Greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic cooperation

Abstract

The invention provides a greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination, and relates to the technical field of agricultural facility environment control. The method comprises the following steps: acquiring temperature, air humidity, wind speed and airflow distribution environment data inside and outside a greenhouse in real time; the environment state is quantified through a dynamic evaporation efficiency coefficient DEEC and an airflow distribution uniformity index ADUI; the flow of a water pump, the rotating speed of a fan and the opening degree of a ventilation system are dynamically adjusted in combination with an improved PID control algorithm, and the weight of an integral term is dynamically adjusted along with the ADUI to enhance the response efficiency; a time sequence prediction model based on LSTM is adopted to fuse historical and real-time data, a parameter correction amount is predicted, and the adaptation capability of the system to a new environment is improved in combination with a transfer learning algorithm; system hardware comprises a low-power-consumption wireless sensor, a frequency conversion actuator and an embedded control terminal. Through combination of multi-parameter perception and cooperative control, accurate temperature and humidity control and energy efficiency optimization are realized, resource consumption is reduced, and automation and intelligence levels of environment adjustment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural facility environmental control, and in particular to a greenhouse intelligent temperature control automatic adjustment optimization method based on dynamic coordination of multiple parameters. Background Art

[0002] Traditional greenhouse cooling systems widely use evaporative cooling technology based on wet curtains (pads) and fans (fans). The principle is to absorb heat through water evaporation to reduce air temperature. However, existing technologies have significant limitations in practical applications. On the one hand, most systems rely on a single environmental parameter (such as temperature or humidity) for control and lack the ability to dynamically and collaboratively respond to multi-dimensional environmental changes. This leads to insufficient temperature and humidity control precision, prone to localized overcooling or overheating, and affects the uniformity of crop growth. On the other hand, the system has weak adaptability to complex environmental conditions (such as seasonal changes, extreme weather, and different crop types), often requiring frequent manual adjustment of operating parameters, which not only increases management costs but also makes it difficult to achieve energy conservation and consumption reduction goals. In addition, traditional methods lack effective dynamic adjustment mechanisms for water resource utilization efficiency and airflow distribution optimization, and often waste resources due to over-reliance on fixed thresholds or empirical parameters.

[0003] In the existing technology, some improvement schemes have attempted to introduce automated control algorithms (such as PID control) or simple feedback mechanisms, but their control logic is still limited to the independent adjustment of a single device, and has failed to achieve multi-dimensional coordinated optimization of wet curtains, fans and ventilation devices. For example, the adjustment of the wet curtain water pump flow is often based only on temperature deviation, while ignoring the dynamic coupling relationship between humidity, wind speed and airflow distribution; the setting of the fan speed also lacks a quantitative assessment of the uniformity of air distribution, resulting in difficulty in balancing cooling effect and energy consumption. At the same time, when traditional systems are deployed in new greenhouses or different climate zones, they require a lot of time for manual debugging and parameter calibration, and lack the ability to quickly adapt, which limits the promotion and application of the technology.

[0004] To address these issues, an intelligent temperature control method that deeply integrates multiple environmental parameters and possesses adaptive optimization capabilities is urgently needed. An ideal solution would dynamically and collaboratively analyze evaporation efficiency, airflow distribution, and environmental trends to achieve coordinated control of the wet curtain, fan, and ventilation system. This approach would also incorporate predictive models and transfer learning techniques to enhance the system's adaptability and deployment efficiency in diverse scenarios. Such technological breakthroughs would not only significantly improve temperature control accuracy and resource utilization efficiency, but also provide more reliable and sustainable environmental control solutions for modern smart agriculture. Summary of the Invention

[0005] In order to solve the technical problems in the existing technology of insufficient dynamic coordinated response capability to multi-dimensional environmental changes, low temperature and humidity control precision leading to uneven local environment, weak adaptability to complex environments and reliance on manual debugging, low resource utilization efficiency, lack of coordinated optimization of independent equipment adjustment, lack of quantitative evaluation of airflow distribution, and poor adaptability of new greenhouse deployment, the present invention provides a greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination.

