviation between the calibrated vector set and the initial error exceeds a preset threshold, the initial error is corrected using a gradient descent algorithm to generate error correction parameters; based on the error correction parameters, the calibrated vector set is fused to generate an adjustment parameter set adapted to the current scene; by adjusting the parameter set and combining it with the dynamic environmental variables of the current scene, real-time data stream processing is used to generate equipment operation control commands; based on the equipment operation control commands, real-time operating status data is obtained, stored in the dynamic environment database, and the adaptability parameters of the current scene are updated.

[0031] In step S3 of this embodiment, the system takes the initial reference parameter set output in step S2 and the dynamic environmental variables of the current scene as input. It first performs deviation determination and, when the threshold is exceeded, carries out a series of deterministic optimization and calibration operations until executable equipment operation control commands are generated and the adaptability parameters are updated. The specific process is as follows: The system first calculates the deviation value between the preliminary parameters and the dynamic environment. For the temperature parameter, the absolute value of the difference between the currently measured temperature and the center value of the user's target temperature range is divided by the width of the user's target temperature range to obtain the relative temperature deviation. For the humidity parameter, the absolute value of the difference between the currently measured humidity and the center value of the user's target humidity range is divided by the width of the user's target humidity range to obtain the relative humidity deviation. For the equipment state consistency parameter, the initial parameters are used... The inconsistency degree is calculated by comparing the operating mode, wind speed, and power of each device with the actual status of the current device item by item, and taking the complement of the consistency ratio. This inconsistency degree is directly used as the relative deviation of the device status. The system calculates a weighted average deviation with a weight of 0.5 for temperature, 0.3 for humidity, and 0.2 for device status, and uses 0.10 as the deviation threshold. The threshold is derived from offline verification of 1,000 real samples between 0.05 and 0.15. Considering user demand satisfaction rate, energy consumption, and convergence time, 0.10 is selected as the optimal fixed value. When the deviation value is greater than or equal to 0.10, the system starts a genetic algorithm to iteratively optimize the initial baseline. The population size of the genetic algorithm is 50, and the initial generation has 1000 samples. The body is generated by symmetrical perturbation of the initial baseline parameters within their respective allowable ranges, with a perturbation not exceeding 5. The fitness function consists of three parts: the weighted average deviation from the dynamic environment, energy consumption per unit time, and temperature and humidity fluctuation amplitude, with weights of 0.5, 0.3, and 0.2 respectively. The selection operator adopts a ternary competition method, selecting the best-fitter candidate from three candidates each time to participate in breeding. Crossover is performed using single-point crossover, and the parameters after crossover are truncated to ensure that they do not exceed the parameter threshold range. The crossover probability is 0.8. Mutation uses uniform perturbation with an amplitude not exceeding 2 of the allowable range of each parameter. The mutation probability is 0.05. The elite retention quantity is 2. The termination condition is reaching 100 generations or 10 consecutive generations with a fitness improvement of less than 1. The optimized algorithm output is... The system then performs a configuration compatibility check, which includes single-device boundary checks and cross-device timing coordination checks. The single-device boundary check verifies each device's set temperature, operating power, fan speed, and the interval between two consecutive start-stop cycles. The cross-device timing coordination check verifies the simultaneous start-up of similar high-power devices according to a staggered start-up rule of at least one minute and checks whether the minimum start-stop interval between different devices meets the requirements. The system scores each item proportionally and deducts points from violations to form a compatibility score between 0 and 1. A score less than 0 indicates a lower compatibility score.At a score of 85, the algorithm returns to the genetic algorithm stage to further optimize by reducing the perturbation amplitude near the boundary. When the score is greater than or equal to 0.85, it enters the weight factor adjustment process. The weight factor is used to emphasize the importance of different output targets in subsequent calibration. The system initializes the temperature stability factor, humidity stability factor, and energy consumption factor to 1. When the temperature fluctuation exceeds 50% of the target temperature range width within a short window, the temperature stability factor is set to 1.2. When the energy consumption per unit time is higher than 10% of the median value of similar scenarios, the energy consumption factor is set to 1.1. The humidity stability factor is set to 1.2 when the humidity fluctuation exceeds 50% of the target humidity range width; otherwise, it remains at 1. After applying the tolerance score and weighting factors, the system calibrates the base vector. The base vector is obtained by arranging the optimized baseline parameter set in a fixed order. The linear regression model uses temperature change, humidity change, and instantaneous energy consumption within a short window as the predicted variables. The training samples are a concatenated set of similar case data and near-end data of the current scene. The window length is 30 minutes, and the number of samples is no less than 60. The error metric is the sum of squared errors, with the errors of temperature, humidity, and energy consumption weighted according to the aforementioned weighting factors. The numerical solution adopts a step-by-step iterative minimization process until the error decreases to less than 1 after two iterations or reaches 200 iterations. The obtained parameters are used to calibrate the base vector. The system corrects each item to obtain the calibrated vector set. Then, it calculates the deviation between the calibrated vector set and the initial error. The initial error is defined as a single value obtained by summing the deviations of the average temperature deviation, average humidity deviation, and energy consumption per unit time relative to the historical median of similar data within one evaluation cycle of the current scenario, weighted at 0.5, 0.3, and 0.2 respectively. The deviation threshold is set to 0.05, which is optimally determined by a grid search within the range of 0.03 to 0.10, taking into account three conditions: a user demand satisfaction rate of no less than 90%, energy consumption per unit time no more than 5 times the median of similar data, and convergence iterations not exceeding 100. When the above deviation is greater than or equal to 0.05, the system... The system initiates a gradient descent algorithm to correct the initial error, with a learning step size of 0.01 and a maximum number of iterations of 200. Each iteration evaluates the sensitivity of the error to changes in each parameter through numerical differentiation and fine-tunes the parameters along the error descent direction, with the adjustment not exceeding 2% of their respective allowable ranges. If the error decrease is less than 1% for five consecutive iterations, the system stops early and outputs the error correction parameters. The system then fuses the error correction parameters with the calibrated vector set to form an adjustment parameter set adapted to the current scenario. The fusion rule is as follows: when the compatibility score is greater than or equal to 0.90, the calibrated vector is used as the main parameter, with a correction ratio not exceeding 20% ​​of the original value; when the compatibility score is between 0.85 and 0.90, the calibrated vector is used as the main parameter, and a correction ratio not exceeding 20% ​​of the original value is added.When the value is between 90 and 90, the error correction parameter is the primary parameter, with a calibration ratio not exceeding 10% of the original value. All fused parameters are truncated within the threshold range to ensure they do not exceed the limits. The system performs real-time data stream processing and generates equipment operation control commands based on the adjusted parameter group and current dynamic environmental variables. The data stream processing runs in 1-second increments, sequentially executing four steps: data cleaning, limit detection, action mapping, and command issuance. Data cleaning removes readings that are zero or have obvious abnormal jumps and fills them with the most recent valid values. Limit detection compares temperature, humidity, and target range boundaries, as well as power and allowable upper limits. Action mapping transforms limit events into specific executable fine-tuning actions, including lowering or raising the set temperature by no more than 1 degree Celsius, adjusting the power percentage by no more than 0.05, adjusting the fan speed by no more than level 1, and forcibly meeting a minimum continuous operating time of no less than 10 minutes. Command issuance follows a coordinated timing sequence for devices that need to be started or stopped simultaneously. A time offset of at least 1 minute is added to suppress instantaneous power peaks. After the command is executed, the system continuously acquires real-time operating status data according to the equipment operation control commands and writes it to the dynamic environment database. Stored fields include timestamp, temperature, humidity, operating mode of each device, set temperature, operating power, fan speed, start / stop time, and energy consumption per unit time. The system calculates and updates adaptability parameters within each 30-minute sliding window. Adaptability parameters consist of three components: the percentage of time within the target range, energy consumption per unit time, and control stability. The percentage of time within the target range must be at least 90%, the energy consumption per unit time must not exceed the historical median of 5, and control stability is considered satisfactory when the temperature and humidity fluctuation amplitude does not exceed 50% of the target range width. The updated adaptability parameters are synchronously fed back to the weighting factor and deviation monitoring to form a closed loop, ensuring that the optimization goals and calibration directions remain consistent in subsequent windows and continuously meet user needs and energy efficiency constraints.

