Intelligent regulation and control method for enhancing atomizer temperature control based on artificial intelligence

Abstract

The invention discloses an intelligent regulation and control method for enhancing atomizer temperature control based on artificial intelligence, and the method comprises the following steps: 1, collecting temperature data, and carrying out the preprocessing of the temperature data, and generating the temperature data of a unified structure; step 2, calculating test statistics by using an improved Kruscal-Wallis algorithm, and generating a temperature fluctuation range; 3, combining environment change data collected in real time with temperature data through a dynamic time warping method, and calculating a dynamic time warping distance; 4, predicting a temperature change trend through a GRU network according to the temperature fluctuation range and the dynamic time warping distance; 5, setting the temperature according to the temperature change trend, and returning to the step 3 when the temperature exceeds the set safety range; and 6, when the temperature exceeding state exceeds a preset time threshold for more than or equal to 3 times, entering a safety protection mode. Through intelligent prediction and dynamic adjustment, temperature control precision improvement and safety protection of the atomizer are achieved.

Description

Technical Field

[0001] This invention relates to the field of temperature control technology, and in particular to an intelligent regulation method for enhancing the temperature control of an atomizer based on artificial intelligence. Background Art

[0002] With the widespread application of nebulizers in medical, industrial, and scientific research fields, higher requirements have been placed on their temperature control accuracy and safety. As a precision device, the operating temperature of a nebulizer has a crucial impact on the atomization effect and the stability of the device. Traditional nebulizer temperature control systems mainly rely on preset temperature settings and adjust the temperature through simple feedback mechanisms. However, with changes in working environmental conditions, such as fluctuations in humidity and air pressure, traditional temperature control methods often cannot adapt to these changes in real time, resulting in low temperature control accuracy and potentially leading to equipment damage or low work efficiency. Traditional methods do not consider the real-time impact of environmental changes; therefore, the temperature of the nebulizer often fluctuates during actual operation, failing to ensure stable operation within a safe temperature range.

[0003] Furthermore, existing technologies are typically based on static control algorithms, failing to effectively adjust to dynamically changing environmental data. Traditional temperature control systems cannot process data collected from different sensors in real time, resulting in slow system response and an inability to accurately predict and adjust the temperature promptly. For example, in high humidity or low pressure environments, traditional systems cannot accurately adjust temperature control parameters, increasing the risk of the temperature exceeding the set range. Under high loads and prolonged operation, such systems often lead to untimely temperature regulation, resulting in overheating or underheating, severely affecting atomization performance, and even causing irreversible damage to the equipment.

[0004] While some existing technologies employ data acquisition and simple control algorithms for temperature regulation, most rely on fixed setpoints and cannot flexibly adapt to varying operating conditions. Even some sensor-based intelligent control systems focus on single feedback mechanisms, lacking comprehensive integration and effective prediction based on different data sources. Due to the failure to effectively integrate the relationship between environmental and temperature changes, these systems often exhibit lag in temperature regulation, preventing equipment from promptly entering protection mode and resulting in overheating or overcooling.

[0005] Therefore, how to provide an artificial intelligence-based intelligent control method for enhancing atomizer temperature is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose an intelligent temperature control method for enhanced atomizers based on artificial intelligence. This invention utilizes an AI-based dynamic temperature control method, combining temperature fluctuation range and environmental data, and employs an improved Kruskal-Wallis algorithm and GRU network to predict temperature change trends in real time, achieving intelligent adjustment of the temperature control system. Compared to traditional temperature control methods, it can more accurately respond to temperature fluctuations and improve temperature control precision. Automatic activation of safety protection measures in case of overheating ensures stable equipment operation, improving equipment safety, reliability, and service life.

[0007] According to an embodiment of the present invention, an intelligent control method for enhancing the temperature of an atomizer based on artificial intelligence includes the following steps:

[0008] Step 1: Collect temperature data of the atomizer in real time using a temperature sensor, and preprocess temperature data from different sources to generate temperature data with a uniform structure;

[0009] Step 2: Use the improved Kruskal-Wallis algorithm to group the uniformly structured temperature data according to the time series, calculate the rank of each group, and calculate the test statistic based on the sum of the ranks of each group to generate the temperature fluctuation range.

[0010] Step 3: Combine the real-time environmental change data with temperature data using the dynamic time warp method to calculate the dynamic time warp distance;

[0011] Step 4: Based on the temperature fluctuation range and dynamic time warping distance, predict the temperature change trend using a GRU network;

[0012] Step 5: Based on the temperature change trend, set the temperature through the atomizer's temperature control system and monitor the atomizer's operating status in real time. When the temperature exceeds the set safe range, return to Step 3.

[0013] Step 6: When the temperature exceeds the preset time threshold three or more times, enter the safety protection mode.