[0006] The technical solutions provided by the present invention are as follows:

[0007] The present invention provides a greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination, comprising:

[0008] S1. Real-time collection of greenhouse environmental parameters, including outdoor temperature T out , indoor temperature T in , wet curtain surface humidity H pad , wind speed V air and air distribution uniformity data;

[0009] S2. Based on the data from step S1, calculate the dynamic evaporation efficiency coefficient (DEEC) and the airflow distribution uniformity index (ADUI). DEEC is used to characterize the dynamic response of the wet curtain evaporation efficiency to changes in environmental parameters; ADUI is used to quantify the distribution uniformity of the fan-driven cold air in the greenhouse.

[0010] S3, according to the real-time values of DEEC and ADUI, dynamically adjust the wet curtain water pump flow Q through the collaborative optimization algorithm water , fan speed R fan And roof ventilation opening O vent , so that the greenhouse temperature and humidity reach the preset target range;

[0011] S4. Through the iterative feedback mechanism, historical data and real-time data are combined to correct and optimize parameters to achieve adaptive adjustment of the temperature control system.

[0012] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0013] (1) In the present invention, through the dual-parameter collaborative analysis of the dynamic evaporation efficiency coefficient (DEEC) and the airflow distribution uniformity index (ADUI), the system can perceive the multi-dimensional changes in the greenhouse environment in real time. Combined with the improved PID algorithm and the LSTM prediction model, it dynamically adjusts the operating parameters of the wet curtain, fan, and ventilation device. This solution breaks through the limitations of the single parameter control of traditional temperature control systems, significantly improves the accuracy of temperature and humidity control, effectively avoids local overcooling or overheating, ensures the uniformity and stability of the crop growth environment, and reduces the loss of equipment caused by frequent start-up and shutdown due to environmental fluctuations.

[0014] (2) In the present invention, the system uses LSTM neural networks and transfer learning technology. Through the fusion analysis of historical data and real-time data, it can automatically correct the control strategy and predict future environmental trends, and can quickly adapt to the complex needs of different seasons, crop types and extreme weather conditions. When a new greenhouse is deployed, the pre-trained model and transfer learning algorithm can complete environmental adaptation in a very short time, reducing manual debugging costs. This adaptability not only improves the robustness of the system, but also expands its universality in greenhouses of different sizes and different climate zones, providing a flexible solution for the intelligent upgrade of modern agricultural facilities.

[0015] (3) In the present invention, by optimizing the coordinated control of the wet curtain evaporation efficiency and airflow distribution, the system significantly reduces the consumption of water resources and electricity while ensuring the cooling effect. For example, in combination with the ASHRAE standard, the water pump flow is dynamically limited to avoid excessive water use caused by ineffective evaporation; using stepless speed regulation fans and variable frequency water pumps, the power output is adjusted according to real-time demand to reduce energy waste. In addition, the combination of low-power wireless sensor networks and efficient control algorithms further reduces the overall operating costs of the system. This technology not only conforms to the development direction of green agriculture, but also provides greenhouse operators with an economically sustainable temperature control management solution. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0017] Figure 1 A flowchart of a greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic collaboration provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0019] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0020] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0021] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0022] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0023] Reference Manual Figure 1 , which shows a flow chart of a greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic collaboration provided by an embodiment of the present invention.

[0024] The embodiment of the present invention provides a greenhouse intelligent temperature control automatic adjustment optimization method based on dynamic coordination of multiple parameters. The processing flow may include the following steps:

[0025] S1. Real-time collection of greenhouse environmental parameters, including outdoor temperature T out , indoor temperature T in , wet curtain surface humidity H pad , wind speed V air and air distribution uniformity data.

[0026] Understandably, traditional greenhouse temperature control systems often suffer from inaccurate control due to single or delayed data collection. For example, relying solely on temperature monitoring may overlook the coupled effects of humidity and airflow. To this end, the system deploys a distributed sensor network covering key areas inside and outside the greenhouse, acquiring real-time data on temperature, wet curtain surface humidity, wind speed, and wind speed in the air distribution ducts. The temperature sensor updates data every 30 seconds, a capacitive humidity sensor is installed in the middle of the wet curtain to capture the core humidity state, and anemometers are distributed at the fan outlet and inside the greenhouse. Combined with embedded sensors in eight ducts, comprehensive airflow dynamics are monitored. All data is transmitted to the central controller in real time via a low-power wireless module to ensure the timeliness and integrity of the input information. For example, if the wind speed in a duct drops abnormally, the system can immediately identify local airflow problems, laying the foundation for subsequent control.