[0032] S4 includes acquiring the adjustment parameter set, combining it with equipment coordination requirements, and using real-time data stream processing to generate an initial temperature control command set; based on the initial temperature control command set, running a heat flux distribution simulation to calculate the heat flux distribution matrix during equipment operation; using the heat flux distribution matrix, iteratively optimizing the control cycle through time series analysis to generate control cycle parameters; based on the control cycle parameters and real-time operating status data, generating a demand response curve; if the peak value of the demand response curve deviates from a preset threshold beyond a standard, then a gradient descent algorithm is used to optimize the curve parameters to obtain an optimized response curve; based on the optimized response curve and environmental variable analysis, generating a temperature gradient mapping; if the deviation of the temperature gradient mapping from the energy efficiency optimization standard exceeds a preset threshold, then the fine-tuning parameter set is adjusted to generate a verified parameter scheme.

[0033] In step S4 of this embodiment, the adjustment parameter set output in step S3 and the clearly defined equipment coordination requirements are used as input. First, a temperature control initial instruction set is generated at a 1-second granularity using real-time data stream processing. The instruction set provides the operating mode, set temperature, fan speed level, power ratio, start time, and stop time for each device. According to the coordination requirements, the number of devices allowed to start in parallel at the same time is limited to no more than 2, and the start interval between two adjacent high-power devices is no less than 1 minute. The power ratio, set temperature, and fan speed level are derived from the adjustment parameter set, and the start and stop times are determined according to the fixed order of "high load first, low load second, main unit first, terminal unit second" in the coordination requirements. The system then runs a heat flux distribution simulation to calculate the heat flux distribution matrix during equipment operation. The calculation process divides the controlled space into 1-meter-sided grid cells in both horizontal and vertical directions, iterating in 60-second time steps. Each cell within the grid is updated at each time step based on three types of deterministic inputs: First, the heat or cooling output from each device to the cell, calculated as an equivalent value per minute by multiplying the power percentage by the device's rated power and then evenly distributed to the corresponding cells in the device's exhaust or heat exchange coverage area. Second, the heat exchange between adjacent cells, differentially distributed according to fixed thermal conductivity and convective heat transfer coefficients. These coefficients are stable constants determined through offline calibration with similar historical cases before the project goes live and do not change during online operation. Third, the heat exchange boundary with the outside environment, represented by the differential heat exchange between the fixed heat transfer coefficients of the exterior walls, windows, floor, and ceiling and the external ambient temperature and humidity. These coefficients are also derived from offline calibration and fixed during online operation. The system sums these three types of inputs at each time step and then updates the temperature of each cell, iterating until a candidate control cycle is covered. The system performs time series analysis and iteratively optimizes the control cycle based on the heat flux distribution matrix. The time series analysis constructs segmented characteristics of growth, stability and decline segments based on the temperature of each grid cell over time. The system evaluates each candidate cycle in an integer minute range of 5 to 30 minutes with a step size of 1 minute. For each candidate cycle, three deterministic indicators are calculated: the time required to reach the target temperature range, the percentage of stable duration within the target range and the corresponding energy consumption per unit time. These indicators are weighted by weights of 0.5, 0.3 and 0.2 respectively to obtain a cycle score. The cycle with the lowest score is taken as the control cycle parameter, and the equipment start-up and shutdown sequence, duty cycle and load switching time under this cycle are recorded as periodic control elements. The system synthesizes the control cycle parameters and real-time operating status data into a demand response curve. The curve is plotted with time on the horizontal axis and the weighted sum of the absolute values ​​of temperature deviation, humidity deviation, and energy consumption per unit time on the vertical axis, with fixed weights of 0.5, 0.3, and 0.2. The data on the vertical axis is smoothed using a 1-minute sliding window. The peak value of the curve is used to measure the response strength of the current periodic control to user demand.The system sets an upper limit of 10 for the deviation of the curve peak value. The deviation is calculated as the percentage deviation of the peak value relative to the target level, and the absolute value is taken. When the deviation is greater than or equal to 10, the gradient descent algorithm is activated to optimize the curve parameters. The curve parameters include start-up advance, stop delay, power ramp-up rate, power fall-down rate, and duty cycle. The learning step size is 0.01, and the maximum number of iterations is 200. In each iteration, the above parameters are finely adjusted by no more than 2 of the original value in the direction of deviation decrease. If the deviation decrease is less than 1 for 5 consecutive iterations, the system stops early. The optimization continues until the peak deviation of the curve is reduced to less than or equal to 5, resulting in the optimized response curve and a set of updated periodic control elements. Based on the optimized response curve and environmental variable analysis, the system generates a temperature gradient mapping. The temperature gradient mapping is based on grid cells and gives the direction and amplitude of temperature gradients from low temperature to high temperature and from low humidity to high humidity within the space. During the generation process, the temperature and humidity differences between adjacent cells are statistically analyzed at the same time step and aggregated into a directional amplitude map of the overall indoor distribution using a fixed window. At the same time, the local extreme value regions caused by the equipment air outlets and heat exchange surfaces and their diffusion range are marked. The system uses clear energy efficiency optimization standards to test the temperature gradient mapping. The standards consist of three hard indicators: the area ratio within the target temperature and humidity range is not less than 80%, the energy consumption per unit time is not higher than the baseline by 5, and the area exceeding the limit caused by extreme high or low temperatures is not higher than 10% of the total area. If the deviation between the temperature gradient mapping and any of the above indicators exceeds a preset threshold, it is determined that the energy efficiency optimization standards are not met. The deviation threshold is fixed at 5 and is determined by the optimal comprehensive score of user demand satisfaction rate, energy consumption and spatial uniformity in a grid search of 1,000 historical samples in the range of 3 to 10. When a condition is not met, the system enters the adjustment process of the fine-tuning parameter group. The fine-tuning parameter group includes four categories: fine-tuning of set temperature, fine-tuning of fan speed, fine-tuning of power ratio, and fine-tuning of control cycle. The specific rules are as follows: when a hot spot exists, the set temperature of the relevant equipment is reduced by 1, the fan speed is increased by 1, and the power ratio is reduced by 0.05. When a cold spot exists, the set temperature is increased by 1, the fan speed is reduced by 1, and the power ratio is increased by 0.05. The duty cycle is increased or decreased by no more than 10 based on the duration of the hot or cold spot. All fine-tuning must meet the constraint that the minimum continuous running time is not less than 10 minutes and threshold range truncation is performed to prevent exceeding the limit. After applying the fine-tuning, the system regenerates the initial temperature control instruction set, recalculates the heat flux distribution matrix, and repeats the time series analysis and demand response curve optimization until the temperature gradient mapping meets the energy efficiency optimization standard or reaches the upper limit of a maximum of 3 rounds of fine-tuning. At the termination, the parameter set that meets the standard is determined as the verified parameter scheme.All parameters involved in the above process have been provided with their sources and determination methods: the grid edge length and time step are used for the spatial and temporal discretization of the heat flux distribution matrix and are fixed at 1 meter and 60 seconds, respectively; the weights, thresholds, and step sizes are determined through offline statistical verification of 1,000 cases before going live and remain fixed values ​​in the online phase; the power ratio, set temperature, fan speed level, start and stop times are derived from the adjustment parameter group and equipment coordination requirements, and are deterministically updated with clear magnitude and sequence in the optimization and fine-tuning steps; the values ​​of the three hard indicators of energy efficiency optimization standards and the deviation threshold are derived from the optimal comprehensive score of historical samples and serve as fixed acceptance criteria; the learning step size, maximum number of iterations, and early stopping condition of gradient descent are fixed control parameters to ensure convergence speed and stability. Through the above step-by-step, single-value, and reproducible calculation and judgment process, the system generates and outputs a verified parameter scheme, ensuring that the temperature control results are consistent with the energy efficiency optimization standards and meet the equipment coordination requirements and user needs.