[0014] Optionally, the temperature data includes atomizer surface temperature, flow rate temperature, ambient temperature, and timestamp, where the timestamp is the acquisition time for each temperature data point.

[0015] Optionally, the preprocessing includes deleting abnormal temperature data points, using Gaussian filtering to remove random noise from the temperature data, filling missing temperature data points with mean interpolation, aligning the collected temperature data from different sources according to timestamps, and normalizing the temperature data using the Min-Max normalization method to generate temperature data with a uniform structure.

[0016] Optionally, the improved Kruskal-Wallis algorithm is used to group the uniformly structured temperature data according to the time series, calculate the rank of each group, and calculate the test statistic based on the sum of the ranks of each group to generate the temperature fluctuation range. Specifically:

[0017] Based on the timestamp, the temperature data with a uniform structure is grouped into time periods, with each group representing the temperature data within the same time period.

[0018] Sort all temperature data within each temperature data group in ascending order of temperature values;

[0019] The improved Kruskal-Wallis algorithm assigns a rank to each sorted temperature data point. If the temperature data points have the same temperature value, then all temperature data points with the same temperature value are assigned the same rank. The rank is the number assigned to each data point after the temperature data is arranged in order.

[0020] For each set of temperature data, sum the ranks of all temperature data points to obtain the total rank of each set of temperature data.

[0021] The improved Kruskal-Wallis algorithm calculates a test statistic based on the sum of the ranks of each temperature data set, and the test statistic is used to detect differences between different temperature data sets.

[0022] The p-value is calculated using the cumulative distribution function of the chi-square distribution, given a test statistic and degrees of freedom, where the degrees of freedom is the number of temperature data sets minus 1, and the p-value represents the significance of the differences between temperature data sets.

[0023] If the p-value is less than the preset significance level of 0.05, the temperature value within the prediction period is calculated by weighted average temperature as the predicted value, and the temperature fluctuation range is calculated by using the standard deviation of the temperature data. The lower limit of the predicted temperature fluctuation range is the predicted value minus the standard deviation, and the upper limit of the predicted range is the predicted value plus the standard deviation.

[0024] If the p-value is greater than or equal to the significance level of 0.05, the temperature fluctuation range is set to the predicted value ± 1℃.

[0025] Optionally, the method of combining real-time collected environmental change data with temperature data using dynamic time warping to calculate the dynamic time warping distance specifically involves:

[0026] The system acquires real-time environmental change data and temperature data through sensors, including humidity, air pressure, and wind speed.

[0027] For temperature data and environmental change data with different sampling frequencies, a linear interpolation method is used to adjust the time interval of the environmental data to be the same as that of the temperature data.

[0028] The dynamic time warp method generates an Euclidean distance matrix by calculating the Euclidean distance between each pair of corresponding time steps in the environmental change data time series and the temperature data time series. Each element in the Euclidean distance matrix represents the Euclidean distance between the temperature data point and the environmental change data point at a certain time step.

[0029] Based on the Euclidean distance matrix, the optimal time alignment path is obtained by dynamically programming from the top left corner to the bottom right corner of the matrix to progressively select the minimum cumulative distance between each pair of data points. The optimal time alignment path represents the best time alignment method between temperature data and environmental change data.

[0030] The dynamic time warping distance is obtained by summing the Euclidean distances along the optimal time alignment path.

[0031] Optionally, the step of predicting the temperature change trend using a GRU network based on the temperature fluctuation range and dynamic time warping distance specifically involves:

[0032] Temperature fluctuation range and dynamic time warp distance are used as time series input data and input into the input layer of the GRU network. The temperature fluctuation range provides the expected range of temperature change, and the dynamic time warp distance provides similarity information between temperature and environmental changes.

[0033] The GRU network performs gating processing on the input temperature fluctuation range and dynamic time warping distance through a gating unit. The gating unit includes an update gate and a reset gate. The gating process includes updating the direction and adjustment magnitude of the temperature prediction according to the value of the Sigmoid activation function at each time step, generating a temperature change trend, which includes the predicted temperature mean and temperature fluctuation range.

[0034] Optionally, the temperature is set according to the temperature change trend through the atomizer's temperature control system, and the atomizer's operating status is monitored in real time. When the temperature exceeds the set safe range, the process returns to step three, specifically:

[0035] Set the target temperature range based on the predicted average temperature and the temperature fluctuation range;

[0036] When the minimum value of the target temperature range is less than the actual temperature of the atomizer, the heating temperature of the atomizer is set to the average value of the target temperature range.

[0037] When the maximum value of the target temperature range is greater than the actual temperature of the atomizer, the temperature of the atomizer cooling device is set to the average value of the target temperature range.