[0027] S2. Based on the data from step S1, calculate the dynamic evaporation efficiency coefficient DEEC and the airflow distribution uniformity index ADUI, where DEEC is used to characterize the dynamic response capability of the wet curtain evaporation efficiency to changes in environmental parameters, and ADUI is used to quantify the distribution uniformity of the fan-driven cold air in the greenhouse.

[0028] It is understandable that a single parameter is difficult to fully reflect the complex dynamics of the wet curtain evaporation efficiency and air distribution. For example, relying solely on temperature difference may lead to inaccurate water pump flow regulation. The system comprehensively evaluates the evaporation efficiency of the wet curtain through the dynamic evaporation efficiency coefficient (DEEC). Its calculation integrates the indoor and outdoor temperature difference, humidity difference and wind speed changes, and corrects the weight coefficient in combination with the wet curtain material characteristics to ensure that the index is more in line with the actual working conditions. At the same time, the airflow distribution uniformity index (ADUI) quantifies the uniformity of cold air diffusion based on the discrete degree and average value of the duct wind speed, and dynamically adjusts the calculation period through the sliding window method. For example, the window length is shortened when there is a sudden wind speed fluctuation, and the response sensitivity to abnormal distribution is improved. These two indicators provide a quantitative basis for the coordinated control of multiple devices, avoiding waste of resources or local environmental imbalance caused by empirical operations.

[0029] S3, according to the real-time values of DEEC and ADUI, dynamically adjust the wet curtain water pump flow Q through the collaborative optimization algorithm water , fan speed R fan And roof ventilation opening O vent , so that the greenhouse temperature and humidity reach the preset target range.

[0030] It is understandable that independently adjusting the wet curtain, fan or ventilation device can easily lead to low system efficiency. For example, excessively increasing the water pump flow may cause water waste, while ignoring the fan speed may cause uneven airflow distribution. The system adjusts the equipment parameters based on the real-time values of DEEC and ADUI: when the evaporation efficiency is high, the water pump flow is increased first to enhance the cooling effect, and the fan speed is dynamically adjusted in combination with the uniformity of airflow distribution. If the distribution uniformity is good, the speed is increased to enhance air circulation; otherwise, the speed is reduced and the ventilation opening is adjusted to avoid local overcooling. The improved control algorithm further incorporates airflow distribution indicators into the adjustment weight. For example, when the uniformity is poor, the radical adjustment of the fan speed is suppressed to ensure smooth convergence of temperature and humidity changes. This multi-dimensional collaborative strategy significantly reduces energy and resource consumption while improving cooling efficiency.

[0031] S4. Through the iterative feedback mechanism, historical data and real-time data are combined to correct and optimize parameters to achieve adaptive adjustment of the temperature control system.

[0032] It is understandable that static control parameters are difficult to adapt to long-term environmental changes or new scenario requirements. For example, seasonal changes or differences in crop types may cause the original model to fail. The system uses a time series prediction model to analyze historical and real-time data to predict the direction of future control parameter corrections. For example, it combines the environmental fluctuation trend of the past 24 hours to predict the water pump flow demand in the next stage. When a new greenhouse is deployed, the pre-trained model is quickly adapted through transfer learning technology, and only short-term data is needed to fine-tune key parameters, significantly shortening the debugging cycle. The system continuously monitors the control effect. If the temperature and humidity deviations continue to exceed the threshold, it automatically triggers model retraining and dynamically updates the control logic. This closed-loop optimization mechanism not only enhances the system's adaptability to different environments, but also reduces the frequency of manual intervention through continuous learning, providing a guarantee for long-term stable operation.