[0034] S5 includes obtaining an initial collaborative control model from the validated parameter scheme, generating a multi-device collaborative instruction set through real-time data stream processing; calculating an energy consumption distribution matrix using the multi-device collaborative instruction set and extracting energy consumption distribution characteristics using time series analysis; constructing a collaborative feedback loop based on the energy consumption distribution characteristics and iteratively updating the control instruction set; if the deviation between the iterated control instruction set and the preset regulation stability threshold exceeds the standard, then using a gradient descent algorithm to optimize the instruction set parameters to obtain an optimized control instruction set; simulating boundary conditions using the optimized control instruction set to generate a boundary condition response matrix; analyzing real-time decision-making efficiency based on the boundary condition response matrix to determine an improved temperature control strategy; and updating the collaborative control model using the improved temperature control strategy to generate the final multi-device collaborative operation scheme.

[0035] In step S5 of this embodiment, the system uses the verified parameter scheme obtained in step S4 as the fixed input of the initial collaborative control model, and generates a multi-device collaborative instruction set in the order of acquisition, cleaning, synthesis, and distribution in real-time data stream processing with a 1-second granularity. Acquisition involves reading temperature, humidity, current operating power, and start / stop status of each device every second. Cleaning involves removing data that is zero or has unreasonable jumps and filling in single-point missing data with the previous valid value. Synthesis involves combining each device's operating mode, set temperature, fan speed, power ratio, start time, stop time, and minimum continuous operating time of no less than 10 minutes, as well as the two constraints of staggered start-up of similar high-power devices of no less than 1 minute, into executable instructions. The system issues commands sequentially, recording timestamps for subsequent comparisons. Based on this, the system calculates an energy consumption distribution matrix within a 60-second time step and a 30-minute evaluation window. The matrix is ​​structured with rows corresponding to time steps, columns corresponding to the energy consumption of each device, and a column representing the total energy consumption. The cell value is the product of the device's operating power and the time step duration, converted to kilowatt-hours. The operating power is obtained by multiplying the device's rated power and its power percentage, and is further corrected using a fixed conversion factor based on the wind speed setting. This conversion factor is a stable constant determined through offline calibration of similar historical cases before deployment. The system performs time series analysis on the energy consumption distribution matrix to extract energy consumption distribution features. The feature set includes peak energy consumption, peak-to-average power ratio, the average absolute value of the energy consumption difference between adjacent time steps, energy consumption variance within a control cycle, cycle consistency index, and energy consumption integral within the evaluation window. The cycle consistency index is a dimensionless value obtained by dividing the average absolute value of the energy consumption difference between each time step within a control cycle and the corresponding time step in the previous cycle by the average energy consumption of that cycle. The control cycle length is derived from the control cycle parameter determined in step S4. The system constructs a collaborative feedback mechanism based on the energy consumption distribution features. The system iteratively updates the instruction set via a feedback loop. The feedback error signal consists of two superimposed parts: one part is the deviation between the target energy consumption pattern and the current energy consumption pattern, defined as a low-peak, low-fluctuation pattern under the premise of meeting the target temperature and humidity range; the other part is the execution-level deviation, i.e., the consistency difference between instruction issuance and device response. The system calculates a single scalar error every 60 seconds and maps it to a deterministic micro-adjustment. The adjustment rules are: the power ratio moves towards reducing the peak value in a step of 0.05; the start-up time is staggered in a step of 1 minute towards reducing the simultaneous start-up direction; and the duty cycle converges in a direction that reduces the cycle consistency index by an amplitude not exceeding 10 from the original value. All adjustments satisfy the constraints that the minimum continuous running time is not less than 10 minutes and the start-up stagger is not less than 1 minute, and threshold truncation is applied to parameters that exceed the limits. After each round of feedback, the system calculates the regulation stability index and compares it with the preset regulation stability threshold. The regulation stability index is obtained by weighted summation of three dimensionless indicators: the total energy consumption variation coefficient, the instruction oscillation rate, and the execution success rate gap, with weights of 0.4, 0.4, and 0, respectively.2. The total energy consumption variation coefficient is the ratio of the standard deviation to the mean of the total energy consumption sequence within a control cycle. The instruction oscillation rate is the ratio of the number of modified instruction entries to the total number of instruction entries