[0038] If the real-time temperature is greater than the upper limit of the target temperature or less than the lower limit of the target temperature, it means that the temperature has exceeded the set safe range. Return to step three for readjustment until the temperature reaches the set safe range.

[0039] Optionally, when the temperature exceeds a preset time threshold, a safety protection mode is entered, specifically as follows:

[0040] The atomizer's real-time temperature is continuously monitored by a temperature sensor. When the temperature exceeds the preset time threshold three or more times, the atomizer enters a safety protection mode.

[0041] The safety protection mode includes suspending equipment operation, activating the backup cooling or heating system, triggering an emergency alarm, and recording alarm data. Suspending equipment operation includes cutting off the power supply, stopping the heating or cooling function, and locking system control. The emergency alarm includes simultaneous flashing indicator lights and a buzzer.

[0042] The alarm data record includes temperature data, the time period exceeding the safe range, the conditions for alarm triggering, the atomizer status, and alarm response time information. The atomizer status includes whether the device is paused or the cooling or heating system is activated.

[0043] The beneficial effects of the present invention are:

[0044] 1. This invention effectively addresses the shortcomings of traditional temperature control systems in adapting to real-time environmental changes and operating conditions by introducing an artificial intelligence-based dynamic temperature control method. Traditional temperature control systems typically rely on preset temperature setpoints and simple feedback mechanisms, lacking real-time response capabilities to environmental changes. This results in significant temperature fluctuations, making it difficult for the system to maintain stable operation within a safe temperature range. In contrast, this invention collects temperature data from various sources in real-time and processes the data using a dynamic time warping method and an improved Kruskal-Wallis algorithm to generate a temperature fluctuation range. This allows for real-time adjustment of the temperature control strategy based on environmental changes, enabling the system to respond more accurately to temperature fluctuations and significantly improving temperature control precision and equipment stability.

[0045] 2. A GRU network is used to predict temperature change trends, and the temperature is set based on the prediction results. This deep learning-based prediction method can accurately predict future temperature change trends based on historical temperature data and real-time environmental data. In this way, the temperature control system can not only understand the current temperature fluctuation range but also anticipate possible future temperature changes, thus making adjustments in advance and avoiding the lag response problem in traditional systems. Compared with traditional control methods based on static setpoints, the dynamic prediction method of this invention significantly improves the accuracy and response speed of temperature regulation.

[0046] 3. This invention automatically activates a safety protection mode when the temperature exceeds the safe range, including suspending equipment operation, activating a backup cooling or heating system, and triggering an emergency alarm. This provides additional safety for the equipment. Traditional temperature control systems often fail to detect temperature deviations from the set range in a timely manner and take appropriate measures, frequently resulting in equipment damage or poor atomization when temperatures are abnormal. This invention, through continuous temperature monitoring and an automatic feedback mechanism when the temperature exceeds the safe range, can quickly take protective measures, ensuring that the equipment is not damaged by overheating or overcooling, significantly improving the safety and reliability of the equipment. Attached Figure Description

[0047] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0048] Figure 1 This is an overall flowchart of an artificial intelligence-based intelligent regulation method for enhancing atomizer temperature control proposed in this invention.

[0049] Figure 2 This is a flowchart illustrating the application of the improved Kruskal-Wallis algorithm in an artificial intelligence-based intelligent control method for enhanced atomizer temperature proposed in this invention.

[0050] Figure 3 This is a flowchart of the dynamic time warp method for an artificial intelligence-based intelligent regulation method for enhanced atomizer temperature control proposed in this invention. Detailed Implementation

[0051] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0052] refer to Figure 1-3 An artificial intelligence-based intelligent control method for enhancing atomizer temperature includes the following steps:

[0053] Step 1: Collect temperature data of the atomizer in real time using a temperature sensor, and preprocess temperature data from different sources to generate temperature data with a uniform structure;

[0054] Step 2: Use the improved Kruskal-Wallis algorithm to group the uniformly structured temperature data according to the time series, calculate the rank of each group, and calculate the test statistic based on the sum of the ranks of each group to generate the temperature fluctuation range.

[0055] Step 3: Combine the real-time environmental change data with temperature data using the dynamic time warp method to calculate the dynamic time warp distance;

[0056] Step 4: Based on the temperature fluctuation range and dynamic time warping distance, predict the temperature change trend using a GRU network;

[0057] Step 5: Based on the temperature change trend, set the temperature through the atomizer's temperature control system and monitor the atomizer's operating status in real time. When the temperature exceeds the set safe range, return to Step 3.

[0058] Step 6: When the temperature exceeds the preset time threshold three or more times, enter the safety protection mode.