[0033] In one possible implementation, the calculation formula for the dynamic evaporation efficiency coefficient DEEC is:

[0034]

[0035] Where ΔT = T out -T in , represents the temperature difference between indoor and outdoor, ΔH=H pad -H air , represents the humidity difference between the wet curtain surface and the air, V min is the minimum effective wind speed threshold of the wind turbine, α and β are weight coefficients, which are calibrated through experiments, where α=0.6·(1-e -0.1·ΔT ), β=0.4·(1+0.2·V air ).

[0036] It should be noted that the calculation of the dynamic evaporation efficiency coefficient DEEC comprehensively evaluates the real-time evaporation efficiency of the wet curtain by integrating the indoor and outdoor temperature difference, humidity difference and wind speed changes. The weight coefficients α and β are calibrated through experiments, and their values are dynamically adjusted with environmental parameters. For example, when the indoor and outdoor temperature difference increases, the α value increases according to the law of the exponential function to strengthen the effect of the temperature difference on the evaporation efficiency; while the β value is linearly positively correlated with the wind speed to ensure accurate representation of the evaporation rate of the wet curtain surface at high wind speeds. In addition, the porosity of the wet curtain material is introduced into the calculation process, and the weight coefficient is modified to adapt to the actual evaporation characteristics of different materials to avoid control deviations caused by material differences.

[0037] In a possible implementation, the calculation formula for the air flow distribution uniformity index ADUI is:

[0038]

[0039] Among them, σ v V is the standard deviation of wind speed in the eight air distribution ducts inside the greenhouse. avg is the average wind speed of the 8 pipes.

[0040] It should be noted that the airflow distribution uniformity index (ADUI) dynamically reflects the uniformity of cold air diffusion by quantifying the ratio of the discrete degree of duct wind speed to the average wind speed. The standard deviation is calculated using a sliding window method, and the default window length is a fixed period. However, when the ADUI falls below the set threshold, the window is automatically shortened to enhance the responsiveness to sudden wind speed fluctuations (such as fan failure or external strong wind interference). For example, when a duct suddenly drops in wind speed due to blockage, shortening the window can quickly capture abnormal data, triggering fan speed adjustment or ventilation opening optimization to avoid temperature imbalances in local areas.

[0041] In a possible implementation, S3 further includes:

[0042] The collaborative optimization algorithm adopts an improved PID control algorithm, and its integral term weight is dynamically adjusted according to ADUI. The specific formula is:

[0043] I t =K i ·(1+0.5·ADUI)·∫e(t)dt

[0044] Among them, K i is the preset integral coefficient, and e(t) is the temperature and humidity deviation value.

[0045] It should be noted that the improved PID control algorithm achieves accuracy and stability in temperature and humidity regulation by dynamically associating the airflow distribution uniformity index ADUI with the integral term weight. When the ADUI is high, the integral term weight increases proportionally, accelerating the convergence of temperature and humidity deviations; conversely, when the ADUI is low, the integral effect is suppressed to avoid overshoot or oscillation caused by uneven airflow. For example, during the hot midday period, if the airflow distribution is well-uniform, the system can quickly increase the fan speed to the target value, shortening the cooling response time; if the distribution is uneven, a gradual adjustment is adopted to ensure smooth environmental changes.

[0046] In a possible implementation, S3 further includes:

[0047] Adjust the wet curtain water pump flow Q water When quoting the evaporative cooling efficiency formula in the ASHRAE standard And limit Q according to the η value water to avoid excessive water consumption.

[0048] It's important to note that the system dynamically limits the maximum flow rate of the wet curtain pump by referencing the evaporative cooling efficiency formula in the ASHRAE standard, preventing water waste caused by ineffective evaporation. When the calculated evaporation efficiency η exceeds a preset threshold, the system automatically limits the pump flow rate to a fixed percentage of the design value. This mechanism is particularly critical in high-temperature, low-humidity environments, maintaining cooling effectiveness while preventing humidity saturation or equipment damage caused by excessive water spraying.

[0049] In a possible implementation, S4 further includes:

[0050] The iterative feedback mechanism adopts a time series prediction model based on LSTM neural network. The input parameters include DEEC, ADUI and historical temperature and humidity data, and the output is the parameter correction for the next cycle optimization.