in one iteration. The execution success rate gap is 1 minus the device acknowledgment success rate. The preset control stability threshold is fixed at 0.10. This threshold is determined optimally by performing a grid search on 1,000 historical samples within the range of 0.05 to 0.15 and considering the comprehensive score of user demand satisfaction rate, energy consumption per unit time, and convergence rounds. When the control stability index is greater than or equal to 0.10, the system initiates gradient descent to optimize the instruction set parameters. The optimization parameters include power ratio, start time offset, stop time offset, and duty cycle. The learning step size is 0.01, and the maximum number of iterations is [not specified]. The early stopping condition is 200, where the stability index decreases by less than 1 for five consecutive iterations, the adjustment range of each parameter in a single iteration does not exceed 2 of its allowable range, and the minimum continuous running time and peak-shifting constraints are always met. After obtaining the optimized control instruction set, the system performs boundary condition simulation and generates a boundary condition response matrix. The boundary conditions include five scenarios: high temperature extreme, low temperature extreme, high humidity extreme, low humidity extreme, and baseline. The high temperature and low temperature extremes are respectively increasing or decreasing the current outdoor measured temperature by 5 degrees Celsius, and the high humidity and low humidity extremes are respectively increasing or decreasing the current outdoor measured humidity by 10%. The system records the total energy consumption, temperature deviation, humidity deviation, and cycle time every minute under each scenario. The consistency index forms the response matrix; the system analyzes real-time decision-making efficiency based on the boundary condition response matrix and determines the improved temperature control strategy. The real-time decision-making efficiency score is obtained by weighting three dimensionless indicators: decision delay ratio, energy saving improvement rate, and instruction change intensity, with weights of 0.3, 0.5, and 0.2 respectively. The decision delay ratio is the ratio of the time required to generate the next round of instruction sets to the length of the control cycle, and is required to be no higher than 0.10. The energy saving improvement rate is the reduction ratio of total energy consumption per unit time relative to the baseline scheme, and is required to be no lower than 0.05. The instruction change intensity is the ratio of the number of modified entries between two rounds of instruction sets to the total number of entries, and is required to be no higher than 0.20. The system enumerates combinations of different power ratio allocations and periodic elements and selects... The combination that simultaneously satisfies all three thresholds and achieves the highest score under all boundary conditions is used as the improved temperature control strategy. Finally, the system updates the collaborative control model with the improved temperature control strategy, clarifying the operating mode, set temperature, wind speed level, power ratio, start and stop times, and duty cycle of each device. It also solidifies two constraints: a minimum continuous operating time of no less than 10 minutes and a start-up staggered operation time of no less than 1 minute. This forms the final multi-device collaborative operation scheme. The sources and determination methods of the above parameters are as follows: wind speed conversion factor, device rated power, and energy consumption conversion rules are calibrated and fixed as constants using historical similar cases before going live; the time step of 60 seconds and the evaluation window of 30 minutes are fixed after online and offline load balancing calculations; and the power ratio step is 0.The parameters 0.05 and 1 minute start-up time step are fixed values ​​after balancing convergence speed and stability on 1,000 samples. The stability threshold of 0.10 and boundary condition perturbation amplitudes of 5 degrees Celsius and 10% represent the optimal point for offline grid search. The gradient descent learning step size of 0.01, maximum number of iterations of 200, and early stopping condition of 5 are fixed control parameters to ensure stable convergence.

[0036] S6 includes: collecting real-time environmental temperature control data through environmental sensor data streams; generating a smooth temperature control data stream using data smoothing processing to obtain a continuous environmental state sequence; if the continuous environmental state sequence deviates from the preset temperature control threshold, retrieving similar cases matching the environmental conditions from the historical case library to obtain a matching case set; extracting the equipment power allocation sequence from the matching case set, and fusing the current equipment power allocation using a weighted average algorithm to obtain an optimized power allocation scheme; constructing a heat flux distribution model based on the optimized power allocation scheme and the collaborative time series to generate a heat flux distribution matrix; processing the heat flux distribution matrix using time series analysis to extract the heat flux change trend to obtain a heat flux trend sequence; iteratively adjusting the periodic parameters using the heat flux trend sequence; if the deviation between the iterative parameters and the preset stable threshold exceeds the range, optimizing the parameters using a gradient descent algorithm to obtain a stable control parameter set; updating the collaborative time series based on the stable control parameter set to generate the final temperature control operation scheme.