[0059] This invention provides an artificial intelligence-based intelligent temperature control method for enhanced atomizers, solving the problem that existing temperature control systems cannot adapt to environmental changes and operating conditions in real time. By collecting temperature data from different sources in real time and using an improved Kruskal-Wallis algorithm to group and process the temperature data, a temperature fluctuation range is generated, thereby achieving accurate temperature prediction. This method dynamically adjusts the temperature control strategy by combining environmental data and temperature fluctuation ranges, enabling the temperature control system to respond to temperature changes and fluctuations in environmental conditions in real time. This significantly improves temperature control accuracy and reduces the inaccuracy or instability caused by fixed setpoints in traditional systems. Simultaneously, the integration of a GRU network for temperature prediction and adjustment enhances the system's intelligence and response speed, improving equipment stability.

[0060] In this embodiment, the temperature data includes atomizer surface temperature, flow rate temperature, ambient temperature, and timestamp, where the timestamp is the time of collection for each temperature data point.

[0061] This invention uses the atomizer's surface temperature, flow rate temperature, ambient temperature, and timestamp as key data inputs to ensure the temperature control system can accurately predict and adjust based on multi-dimensional temperature data. By precisely recording the acquisition time of each temperature data point, the system can track the dynamic process of temperature changes and adjust the temperature control strategy in real time. Temperature data from different sources not only provides more comprehensive temperature information, but also, through the introduction of timestamps, the system can perform time alignment and synchronization of temperature data, ensuring data accuracy and consistency. This approach enables the temperature control system to respond promptly to temperature changes and fluctuations in environmental factors, significantly improving the system's intelligence and real-time response capabilities, thereby ensuring the temperature stability and safety of the atomizer under different operating environments.

[0062] In this embodiment, the preprocessing includes deleting abnormal temperature data points, using Gaussian filtering to remove random noise from the temperature data, filling missing temperature data points with mean interpolation, aligning the collected temperature data from different sources according to timestamps to ensure that different types of data can be analyzed synchronously, and normalizing the temperature data using the Min-Max normalization method to generate temperature data with a uniform structure.

[0063] This invention employs meticulous preprocessing of temperature data, utilizing Gaussian filtering to remove noise, mean interpolation to fill missing data, and time alignment to ensure synchronization of data from different sources. Furthermore, it normalizes the data using the Min-Max standardization method, ensuring the consistency and accuracy of the temperature data. This processed temperature data eliminates random errors, handles missing data, and eliminates scale differences between different data sources, thereby making subsequent algorithms, such as Kruskal-Wallis testing and GRU network processing, more efficient and accurate. By optimizing the data preprocessing process, this invention provides a solid data foundation for high-precision temperature control and temperature trend prediction, effectively improving the overall system performance.

[0064] In this embodiment, the improved Kruskal-Wallis algorithm is used to group the uniformly structured temperature data according to the time series, calculate the rank of each group, and calculate the test statistic based on the sum of the ranks of each group to generate the temperature fluctuation range. Specifically:

[0065] Based on the timestamp, the temperature data with a uniform structure is grouped into time periods, with each group representing the temperature data within the same time period.

[0066] Sort all temperature data within each temperature data group in ascending order of temperature values;

[0067] The improved Kruskal-Wallis algorithm assigns a rank to each sorted temperature data point. If the temperature data points have the same temperature value, then all temperature data points with the same temperature value are assigned the same rank. If there are two temperature data points with the same temperature value, ranked in the 3rd and 4th positions respectively, then their ranks are both (3+4) / 2 = 3.5. The rank is the number assigned to each temperature data point after the temperature data is arranged in order.

[0068] For each set of temperature data, sum the ranks of all temperature data points to obtain the total rank of each set of temperature data.

[0069] The improved Kruskal-Wallis algorithm calculates the test statistic H based on the sum of the ranks of each set of temperature data:

[0070]

[0071] Where N is the total number of temperature data points, k is the number of temperature data sets, and T i Let n be the sum of the ranks of the i-th temperature data set. i The number of temperature data points in the i-th temperature data group;

[0072] The test statistic is used to detect differences between different temperature data sets;

[0073] The p-value is calculated using the cumulative distribution function of the chi-square distribution, given a test statistic and degrees of freedom equal to the number of temperature data sets minus one. The p-value represents the significance of the differences between the temperature data sets.

[0074]

[0075] Where P(X≤H) represents the cumulative probability of the chi-square distribution, df is the degree of freedom, which is equal to k-1, Γ is the standard part of the chi-square distribution, and t is the integral variable;

[0076] If the p-value is less than the preset significance level of 0.05, it indicates that the temperature difference between different temperature data groups is significant. The temperature value within the prediction period is calculated by weighted temperature mean as the predicted value, and the temperature fluctuation range is calculated by using the standard deviation of the temperature data. The lower limit of the predicted temperature fluctuation range is the predicted value minus the standard deviation, and the upper limit of the predicted range is the predicted value plus the standard deviation.