[0051] It should be noted that the LSTM neural network-based time series prediction model analyzes historical DEEC, ADUI, and environmental data to predict future control parameter corrections. The model inputs include temperature, humidity, wind speed, and control records from past fixed periods, and outputs adjustments for parameters such as pump flow and fan speed for the next period. Training utilizes multi-scenario datasets (such as seasonal variations and extreme weather), and transfer learning techniques are used to quickly adapt pre-trained models to the new greenhouse. For example, by freezing underlying network parameters and fine-tuning the top-level structure using only short-term data from the new environment, this significantly reduces debugging costs.

[0052] In a possible embodiment, the porosity P of the wet curtain material is introduced into the calculation process of the weight coefficients α and β. pad As a correction factor, the correction formula is:

[0053] α=0.6·(1-e -0.1·ΔT )·(1+0.15·P pad )

[0054] β=0.4·(1+0.2·V air )·(1-0.1·P pad ).

[0055] It should be noted that the porosity of the wet curtain material is introduced as a correction factor in the DEEC weight coefficient calculation to adapt to the evaporation characteristics of different materials (such as cellulose and synthetic fibers). For example, a wet curtain with a higher porosity will increase the weight of the α value, strengthening the impact of temperature difference on evaporation efficiency; while a wet curtain with a lower porosity will reduce the weight of the β value to avoid misjudgment of evaporation rate at high wind speeds. Material parameters are called through a preset database. Users only need to select the wet curtain type, and the system will automatically match the corresponding porosity value to simplify the operation process.

[0056] In one possible implementation, the wind speed standard deviation σ v The calculation adopts the sliding window method with a window length of 10 minutes. The window length is dynamically adjusted according to the ADUI value. When ADUI < 0.8, the window length is shortened to 5 minutes.

[0057] It should be noted that the calculation of wind speed standard deviation uses a dynamic sliding window method, and the window length is adjusted in real time based on the ADUI value. When the ADUI falls below a certain threshold, the window period is shortened from the default period to quickly respond to sudden wind speed fluctuations (such as fan failure or pipe blockage). For example, if the wind speed in a certain pipe drops sharply in a short period of time, shortening the window can immediately identify the anomaly and trigger the fan speed to increase, avoiding localized temperature increases. This mechanism maintains uniform airflow distribution while improving the system's fault tolerance to transient disturbances.

[0058] In a possible implementation, S3 further includes:

[0059] Roof ventilation opening O vent The adjustment strategy is: if DEEC>0.7 and ADUI>0.9, then O ven Increase by 10%; if DEEC < 0.4 or ADUI < 0.7, then O vent 5% reduction.

[0060] It's important to note that the roof ventilation opening adjustment strategy is based on a real-time combined logic of DEEC and ADUI. When both DEEC and ADUI reach preset thresholds, the opening is proportionally increased to enhance air exchange efficiency. When DEEC or ADUI falls below a specific threshold, the opening is reduced to stabilize the internal environment. This strategy, through quantitative thresholds and proportional adjustments, balances cooling efficiency and energy consumption. For example, ventilation is reduced when evaporation efficiency is low to prevent heat backflow, while maintaining a stable airflow distribution.

[0061] In one possible implementation, the training data for the LSTM neural network includes greenhouse operating data under different seasons, crop types, and weather conditions, and is quickly adapted to new greenhouse environments through a transfer learning algorithm.

[0062] It should be noted that the LSTM neural network's training data covers a wide range of crop types, seasonal variations, and extreme weather conditions, ensuring the model's generalization capabilities. Transfer learning technology freezes the underlying network parameters (such as the time series feature extraction layer) of the pre-trained model in new greenhouse deployments, fine-tuning only the parameters of the top, fully connected layer, and adapting the model using short-term operational data from the new greenhouse. For example, in the deployment of a newly built greenhouse, the fine-tuned model significantly reduced prediction error, shortening the cycle required for traditional manual debugging and improving the efficiency of technology promotion.