[0037] In step S6 of this embodiment, the closed-loop processing of detection, retrieval, fusion, modeling, analysis, optimization, and output is completed according to a determined order and fixed parameters, with environmental sensor data stream as the sole input source and historical case library and collaborative time series as references. Specifically: First, data acquisition and smoothing processing: the system continuously reads temperature and humidity data with a sampling granularity of 1 second, and performs a sliding average of each moment and the previous 59 seconds of 60 sampling points to obtain smoothed temperature and smoothed humidity, forming a continuous environmental state sequence arranged in ascending order of time; the smoothing window length of 60 is determined based on the grid search results of 1,000 sets of historical scenes in the range of 30 to 90, and is fixed by comprehensively scoring the three indicators of deviation detection rate, noise suppression ratio and convergence time. The sampling granularity of 1 second is a fixed value to meet the second-level control response. Second, for control deviation determination, the system sets preset temperature control thresholds for continuous environmental state sequences. The temperature threshold is 1 degree Celsius, and the humidity threshold is 5 percent, with weights configured as 0.6 for temperature and 0.4 for humidity. The control deviation value is calculated every second. The calculation method is to multiply the contribution of the smoothed temperature ratio to the temperature threshold by the temperature weight, and add the contribution of the smoothed humidity ratio to the humidity threshold by the humidity weight to obtain a single scalar. When the scalar is greater than or equal to 1, a control deviation is determined and a similar case search is triggered. The values ​​of the temperature threshold and humidity threshold are fixed by selecting the optimal scoring point from 1,000 historical samples in a grid search of temperatures from 0.5 to 2 degrees Celsius and humidity from 3 to 10 percent, with the constraints of a false alarm rate not exceeding 0.05, a false negative rate not exceeding 0.05, and an average convergence cycle not exceeding 10. Third, similar case retrieval: The system filters cases in the historical case database based on the current smoothed temperature and smoothed humidity as conditions. Cases with a temperature difference of no more than 1 degree Celsius and a humidity difference of no more than 5% with the current environment are selected to form a matching case set. If there are more than 10 cases that meet the conditions, they are sorted by the sum of temperature and humidity differences from smallest to largest and the top 10 are selected. If there are fewer than 3 cases, the threshold is relaxed to a temperature of 2 degrees Celsius and a humidity of 10% until the number is no less than 3. The upper and lower limits of this number are in the range of 5 to 20. After offline evaluation, they are fixed at 10 and 3 to balance the representativeness of the sample and the computational cost. Fourth, power allocation fusion: The system extracts the power allocation sequence of each device from the matching case set and performs two-level weighted fusion to generate an optimized power allocation scheme: The first level is the weighting within the sample, specifically, the sum of the temperature and humidity differences of each case is normalized to the sample weight, and the smaller the difference, the greater the weight; The second level is the fusion weighting of history and the current, the weight of the device power allocation at the current moment is fixed at 0.6, the weight of the historical fusion result is fixed at 0.4, and the sum of the two is 1. This ratio is determined by the joint optimality of historical playback experiments with an energy consumption reduction ratio of not less than 0.05 and a comfort deviation time ratio of not more than 0.10.Fifth, heat flux distribution modeling: The system constructs a heat flux distribution model and generates a heat flux distribution matrix using the optimized power allocation scheme and collaborative time series as inputs. Spatial discretization uses a 1-meter side grid with a time step of 60 seconds. At each time step, the system calculates the sum of three types of deterministic heat terms for each grid cell and converts them into temperature changes to update the matrix. The three types of heat terms are: sensible heat injected according to the optimized power allocation, heat exchange between adjacent grids calculated using fixed thermal conductivity and convection coefficients, and heat exchange with the outer boundary calculated using a fixed heat transfer coefficient. These coefficients are fixed after offline calibration using similar historical cases before going online. Sixth, heat flux time series analysis: The system performs statistical analysis on the heat flux distribution matrix at 60-second time steps. The median of the temperature increment of all grids at each time step is taken as the global rate of change and smoothed by a 1-minute sliding motion to obtain the heat flux trend sequence. At the same time, the start and end times and durations of the rising, falling, and plateau segments of the trend are recorded for subsequent periodic parameter iterations. The statistical caliber is determined by offline comparison of spatial uniformity error and sensitivity to outliers to suppress local extreme value interference. Seventh, the system iterates the cycle parameters, which include duty cycle, startup order, and adjacent start-stop interval. The system updates these parameters round by round according to the following rules: When the duration of the upward trend is greater than or equal to 5 minutes and the absolute value of the global rate of change exceeds half of the target bandwidth, the duty cycle of high-load devices is reduced by no more than 10, and devices that start simultaneously are staggered by a time offset of no less than 1 minute; when the duration of the downward trend is greater than or equal to 5 minutes and the absolute value of the global rate of change exceeds half of the target bandwidth, the duty cycle of low-load devices is increased by no more than 10, and their startup time is advanced by no more than 1 minute; the minimum value of the adjacent start-stop interval is always no less than 10 minutes to prevent frequent start-stops; the values ​​of the above three thresholds, namely 5 minutes, 50% and 1 minute and 10 minutes, are fixed by grid search of 1,000 samples to achieve the conditions that the number of convergence rounds does not exceed 10 and the energy consumption variance is no higher than 0.8 times the baseline.Eighth, stability assessment and gradient descent optimization: After each iteration, the system calculates a stability index, which is the weighted average of temperature and humidity deviations within a control cycle plus the normalized value of energy consumption per unit time. The weights are 0.5 for temperature, 0.3 for humidity, and 0.2 for energy consumption. The preset stability threshold is fixed at 0.10. The value is derived from the optimal point of three constraints: a grid search satisfaction rate of no less than 0.90 in the range of 0.05 to 0.15, energy consumption per unit time not exceeding 5% of the median value of similar devices, and instruction change intensity not exceeding 0.20. When the stability index is greater than or equal to 0.10, gradient descent optimization is triggered. The duty cycle, start time offset, and adjacent start-stop interval are the optimized parameters. The learning step size is 0.01, the maximum number of iterations is 200, and the early stopping condition is that the stability index decreases by less than 1 for 5 consecutive iterations, and the adjustment range of each parameter in a single iteration does not exceed 2% of its allowable range and the minimum continuous running time is not less than 10 minutes and the peak start-up stagger of similar devices is not less than 1 minute. Ninth, collaborative time series update and scheme output: The system updates the collaborative time series item by item with a stable control parameter set, generating the start time, stop time and duty cycle of each device in the next control cycle. At the same time, it determines the power distribution trajectory with an optimized power distribution scheme, forming the final temperature control operation scheme and archiving it with timestamp, temperature and humidity deviation, power distribution, duty cycle and stability index as items for subsequent reuse.

[0038] S7 includes real-time monitoring of preliminary operating results through the equipment collaboration system, extracting the changing trends of operating results using time series analysis to obtain an operating trend sequence; if the deviation between the operating trend sequence and the user demand threshold exceeds a preset range, then obtaining historical trend cases similar to the current operating trend sequence from the historical operating database to obtain a historical trend set; extracting equipment operating adjustment parameters based on the historical trend set, and fusing the current operating trend sequence using a support vector machine algorithm to obtain an adjustment parameter set; updating the operating configuration of the equipment collaboration system using the adjustment parameter set to generate an updated temperature control operating scheme and obtain optimized operating results; if the optimized operating results still deviate from the user demand threshold, then using a gradient descent algorithm to iteratively optimize the adjustment parameter set to obtain a stable operating parameter set; and reconfiguring the equipment collaboration system based on the stable operating parameter set to generate the final temperature control execution result.