[0077] If the p-value is greater than or equal to the significance level of 0.05, it indicates that there is no significant difference in temperature change between groups, and the temperature fluctuation range is set to the predicted value ±1℃.

[0078] This invention utilizes an improved Kruskal-Wallis algorithm to group, calculate ranks, and perform statistical tests on temperature data with a uniform structure, effectively overcoming the limitations of traditional methods that cannot dynamically adjust temperature control parameters in real time. By grouping time-series data, calculating the rank of each group, and generating temperature fluctuation ranges, this algorithm accurately reflects the fluctuation trends of the temperature data. By introducing the cumulative distribution function of the chi-square distribution to calculate the p-value, this invention can determine the stability of temperature changes based on significant differences. If the temperature change is significant, the system can accurately predict the future temperature fluctuation range, thus ensuring the accuracy of temperature prediction. This method significantly improves the stability of temperature prediction and reduces temperature control errors.

[0079] In this embodiment, the method of combining real-time collected environmental change data with temperature data using dynamic time warping to calculate the dynamic time warping distance specifically involves:

[0080] The system acquires real-time environmental change data and temperature data through sensors, including humidity, air pressure, and wind speed.

[0081] For temperature data and environmental change data with different sampling frequencies, a linear interpolation method is used to adjust the time interval of the environmental data to be the same as that of the temperature data, ensuring that the timestamps of the temperature data and the environmental data are consistent.

[0082] The dynamic time warp method generates an Euclidean distance matrix by calculating the Euclidean distance between each pair of corresponding time steps in the environmental change data time series and the temperature data time series. Each element in the Euclidean distance matrix represents the Euclidean distance between the temperature data point and the environmental change data point at a certain time step.

[0083] Based on the Euclidean distance matrix, the optimal time alignment path is obtained by dynamically programming from the top left corner to the bottom right corner of the matrix to progressively select the minimum cumulative distance between each pair of data points. The optimal time alignment path represents the best time alignment method between temperature data and environmental change data.

[0084] The dynamic time warping distance is obtained by summing the Euclidean distances along the optimal time alignment path.

[0085] This invention combines environmental change data with temperature data using a dynamic time warp method to calculate the dynamic time warp distance, solving the problem that traditional temperature control methods cannot fully consider the relationship between temperature changes and environmental factors. The dynamic time warp method uses an Euclidean distance matrix and a dynamic programming algorithm to accurately find the optimal time alignment between temperature data and environmental change data. Utilizing the dynamic time warp distance, the system can adjust its temperature control strategy in real time to adapt to dynamic changes in temperature and environment, thereby improving the response speed and accuracy of the temperature control system. Through this method, this invention effectively solves the problem of temperature fluctuations caused by changes in environmental factors, enhancing the intelligence and adaptability of the temperature control system.

[0086] In this embodiment, the step of predicting the temperature change trend using a GRU network based on the temperature fluctuation range and the dynamic time warp distance specifically involves:

[0087] Temperature fluctuation range and dynamic time warp distance are used as time series input data and input into the input layer of the GRU network. The temperature fluctuation range provides the expected range of temperature change, and the dynamic time warp distance provides similarity information between temperature and environmental changes.

[0088] The GRU network performs gating processing on the input temperature fluctuation range and dynamic time warping distance through a gating unit. The gating unit includes an update gate and a reset gate. The gating process includes updating the direction and adjustment magnitude of the temperature prediction according to the value of the Sigmoid activation function at each time step, generating a temperature change trend, which includes the predicted temperature mean and temperature fluctuation range.

[0089] This invention utilizes a GRU network to predict temperature change trends based on temperature fluctuation range and dynamic time warping distance, further optimizing the intelligent adjustment capability of the temperature control system. As a deep learning method, the GRU network can effectively process time-series data and capture the long-term dependency between temperature and environmental changes. By inputting the temperature fluctuation range and dynamic time warping distance, the GRU network can accurately predict future temperature change trends and adjust the temperature setpoint of the temperature control system based on the predicted average temperature. This prediction and adjustment mechanism overcomes the shortcomings of traditional temperature control systems that cannot respond promptly to environmental changes and temperature fluctuations, significantly improving the accuracy and real-time performance of the temperature control system.

[0090] In this embodiment, the temperature is set according to the temperature change trend through the atomizer's temperature control system, and the atomizer's operating status is monitored in real time. When the temperature exceeds the set safe range, the process returns to step three, specifically:

[0091] Set the target temperature range based on the predicted average temperature and the temperature fluctuation range;

[0092] When the minimum value of the target temperature range is less than the actual temperature of the atomizer, the heating temperature of the atomizer is set to the average value of the target temperature range.