[0063] In the specific implementation provided in this embodiment, first, the temperature data inside and outside the greenhouse are collected in real time by distributed temperature sensors, including the outdoor temperature T out and indoor temperature T in , updated every 30 seconds to ensure the timeliness of the data. pad It is measured using a high-precision capacitive humidity sensor installed in the middle of the wet curtain to reflect the humidity status of its core area. air Anemometers installed at the fan outlet and at key points within the greenhouse, combined with wind speed sensors embedded in eight air slurves, provide real-time wind speed data for each duct, which is then used to calculate air distribution uniformity. All sensor data is transmitted to a central controller via a LoRa wireless module, ensuring low power consumption and stable long-distance communication.

[0064] In the specific implementation provided in this embodiment, the calculation of the dynamic evaporation efficiency coefficient DEEC is based on the real-time changes of environmental parameters, and its formula is:

[0065]

[0066] Where ΔT represents the temperature difference between indoor and outdoor out -T in , ΔH is the surface humidity of the wet curtain H pad The average humidity inside the greenhouse Hair The difference, V min The minimum effective wind speed threshold of the fan is set to 0.5m / s by default to ensure the basic airflow requirements. The weight coefficients α and β are calibrated through experiments. The specific formula is:

[0067] α=0.6·(1-e -0.1·ΔT )·(1+0.15·P pad )

[0068] β=0.4·(1+0.2·V air )·(1-0.1·P pad )

[0069] Among them, P pad The porosity of the wet curtain material must be obtained from the material parameter table. For example, the porosity of cellulose wet curtains is usually 0.35 to 0.45. DEEC's real-time calculation can dynamically reflect the responsiveness of the wet curtain's evaporation efficiency to changes in temperature, humidity, and wind speed, providing a core basis for subsequent control strategies.

[0070] In the specific implementation provided in this embodiment, the quantification of the air distribution uniformity index ADUI depends on the wind speed data of the air distribution duct, and its calculation formula is:

[0071]

[0072] where σ v is the standard deviation of wind speed in 8 pipes, V avg is the average wind speed. Standard deviation σ v The calculation adopts the sliding window method, and the default window length is 10 minutes. However, if the real-time ADUI value is lower than 0.8 (indicating a high risk of uneven distribution), the window length is automatically shortened to 5 minutes to improve data sensitivity. For example, if the wind speeds of each duct at a certain moment are 1.2m / s, 1.5m / s, 1.1m / s, 1.4m / s, 1.3m / s, 1.6m / s, 1.0m / s, and 1.5m / s, then V avg is 1.325m / s, σ v The ADUI is 0.207, and 1-(0.207 / 1.325)≈0.844, indicating that the airflow distribution is relatively uniform. Real-time monitoring of ADUI can effectively avoid local overcooling or overheating, ensuring a consistent crop growth environment.

[0073] In the specific implementation provided in this embodiment, the core of the collaborative optimization control strategy is to dynamically adjust the wet curtain water pump flow Q water , fan speed R fan And roof ventilation opening O vent For the control of water pump flow, it is necessary to combine the evaporative cooling efficiency formula in the ASHRAE standard

[0074]

[0075] Where T wet is the wet bulb temperature, which is calculated indirectly through the temperature and humidity sensor. When the η value exceeds 0.8, the system automatically limits Q water Do not exceed 80% of the maximum design flow rate to prevent excessive water consumption. For example, if the maximum design flow rate is 100L / min, the water pump flow rate is limited to 80L / min. The fan speed is regulated using an improved PID (Proportional-Integral-Derivative) control algorithm, with the integral term weight dynamically associated with the ADUI. The specific formula is:

[0076] I t =K i ·(1+0.5·ADUI)·∫e(t)dt

[0077] where K i The default value is 0.2, where e(t) is the deviation between the real-time temperature and humidity and the target value. When the ADUI is high (e.g., 0.9), the integral term weight increases to 0.2 × 1.45 = 0.29, thereby speeding up the adjustment process. Conversely, when the ADUI is low, the integral effect is reduced to avoid system oscillation.