[0039] In step S7 of this embodiment, the preliminary operating results of the equipment collaboration system are used as the sole input, and the entire process calculation is performed in a fixed order with determined parameters: First, real-time monitoring and trend extraction are performed. The monitoring indicators are space temperature, space humidity, equipment power output, equipment start / stop status, and user demand completion rate. The sampling granularity is 1 minute. The system smooths each indicator using a moving average with a length of 5 sampling points and obtains the 1-minute change rate by differencing adjacent points. The smoothing window length of 5 and the sampling granularity of 1 minute are selected and fixed by combining the user demand satisfaction rate, energy consumption, and convergence time in a grid search of 1,000 historical scenarios within the range of windows 3 to 10 and granularities of 30 seconds to 2 minutes. The system then calculates the temperature change rate, humidity... The rate of change in temperature, the proportion of power fluctuation, and the rate of change in demand completion rate are arranged in ascending order over time to form an operational trend sequence. The proportion of power fluctuation is calculated by taking the average absolute value of the difference between adjacent readings as the base value and then dividing it by this base value to obtain a dimensionless result. Subsequently, deviation judgment is performed. The user demand threshold consists of three hard indicators: absolute temperature deviation not exceeding 1 degree Celsius, absolute humidity deviation not exceeding 5% percentage, and demand completion rate not lower than 90%. These three thresholds are determined based on a grid search of 1,000 historical samples within the ranges of 0.5 to 2 degrees Celsius, 3 to 10% humidity, and 80 to 95% demand completion rate, with a false alarm rate not exceeding 0.05, a false negative rate not exceeding 0.05, and an average convergence cycle not exceeding 10. The optimal point is determined; the system compares the prediction deviation of the trend sequence extrapolated to the next 5-minute window with the aforementioned thresholds at each 1-minute interval. If any one of these thresholds is exceeded, the system is deemed to have exceeded the preset range and enters the historical trend retrieval process. The admission criteria for historical trend retrieval are: temperature change rate difference not exceeding 20%, humidity change rate difference not exceeding 20%, power fluctuation ratio difference not exceeding 20%, and demand completion rate change rate difference not exceeding 10%. The difference is expressed as the percentage of the absolute value of the difference between the mean of the current trend sequence in the most recent 5 minutes and the mean of the corresponding duration of the historical trend segment, divided by the historical mean. Segments that meet all four conditions form a historical trend set. If there are more than 20 candidate segments, the top 20 are selected by sorting them in ascending order of the sum of the four differences. If there are fewer than 5, the 20% threshold for each of the three items will be increased to 30% and the 10% threshold will be increased to 15% until the number is no less than 5. The system extracts equipment operation adjustment parameters segment by segment from the historical trend set. The parameters include the set temperature correction amount, power distribution correction ratio, start-stop time offset amount, and wind speed adjustment amount. The current operation trend sequence is fused using the support vector machine algorithm to obtain the adjustment parameter set for the current moment. The specific process is as follows: the mean and variance of the temperature change rate, humidity change rate, power fluctuation ratio, and demand completion rate change rate in the last 5 minutes are normalized and used as features. The above four types of adjustment parameters are used as targets. Support vector regression with radial basis function is used for modeling. The penalty coefficient is fixed at 10 and the kernel width is fixed at 0.5. The loss insensitivity interval is fixed at 0.1. The three hyperparameters are solidified by selecting the optimal score from the grid set through five-fold cross-validation of 1,000 historical segments. The training samples are the historical trend set, the prediction input is the current trend features, and the output is the set of adjustment parameters. The system performs boundary truncation on the output to ensure that the magnitude of the set temperature correction does not exceed 2 degrees Celsius, the magnitude of the power distribution correction ratio does not exceed 0.10, the magnitude of the start-stop time offset does not exceed 2 minutes, and the magnitude of the wind speed adjustment does not exceed 1 level. The system uses the set of adjustment parameters to update the operation configuration of the equipment coordination system, generates the updated temperature control operation plan, and implements it in the next control cycle to obtain the optimized operation result. If there is still a deviation between the optimized operation result and the user's demand threshold, the system performs gradient descent iterative optimization on the set of adjustment parameters. The optimization objective is to minimize the weighted sum of temperature deviation, humidity deviation, and demand completion rate gap, with the three weights fixed. The weights are set as follows: temperature 0.5, humidity 0.3, and demand completion rate 0.2. These weights are optimally determined in offline playback based on a comprehensive score of comfort and energy consumption. The learning step size is fixed at 0.01, and the maximum number of iterations is fixed at 200. Early stopping is achieved when the target decrease is less than 1 for five consecutive iterations. The adjustment of each parameter in a single iteration does not exceed 2% of its allowable range, and the same boundary truncation is performed after each update to ensure physical feasibility. After gradient descent is completed, a stable operating parameter set is obtained. Based on this, the system reconfigures the equipment coordination system and generates the final temperature control execution result. The final result requires that within a continuous 30-minute sliding window, the absolute value of the temperature deviation is not higher than 1 degree Celsius, the absolute value of the humidity deviation is not higher than 5%, the demand completion rate is not lower than 90%, and the energy consumption per unit time is not higher than 5% of the baseline before the update. If any indicator fails to meet the standard, the gradient descent process is repeated with the stable operating parameter set as the new initial value, but a maximum of three additional rounds are added to ensure project executability.

[0040] like Figure 2As shown, a multi-device temperature control collaborative adjustment system is also provided to implement the steps of the multi-device temperature control collaborative adjustment method. The system includes a data acquisition module, which collects multi-dimensional variable data of the current temperature control scenario, including temperature, humidity, device status, and user needs, and calculates the matching value with cases in a historical experience database using the cosine similarity method to obtain a set of similar historical cases; a parameter extraction module, which extracts the device collaborative configuration parameters of the case with the highest matching value in the set of similar historical cases as an initial benchmark, and integrates device power allocation, collaborative time sequence, parameter threshold range, and historical load mode to determine the initial parameter adjustment basis; and a parameter optimization module, which optimizes the initial benchmark parameters using a genetic algorithm if the difference between the initial parameter adjustment basis and the current dynamic environmental variables exceeds a preset threshold, incorporating configuration compatibility verification, benchmark weight factors, adjustment basis vectors, and initial error calibration to obtain an adjustment parameter set adapted to the current scenario. The simulation and verification module, after obtaining the adjusted parameter set, simulates the temperature control process for equipment collaboration needs, integrates simulated heat flow distribution and control cycle iteration, as well as demand response curves and temperature gradient mapping, and determines whether the simulation results meet the energy efficiency optimization standards to obtain the verified parameter scheme; the model update module updates the multi-equipment collaborative control model with the verified parameter scheme, combines process energy consumption estimation and collaborative feedback loop, as well as control stability verification and simulated boundary conditions, to determine the temperature control strategy after improving real-time decision-making efficiency; the strategy monitoring module monitors environmental changes through the temperature control strategy, and if control deviation occurs, it re-obtains similar cases from historical experience to supplement and adjust, integrates equipment power allocation and collaborative time sequence, as well as simulated heat flow distribution and control cycle iteration, to obtain stable parameter output; the execution module deploys the stable parameter output to the actual equipment collaborative system, determines whether user needs are met, and obtains the final temperature control execution result.