[0093] When the maximum value of the target temperature range is greater than the actual temperature of the atomizer, the temperature of the atomizer cooling device is set to the average value of the target temperature range, and the cooling flow rate is increased.

[0094] If the real-time temperature is greater than the upper limit of the target temperature or less than the lower limit of the target temperature, it means that the temperature has exceeded the set safe range. Return to step three for readjustment until the temperature reaches the set safe range.

[0095] This invention uses a temperature control system to monitor the atomizer's temperature in real time and detects temperature deviations within the target temperature range, enabling timely detection of temperatures exceeding the set safe range. By returning to step three, the system re-predicts the temperature change trend based on the temperature fluctuation range and dynamic time warping distance, and adjusts the temperature control parameters accordingly. This mechanism effectively solves the problem of traditional temperature control systems' inability to flexibly adapt to temperature fluctuations and ensures that the temperature remains stable within a safe range in rapidly changing environments. The temperature control system can adjust the temperature setting through real-time feedback, thereby improving temperature control accuracy and reducing the risks posed by temperature anomalies to the equipment.

[0096] In this embodiment, the step of entering the safety protection mode when the temperature exceeds a preset time threshold is specifically as follows:

[0097] The atomizer's real-time temperature is continuously monitored by a temperature sensor. When the temperature exceeds the preset time threshold three or more times, the atomizer enters a safety protection mode.

[0098] The safety protection mode includes suspending equipment operation, activating the backup cooling or heating system, triggering an emergency alarm, and recording alarm data. Suspending equipment operation includes cutting off the power supply, stopping the heating or cooling function, and locking system control. The emergency alarm includes simultaneous flashing indicator lights and a buzzer.

[0099] The alarm data record includes temperature data, the time period exceeding the safe range, the conditions for alarm triggering, the atomizer status, and alarm response time information. The atomizer status includes whether the device is paused or the cooling or heating system is activated.

[0100] This invention employs a safety protection mode that automatically triggers protective measures when the temperature continuously exceeds the safe range, avoiding the problem of traditional temperature control systems failing to take timely action in the event of temperature anomalies. By suspending equipment operation, activating backup cooling or heating systems, triggering emergency alarms, and recording alarm data, this invention provides comprehensive safety assurance for the atomizer. Alarm data recording provides important reference for subsequent fault analysis and equipment maintenance, making the system more reliable and transparent. This safety protection mode effectively reduces the risk of equipment damage due to abnormal temperatures, ensuring that the equipment maintains safe and stable operation even in abnormal environments.

[0101] Example 1:

[0102] To verify the feasibility of this invention in practice, it was applied to the intensive care unit (ICU) of a medical institution, where a high-precision nebulizer was used to provide nebulization therapy to patients. The temperature and humidity in this environment often fluctuate significantly, and traditional temperature control systems cannot respond to these changes in real time. Consequently, the nebulizer temperature frequently deviates from the set range, leading to equipment malfunction or reduced treatment effectiveness. Traditional temperature control systems typically rely on preset temperature settings and adjust through simple feedback mechanisms. This approach is ineffective in handling dynamically changing environmental conditions. For example, when the ambient humidity is too high or too low, the temperature control system cannot adjust based on real-time environmental data, resulting in poor temperature control accuracy and failing to meet the high-precision requirements of nebulizers in sensitive applications such as intensive care. To address this problem, the hospital decided to introduce an artificial intelligence-based temperature control system, and the intelligent control method of this invention was applied to the nebulizer temperature control.

[0103] The hospital's temperature control system collects multiple temperature data sources from the nebulizer in real time, including nebulizer surface temperature, flow rate temperature, and ambient temperature. Each data point is timestamped to ensure the system can track and analyze changes in each temperature data point in real time. After data acquisition, the system preprocesses the temperature data from different sources, including removing outliers, using Gaussian filtering to remove noise, and performing mean interpolation on missing data to generate temperature data with a uniform structure. After data preprocessing, the system analyzes the temperature data using an improved Kruskal-Wallis algorithm. This algorithm first groups the data according to the time series and calculates the rank of each group. Through rank calculation and test statistics, the system can determine the temperature fluctuation range, providing a basis for subsequent temperature control decisions. The temperature fluctuation range is calculated based on the sum of the ranks of each group and the calculation of the statistics, and the generated temperature fluctuation range reflects the current temperature change trend of the system.

[0104] Next, the system uses the Dynamic Time Warp (DTW) method to combine environmental change data with temperature data to calculate the dynamic time warp distance. The DTW algorithm calculates the Euclidean distance between temperature data and environmental change data to help the system determine the relationship between temperature and environmental changes. When changes in ambient temperature and humidity have a significant impact on temperature, the DTW algorithm makes more precise adjustments to the temperature changes. Through this method, the system can accurately capture the time alignment between temperature changes and environmental changes and generate the dynamic time warp distance, providing important background data for predicting future temperature trends.