[0078] In the specific implementation provided in this embodiment, the roof ventilation opening adjustment strategy is formulated based on the real-time values of DEEC and ADUI. If DEEC>0.7 and ADUI>0.9, it indicates that the evaporation efficiency is high and the air flow distribution is uniform, the system will be 0 vent Increase by 10% to further improve air exchange efficiency; if DEEC < 0.4 or ADUI < 0.7, it is determined that the system has the risk of inefficiency or uneven distribution. ven Reduce by 5% to stabilize the internal environment. For example, if the initial ventilation opening is 50%, it will increase to 55% after an increase and decrease to 47.5% after a decrease. This strategy combines logical judgment with quantitative adjustment to achieve dynamic collaborative optimization of multiple parameters.

[0079] In the specific implementation provided in this embodiment, the adaptive feedback and prediction model adopts a long short-term memory (LSTM) neural network, the input parameters include DEEC, ADUI and the temperature and humidity data of the past 24 hours, and the output is the Q of the next cycle. water 、R fan and O ventCorrection amount. The model training data covers seasonal changes, multiple crops (such as tomatoes and cucumbers) and extreme weather scenarios (such as high temperatures of 40°C and humidity exceeding 90%) to ensure its generalization ability. During training, the mean square error (MSE) is used as the loss function, the optimizer selects Adam, the learning rate is set to 0.001, the batch size is 32, and the training cycle is 200 times. The application of transfer learning technology enables the pre-trained model to be loaded and the underlying network parameters to be frozen when a new greenhouse is deployed, and only the top fully connected layer is fine-tuned through the operating data of the first 7 days of the new greenhouse, which significantly shortens the adaptation time. For example, when the new greenhouse is deployed in summer, the model can be fine-tuned within 3 hours, and the prediction error is controlled within 5%.

[0080] In terms of the system hardware configuration in the specific implementation provided in this embodiment, the sensor network consists of temperature, humidity and wind speed sensors, all of which realize wireless data transmission through LoRa modules. The communication distance can reach 2 kilometers, which is suitable for large greenhouses. The actuators include variable frequency water pumps, stepless speed control fans and roof ventilation devices driven by stepper motors. The central controller uses an embedded industrial computer (CPU quad-core 1.8GHz, memory 4GB) running a real-time Linux system to ensure millisecond response. For example, when T is detected, in When the temperature suddenly rises by 2°C, the controller can calculate the new DEEC and ADUI within 1 second and complete the linkage adjustment of the water pump, fan and ventilation opening within 5 seconds.

[0081] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0082] (1) In the present invention, through the dual-parameter collaborative analysis of the dynamic evaporation efficiency coefficient (DEEC) and the airflow distribution uniformity index (ADUI), the system can perceive the multi-dimensional changes in the greenhouse environment in real time. Combined with the improved PID algorithm and the LSTM prediction model, it dynamically adjusts the operating parameters of the wet curtain, fan, and ventilation device. This solution breaks through the limitations of the single parameter control of traditional temperature control systems, significantly improves the accuracy of temperature and humidity control, effectively avoids local overcooling or overheating, ensures the uniformity and stability of the crop growth environment, and reduces the loss of equipment caused by frequent start-up and shutdown due to environmental fluctuations.

[0083] (2) In the present invention, the system uses LSTM neural networks and transfer learning technology. Through the fusion analysis of historical data and real-time data, it can automatically correct the control strategy and predict future environmental trends, and can quickly adapt to the complex needs of different seasons, crop types and extreme weather conditions. When a new greenhouse is deployed, the pre-trained model and transfer learning algorithm can complete environmental adaptation in a very short time, reducing manual debugging costs. This adaptability not only improves the robustness of the system, but also expands its universality in greenhouses of different sizes and different climate zones, providing a flexible solution for the intelligent upgrade of modern agricultural facilities.

[0084] (3) In the present invention, by optimizing the coordinated control of the wet curtain evaporation efficiency and airflow distribution, the system significantly reduces the consumption of water resources and electricity while ensuring the cooling effect. For example, in combination with the ASHRAE standard, the water pump flow is dynamically limited to avoid excessive water use caused by ineffective evaporation; using stepless speed regulation fans and variable frequency water pumps, the power output is adjusted according to real-time demand to reduce energy waste. In addition, the combination of low-power wireless sensor networks and efficient control algorithms further reduces the overall operating costs of the system. This technology not only conforms to the development direction of green agriculture, but also provides greenhouse operators with an economically sustainable temperature control management solution.