[0041] The system operates in a closed loop following a modular data flow sequence: The data acquisition module generates multi-dimensional data frames containing temperature, humidity, device on / off status, operating mode, fan speed, set temperature, and user target from temperature and humidity sensors and device feedback at a second-level granularity. After time alignment and normalization, it calculates the cosine similarity between the current scene and historical case entries, outputting a set of similar cases containing case number, matching value, and key parameter index. The parameter extraction module receives this set, selects the first case in descending order of matching value, reads its collaborative configuration parameter group (including power allocation ratio, start / stop sequence, minimum start / stop interval, allowable set temperature and power range, and historical load mode curve), uses it as the initial benchmark, and aligns it item by item with the current device list to generate a preliminary parameter adjustment basis containing "device—power—timing—threshold—load mode". The parameter optimization module continuously compares the initial parameter adjustment baseline with the current dynamic environmental variables. When the difference exceeds a fixed threshold, a genetic algorithm search is triggered. The search minimizes energy consumption, temperature and humidity deviations, and fluctuations. Iteratively, configuration compatibility verification (boundary and temporal peak-shifting checks), weight factor adjustment (temperature, humidity, and energy consumption focus), basic vector construction, and initial error calibration are performed. Upon convergence, the module outputs an adjustment parameter set that satisfies physical and temporal constraints. The simulation verification module uses the adjustment parameter set as input to generate a heat flux distribution matrix on a discrete spatial grid and a fixed time step. It performs multiple rounds of control cycle iterations, combining cycle length and duty cycle, to obtain the demand response curve and temperature gradient mapping. This is then compared item by item with energy efficiency optimization standards. If all standards are met, a verified parameter scheme is formed. The model update module writes this scheme into the collaborative control model, recalculates energy consumption distribution and stability indicators, and verifies decision delay and command change intensity under boundary conditions, producing an improved temperature control strategy. During strategy execution, the strategy monitoring module continuously monitors the environmental and equipment status. When a control deviation reaches a preset threshold, it triggers rapid compensation based on historical similar cases, reusing power allocation and coordination timing, and combining heat flow and periodic iteration to form stable parameter output. The execution module then generates equipment-side instruction entries, issuing the mode, set temperature, fan speed, and start / stop times for each device. Subsequently, it monitors the user demand fulfillment rate and records energy consumption and stability results, forming the final temperature control execution result and writing it back to the database to update the similar case index.

[0042] This system uses a standardized "collection-extraction-optimization-verification-update-monitoring-execution" chain to couple historical experience with the real-time environment in a closed loop, achieving full automation from initial parameter acquisition to verification, online calibration, and deployment. It rapidly identifies highly relevant experiences using cosine similarity, ensures parameter convergence under physical and temporal constraints through genetic algorithms and compatibility checks, performs energy efficiency and stability checks before deployment using heat flux and periodic iteration, and achieves online deviation correction through strategy monitoring. Ultimately, it synergistically improves temperature control accuracy, energy efficiency, and response speed, and possesses reusability and scalability across devices and scenarios.

[0043] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-device temperature control and coordinated regulation method, characterized in that, include: S1. By collecting multi-dimensional variable data of the current temperature control scenario, including temperature, humidity, equipment status and user needs, the cosine similarity method is used to calculate the matching value with cases in the historical experience database to obtain a set of similar historical cases; S2. Based on the case with the highest matching value in the set of similar historical cases, extract its equipment collaborative configuration parameters as the initial benchmark, integrate the equipment power allocation and collaborative timing sequence, as well as the parameter threshold range and historical load mode, to determine the initial parameter adjustment basis; S3. If the difference between the initial parameter adjustment base and the current dynamic environmental variables exceeds a preset threshold, the initial baseline parameters are optimized through a genetic algorithm, incorporating configuration compatibility verification and baseline weight factors, as well as adjusting the base vector and initial error calibration, to obtain an adjustment parameter set adapted to the current scenario. S4. After obtaining the adjustment parameter set, simulate the temperature control process according to the equipment coordination requirements, integrate the simulated heat flow distribution and control cycle iteration, as well as the demand response curve and temperature gradient mapping, and determine whether the simulation results meet the energy efficiency optimization standards to obtain the verified parameter scheme. S5. Update the multi-device collaborative control model using the verified parameter scheme, and combine process energy consumption estimation, collaborative feedback loop, regulation stability test and simulated boundary conditions to determine the temperature control strategy after improving real-time decision efficiency. S6. Monitor environmental changes through temperature control strategies. If control deviations occur, supplement and adjust by retrieving similar cases from historical experience, and integrate equipment power distribution and coordinated timing sequences with simulated heat flow distribution and control cycle iteration to obtain stable parameter output. S7. Based on the stable parameter output, deploy it to the actual equipment collaborative system, determine whether the user's needs are met, and obtain the final temperature control execution result.

2. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S1 includes: By collecting temperature data, humidity data, and equipment status in the temperature control scenario in real time through sensors, and combining them with user needs, a multidimensional variable dataset is formed. A preprocessing method is used to standardize the multidimensional variable dataset to obtain a normalized dataset. The matching degree value between the normalized dataset and each case in the historical experience database is calculated using the cosine similarity algorithm to obtain a set of matching degree values; If the matching degree value exceeds the preset threshold, the corresponding case is extracted from the historical experience database to generate a preliminary set of similar cases; The initial set of similar cases is sorted according to user needs to obtain an optimized set of similar cases; By analyzing the device status and historical experience in the optimized set of similar cases, recommended parameter configurations for temperature control scenarios are generated. The recommended parameter configuration is used to adjust the equipment in the current temperature control scenario to obtain the real-time optimized equipment operating status.

3. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S2 includes: The case with the highest matching value is obtained from the set of similar cases. The device collaboration configuration parameters contained in the case with the highest matching value are extracted to determine the initial baseline configuration. Based on the initial baseline configuration, analyze the device power allocation and collaborative timing sequence, calculate the load balancing coefficient, and obtain the optimized device allocation scheme. By adopting the optimized equipment allocation scheme and combining it with parameter threshold ranges, dynamic adjustment rules for equipment operating status are generated. If the deviation between the dynamic adjustment rules and the historical load patterns exceeds a preset threshold, relevant load characteristics are extracted from historical experience data, and the adjustment rules are updated. Based on the updated adjustment rules, generate real-time equipment collaborative control instructions and determine equipment operating parameters; By combining real-time equipment operating parameters with historical experience data, parameter optimization results are generated to obtain the final equipment control configuration; Operational status data is extracted from the final equipment control configuration, stored in the historical experience database, and the set of similar cases is updated.

4. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S3 includes: If the deviation between the initial parameters and the dynamic environment exceeds a preset threshold, the deviation value is obtained through deviation calculation, and the initial benchmark is iteratively optimized using a genetic algorithm to obtain the optimized benchmark parameter set. By configuring compatibility verification, the adaptability of the optimized baseline parameter set to the current scenario is analyzed, and the compatibility verification results are generated by adjusting the weight factors. Based on the compatibility verification results, the weighting factors are integrated, and a linear regression model is used to calibrate the basic vectors to obtain the calibrated vector set. If the deviation between the calibrated vector set and the initial error exceeds a preset threshold, the initial error is corrected using the gradient descent algorithm to generate error correction parameters. Based on the error correction parameters, the calibrated vector set is fused to generate an adjustment parameter set adapted to the current scenario; By adjusting the parameter group and combining it with the dynamic environmental variables of the current scenario, real-time data stream processing is used to generate equipment operation control commands; Based on the equipment operation control commands, real-time operating status data is acquired, stored in the dynamic environment database, and the adaptability parameters for the current scenario are updated.

5. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S4 includes: The set of adjustment parameters is obtained, and combined with the equipment coordination requirements, real-time data stream processing is used to generate the initial set of temperature control instructions. Based on the initial temperature control instruction set, run the heat flow distribution simulation and calculate the heat flow distribution matrix during equipment operation; By using the heat flux distribution matrix and time series analysis, the control cycle is iteratively optimized to generate control cycle parameters; Based on the control cycle parameters and real-time operating status data, a demand response curve is generated. If the peak value of the demand response curve deviates from the preset threshold by more than the standard, the gradient descent algorithm is used to optimize the curve parameters to obtain the optimized response curve. A temperature gradient mapping is generated by combining the optimized response curve with environmental variable analysis. If the deviation between the temperature gradient mapping and the energy efficiency optimization standard exceeds a preset threshold, the parameter group is fine-tuned to generate a verified parameter scheme.

6. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S5 includes: The initial collaborative control model is obtained from the verified parameter scheme, and a multi-device collaborative instruction set is generated through real-time data stream processing. The energy consumption distribution matrix is ​​calculated by multi-device collaborative instruction set, and the energy consumption distribution characteristics are extracted by time series analysis. A collaborative feedback loop is constructed based on the energy consumption distribution characteristics, and the control instruction set is iteratively updated. If the deviation between the iterative control instruction set and the preset regulation stability threshold exceeds the standard, the gradient descent algorithm is used to optimize the instruction set parameters to obtain the optimized control instruction set. The boundary conditions are simulated using the optimized control instruction set, and the boundary condition response matrix is ​​generated. Based on the analysis of the boundary condition response matrix, the real-time decision-making efficiency is determined, and the improved temperature control strategy is identified. The collaborative control model is updated by improving the temperature control strategy, and the final multi-device collaborative operation scheme is generated.

7. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S6 includes: Real-time environmental temperature control data is collected by environmental sensor data streams, and smooth temperature control data streams are generated by data smoothing processing to obtain a continuous environmental state sequence. If a continuous environmental state sequence deviates from the preset temperature control threshold, then similar cases matching the environmental conditions are retrieved from the historical case database to obtain a set of matching cases. By extracting the device power allocation sequence from the matching case set, and using a weighted average algorithm to fuse the current device power allocation, an optimized power allocation scheme is obtained; Based on the optimized power allocation scheme and the cooperative time series, a heat flux distribution model is constructed, and a heat flux distribution matrix is ​​generated. Time series analysis was used to process the heat flux distribution matrix, extract the heat flux variation trend, and obtain the heat flux trend sequence; The periodic parameters are iteratively adjusted by the heat flow trend sequence. If the deviation between the iterative parameters and the preset stable threshold exceeds the range, the gradient descent algorithm is used to optimize the parameters to obtain a stable control parameter set. The final temperature control operation plan is generated by updating the coordinated time series based on the stable control parameter set.

8. The multi-device temperature control and coordinated adjustment method according to claim 1, characterized in that: S7 includes: The preliminary operating results are monitored in real time through the equipment collaboration system, and the changing trend of the operating results is extracted by time series analysis to obtain the operating trend sequence. If the deviation between the running trend sequence and the user demand threshold exceeds the preset range, then historical trend cases similar to the current running trend sequence are obtained from the historical running database to obtain a historical trend set; Based on the historical trend set, the equipment operation adjustment parameters are extracted, and the current operation trend sequence is fused using the support vector machine algorithm to obtain the adjustment parameter set.

9. A multi-device temperature control and coordinated adjustment method according to claim 8, characterized in that: The S7 also includes: By adjusting the parameter set to update the operating configuration of the equipment coordination system, an updated temperature control operating scheme is generated, resulting in optimized operating results. If the optimized results still deviate from the user's required threshold, the gradient descent algorithm is used to iteratively optimize the set of adjustment parameters to obtain a stable set of operating parameters. The equipment coordination system is reconfigured based on the stable operating parameter set to generate the final temperature control execution result.

10. A multi-device temperature control and coordination system, used to implement the steps of the multi-device temperature control and coordination method according to any one of claims 1-9, characterized in that, The system includes: The data acquisition module collects multi-dimensional variable data of the current temperature control scenario, including temperature, humidity, equipment status, and user needs. It then uses the cosine similarity method to calculate the matching value with cases in the historical experience database, thus obtaining a set of similar historical cases. The parameter extraction module extracts the device collaborative configuration parameters of the case with the highest matching value in the set of similar historical cases as an initial benchmark. It integrates the device power allocation and collaborative timing sequence, as well as the parameter threshold range and historical load mode to determine the initial parameter adjustment basis. The parameter optimization module optimizes the initial baseline parameters by using a genetic algorithm if the difference between the initial parameter adjustment base and the current dynamic environmental variables exceeds a preset threshold. This optimization incorporates configuration compatibility verification and baseline weight factors, as well as adjustments to the base vector and initial error calibration, to obtain a set of adjustment parameters adapted to the current scenario. The simulation verification module, after obtaining the adjustment parameter set, simulates the temperature control process according to the equipment coordination requirements, integrates the simulated heat flow distribution and control cycle iteration, as well as the demand response curve and temperature gradient mapping, and judges whether the simulation results meet the energy efficiency optimization standards to obtain the verified parameter scheme. The model update module updates the multi-device collaborative control model using the validated parameter scheme. It combines process energy consumption estimation, collaborative feedback loop, regulation stability test, and simulated boundary conditions to determine the temperature control strategy after improving real-time decision-making efficiency. The strategy monitoring module monitors environmental changes through temperature control strategies. If control deviations occur, it retrieves similar cases from historical experience to supplement and adjust the system. It integrates equipment power allocation and coordinated timing sequences with simulated heat flow distribution and iterative control cycles to obtain stable parameter outputs. The execution module outputs stable parameters and deploys them to the actual equipment collaborative system to determine whether user needs are met and obtain the final temperature control execution result.

Images