[0105] Based on the previously calculated temperature fluctuation range and dynamic time warping distance, the system predicts temperature change trends using a GRU (Gated Recurrent Unit) network. The GRU network can process time-series data and capture the long-term dependence between temperature and environmental factors. The temperature fluctuation range and dynamic time warping distance, as input features, provide the expected range of temperature changes and the correlation between temperature and environmental factors. The GRU network uses this information to predict future temperature change trends. Through network processing, the system can predict the average temperature and fluctuation range over a future period, thus providing a basis for subsequent temperature control decisions. The prediction results include upper and lower limits for future temperatures, which the temperature control system uses to set the temperature. This process ensures that the temperature control system can identify temperature change trends in advance and make corresponding adjustments.

[0106] After predicting temperature change trends, the temperature control system sets a target temperature based on the average temperature and fluctuation range. When there is a deviation between the target temperature and the actual temperature, the system adjusts accordingly. If the current temperature is below the lower limit of the target temperature, the system activates the heating device; if the temperature is above the upper limit of the target temperature, the system activates the cooling device. To ensure temperature stability, the system also monitors temperature changes in real time. When the temperature exceeds the set safety range, the system triggers a feedback mechanism and returns to the data adjustment step for correction until the temperature returns to the set range.

[0107] When the temperature continuously exceeds the set safe range, the temperature control system will automatically activate the safety protection mode. This mode includes suspending equipment operation, activating the backup cooling or heating system, triggering an emergency alarm, and recording alarm data. When the temperature is abnormal, the system will cut off the power, stop the heating or cooling function, and lock system control to ensure the equipment is not damaged due to the abnormal temperature. Simultaneously, the system will trigger an emergency alarm via flashing indicator lights and a buzzer to alert personnel to respond promptly. Alarm data will include temperature data, the time period exceeding the safe range, the conditions for alarm triggering, and the system status. This data will be recorded and stored to provide a basis for subsequent fault analysis.

[0108] During implementation, the hospital compared the temperature control performance of traditional temperature control systems with that of the method of this invention. The temperature control system data comparison table is shown below:

[0109] Table 1 Comparison of Temperature Control System Data

[0110]

[0111] As can be seen from Table 1, the accuracy of the temperature control system is significantly improved after adopting the method of the present invention, especially in the control of surface temperature and flow temperature, where the temperature control error is reduced by more than 70%. At the same time, the number of over-temperature events within the safe range is also greatly reduced, and the safety of the equipment is significantly guaranteed.

[0112] As demonstrated in this embodiment, the AI-based temperature control method has significant advantages in adapting to real-time environmental changes and temperature fluctuations. Compared to traditional temperature control systems, the method of this invention can not only accurately predict temperature change trends but also adjust the temperature control system in real time, ensuring the temperature stability of the device under various environmental conditions. Simultaneously, the automatic safety protection mechanism effectively prevents equipment damage caused by abnormal temperatures, improving the atomizer's efficiency and safety.

[0113] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent temperature control of an enhanced atomizer based on artificial intelligence, characterized in that, Includes the following steps: Step 1: Collect temperature data of the atomizer in real time using a temperature sensor, and preprocess temperature data from different sources to generate temperature data with a uniform structure; Step 2: Use the improved Kruskal-Wallis algorithm to group the uniformly structured temperature data according to the time series, calculate the rank of each group, and calculate the test statistic based on the sum of the ranks of each group to generate the temperature fluctuation range. Step 3: Combine the real-time environmental change data with temperature data using the dynamic time warp method to calculate the dynamic time warp distance; Step 4: Based on the temperature fluctuation range and dynamic time warping distance, predict the temperature change trend using a GRU network; Step 5: Based on the temperature change trend, set the temperature through the atomizer's temperature control system and monitor the atomizer's operating status in real time. When the temperature exceeds the set safe range, return to Step 3. Step 6: When the temperature exceeds the preset time threshold three or more times, enter the safety protection mode.

2. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, The temperature data includes atomizer surface temperature, flow rate temperature, ambient temperature, and timestamp, where the timestamp is the time of collection for each temperature data point.

3. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, The preprocessing includes deleting abnormal temperature data points, using Gaussian filtering to remove random noise from the temperature data, filling missing temperature data points with mean interpolation, aligning the collected temperature data from different sources according to timestamps, and normalizing the temperature data using the Min-Max normalization method to generate temperature data with a uniform structure.

4. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, The improved Kruskal-Wallis algorithm is used to group the uniformly structured temperature data according to the time series, calculate the rank of each group, and calculate the test statistic based on the sum of the ranks of each group to generate the temperature fluctuation range. Specifically: Based on the timestamp, the temperature data with a uniform structure is grouped into time periods, with each group representing the temperature data within the same time period. Sort all temperature data within each temperature data group in ascending order of temperature values; The improved Kruskal-Wallis algorithm assigns a rank to each sorted temperature data point. If the temperature data points have the same temperature value, then all temperature data points with the same temperature value are assigned the same rank. The rank is the number assigned to each data point after the temperature data is arranged in order. For each set of temperature data, sum the ranks of all temperature data points to obtain the total rank of each set of temperature data. The improved Kruskal-Wallis algorithm calculates a test statistic based on the sum of the ranks of each temperature data set, and the test statistic is used to detect differences between different temperature data sets. The p-value is calculated using the cumulative distribution function of the chi-square distribution, given a test statistic and degrees of freedom, where the degrees of freedom is the number of temperature data sets minus 1, and the p-value represents the significance of the differences between temperature data sets. If the p-value is less than the preset significance level of 0.05, the temperature value within the prediction period is calculated by weighted average temperature as the predicted value, and the temperature fluctuation range is calculated by using the standard deviation of the temperature data. The lower limit of the predicted temperature fluctuation range is the predicted value minus the standard deviation, and the upper limit of the predicted range is the predicted value plus the standard deviation. If the p-value is greater than or equal to the significance level of 0.05, the temperature fluctuation range is set to the predicted value ± 1℃.

5. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, The method of combining real-time collected environmental change data with temperature data using dynamic time warp to calculate the dynamic time warp distance is as follows: The system acquires real-time environmental change data and temperature data through sensors, including humidity, air pressure, and wind speed. For temperature data and environmental change data with different sampling frequencies, a linear interpolation method is used to adjust the time interval of the environmental data to be the same as that of the temperature data. The dynamic time warp method generates an Euclidean distance matrix by calculating the Euclidean distance between each pair of corresponding time steps in the environmental change data time series and the temperature data time series. Each element in the Euclidean distance matrix represents the Euclidean distance between the temperature data point and the environmental change data point at a certain time step. Based on the Euclidean distance matrix, the optimal time alignment path is obtained by dynamically programming from the top left corner to the bottom right corner of the matrix to progressively select the minimum cumulative distance between each pair of data points. The optimal time alignment path represents the best time alignment method between temperature data and environmental change data. The dynamic time warping distance is obtained by summing the Euclidean distances along the optimal time alignment path.

6. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, The method of predicting temperature change trends using a GRU network based on temperature fluctuation range and dynamic time warping distance is as follows: Temperature fluctuation range and dynamic time warp distance are used as time series input data and input into the input layer of the GRU network. The temperature fluctuation range provides the expected range of temperature change, and the dynamic time warp distance provides similarity information between temperature and environmental changes. The GRU network performs gating processing on the input temperature fluctuation range and dynamic time warping distance through a gating unit. The gating unit includes an update gate and a reset gate. The gating process includes updating the direction and adjustment magnitude of the temperature prediction according to the value of the Sigmoid activation function at each time step, generating a temperature change trend, which includes the predicted temperature mean and temperature fluctuation range.

7. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, The temperature is set according to the temperature change trend through the atomizer's temperature control system, and the atomizer's operating status is monitored in real time. When the temperature exceeds the set safe range, the process returns to step three, specifically: Set the target temperature range based on the predicted average temperature and the temperature fluctuation range; When the minimum value of the target temperature range is less than the actual temperature of the atomizer, the heating temperature of the atomizer is set to the average value of the target temperature range. When the maximum value of the target temperature range is greater than the actual temperature of the atomizer, the temperature of the atomizer cooling device is set to the average value of the target temperature range. If the real-time temperature is greater than the upper limit of the target temperature or less than the lower limit of the target temperature, it means that the temperature has exceeded the set safe range. Return to step three for readjustment until the temperature reaches the set safe range.

8. The intelligent control method for enhanced atomizer temperature based on artificial intelligence according to claim 1, characterized in that, When the temperature exceeds a preset time threshold, a safety protection mode is entered, specifically as follows: The atomizer's real-time temperature is continuously monitored by a temperature sensor. When the temperature exceeds the preset time threshold three or more times, the atomizer enters a safety protection mode. The safety protection mode includes suspending equipment operation, activating the backup cooling or heating system, triggering an emergency alarm, and recording alarm data. Suspending equipment operation includes cutting off the power supply, stopping the heating or cooling function, and locking system control. The emergency alarm includes simultaneous flashing indicator lights and a buzzer. The alarm data record includes temperature data, the time period exceeding the safe range, the conditions for alarm triggering, the atomizer status, and alarm response time information. The atomizer status includes whether the device is paused or the cooling or heating system is activated.

Images