[0085] The above content is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0086] There are a few points to note:

[0087] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0088] (2) For the sake of clarity, the thickness of layers or regions in the drawings used to describe the embodiments of the present invention are exaggerated or reduced, that is, these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region, or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or intervening elements may be present.

[0089] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to form new embodiments.

[0090] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination, characterized in that: include: S1. Real-time collection of greenhouse environmental parameters, including outdoor temperature T out , indoor temperature T in , wet curtain surface humidity H pad , wind speed V air and air distribution uniformity data; S2. Based on the data from step S1, calculate the dynamic evaporation efficiency coefficient DEEC and the airflow distribution uniformity index ADUI, where DEEC is used to characterize the dynamic response of the wet curtain evaporation efficiency to changes in environmental parameters; ADUI is used to quantify the distribution uniformity of the fan-driven cold air in the greenhouse; S3, according to the real-time values of DEEC and ADUI, dynamically adjust the wet curtain water pump flow Q through the collaborative optimization algorithm water , fan speed R fan And roof ventilation opening O vent , so that the greenhouse temperature and humidity reach the preset target range; S4. Through the iterative feedback mechanism, historical data and real-time data are combined to correct and optimize parameters to achieve adaptive adjustment of the temperature control system.

2. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 1 is characterized in that: include: The calculation formula of the dynamic evaporation efficiency coefficient DEEC is: Where ΔT = T out -T in , represents the temperature difference between indoor and outdoor, ΔH=H pad -H air , represents the humidity difference between the wet curtain surface and the air, V min is the minimum effective wind speed threshold of the wind turbine, α and β are weight coefficients, which are calibrated through experiments, where α=0.6·(1-e -0.1·ΔT ), β=0.4·(1+0.2·V air ).

3. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 1 is characterized in that: include: The calculation formula of the air distribution uniformity index ADUI is: Among them, σ v V is the standard deviation of wind speed in the eight air distribution ducts inside the greenhouse. avg is the average wind speed of the 8 pipes.

4. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 1 is characterized in that: Said S3 further comprises: The collaborative optimization algorithm adopts an improved PID control algorithm, and its integral term weight is dynamically adjusted according to ADUI. The specific formula is: I t =K i ·(1+0.5·ADUI)·∫e(t)dt Among them, K i is the preset integral coefficient, and e(t) is the temperature and humidity deviation value.

5. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 1 is characterized in that: Said S3 further comprises: Adjust the wet curtain water pump flow Q water When quoting the evaporative cooling efficiency formula in the ASHRAE standard And limit Q according to the η value water to avoid excessive water consumption.

6. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 1 is characterized in that: Said S4 further comprises: The iterative feedback mechanism adopts a time series prediction model based on LSTM neural network. The input parameters include DEEC, ADUI and historical temperature and humidity data, and the output is the correction value of the optimization parameters for the next cycle.

7. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 2 is characterized in that: include: In the calculation process of the weight coefficients α and β, the porosity Ppad of the wet curtain material is introduced as a correction factor. The correction formula is: α=0.6·(1-e -0.1·ΔT )·(1+0.15·P pad ) β=0.4·(1+0.2·V air )·(1-0.1·P pad )。 8. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 3 is characterized in that: include: The wind speed standard deviation σ v The calculation adopts the sliding window method with a window length of 10 minutes. The window length is dynamically adjusted according to the ADUI value. When ADUI < 0.8, the window length is shortened to 5 minutes.

9. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 1 is characterized in that: Said S3 further comprises: Roof ventilation opening O vent The adjustment strategy is: if DEEC>0.7 and ADUI>0.9, then O vent Increase by 10%; if DEEC < 0.4 or ADUI < 0.7, then O vent 5% reduction.

10. The greenhouse intelligent temperature control automatic adjustment optimization method based on multi-parameter dynamic coordination according to claim 6 is characterized in that: include: The training data of the LSTM neural network includes greenhouse operation data under different seasons, crop types and weather conditions, and it quickly adapts to new greenhouse environments through a transfer learning algorithm.

Images