A root temperature intelligent control system for hydroponic crops and its implementation method

Abstract

The present invention relates to the field of temperature regulation technology, specifically to an intelligent root temperature control system for hydroponic crops and its implementation method. The system includes: a root oxygen consumption monitoring module, an osmotic pressure status monitoring module, a root metabolic status analysis module, a root temperature parameter reconstruction module, and a nutrient solution circulation linkage module. In the present invention, the root metabolic rhythm state is determined by calculating the gradient difference of the oxygen consumption rate and analyzing the trend. A dual-parameter logical judgment is performed to identify the risk of cell damage based on the amplitude of the osmotic pressure change and the oxygen consumption trend. A gray correlation analysis is performed based on the abnormal trigger temperature and the stable state temperature to select the target root temperature. A control instruction is generated and a PID algorithm is called to adjust the pump speed based on real-time circulating flow rate data. This achieves closed-loop dynamic matching between the root temperature setpoint and the flow rate behavior, enhances the adaptability of the root temperature response, improves the real-time control capability of metabolic imbalance, and effectively improves the stability of the root system and the water and nutrient absorption efficiency in the hydroponic environment.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature regulation, and in particular to an intelligent root temperature control system for hydroponic crops and an implementation method thereof. Background Art

[0002] The field of temperature regulation encompasses precise management of key plant growth parameters within agricultural environmental control. Its core focus is on real-time monitoring of environmental variables through sensor networks and dynamic adjustment of equipment operating status based on feedback mechanisms. This field systematically integrates the technical systems of temperature acquisition devices, control units, and actuators, focusing on addressing the impact of fluctuating environmental parameters on crop physiological activities in facility agriculture. In particular, stable control of root temperature in hydroponic systems directly impacts nutrient solution absorption efficiency and plant metabolic rate. Existing technologies often use fixed threshold control or manual intervention to regulate water temperature, lacking the ability to adapt to dynamic changes in the root microenvironment.

[0003] One such system, an intelligent root temperature control system for hydroponic crops, uses distributed temperature probes to collect real-time rhizosphere water temperature data, builds a multi-node temperature field model based on the structural characteristics of the cultivation container, and triggers a gradient temperature control strategy based on preset threshold intervals. The system specifically includes a flow rate regulator for the circulating water circuit, a heat exchanger power control unit, and a redundant temperature compensation mechanism. It uses a temperature difference prediction method based on time series analysis to drive the actuator linkage, achieves uniform distribution of the root temperature field through fluid dynamics optimization design, and utilizes a hierarchical control architecture to separate the logical processing of the data acquisition layer from the device driver layer.

[0004] Existing technologies mostly rely on temperature field models and time series prediction methods to execute control instructions. Although multi-node control of root temperature is achieved, there is a lag problem in identifying abnormal root metabolism. The fundamental reason is that the control mechanism is usually triggered only based on the trend of water temperature changes itself, and does not introduce a real-time feedback path for the physiological state of the crop. As a result, in the early stages of temperature fluctuations, the system is difficult to capture the functional change signals of the root system in time, resulting in response delays. For example, in the high temperature stage, although the root temperature has not exceeded the set upper limit, the root system has already produced metabolic imbalance, and the system still maintains a slow adjustment state, thereby exacerbating root damage. In addition, the existing technology mostly uses static thresholds to set the temperature range, which cannot adapt to the dynamic changes in the root system. It is easy to cause frequent misadjustments or idle control due to inaccurate set values, which reduces the sensitivity and regulation efficiency of the control system. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, the present invention provides an intelligent root temperature control system for hydroponic crops and its implementation method. The technical solution is as follows:

[0006] In one aspect, a root temperature intelligent control system for hydroponic crops is provided, the system comprising:

[0007] The root oxygen consumption monitoring module collects root oxygen concentration data in real time, calculates the root oxygen consumption rate gradient, determines the trend state of the oxygen consumption gradient based on a multivariable nonlinear coupling model, and transmits the data to the osmotic pressure status monitoring module;

[0008] The osmotic pressure status monitoring module obtains the cell osmotic pressure change data, performs difference calculation based on the oxygen consumption gradient trend state, obtains the dynamic offset of the osmotic pressure change, and transmits it to the root metabolism status analysis module;

[0009] The root metabolic state analysis module, based on the oxygen consumption gradient trend state and the osmotic pressure change dynamic offset, assigns weights by pore connectivity and root depth, inputs a multi-condition state machine model to determine the damage risk level, obtains a metabolic abnormality determination result, and transmits it to the root temperature parameter reconstruction module;

[0010] The root temperature parameter reconstruction module compares the metabolic abnormality trigger temperature with the root system temperature before the abnormal state occurs in the metabolic abnormality state determination result, generates a root system temperature control instruction, and transmits it to the nutrient solution circulation linkage module;

[0011] The nutrient solution circulation linkage module receives the root temperature control instruction, obtains the circulation flow rate data in real time, and calls the PID control algorithm to adjust the circulation pump speed.

[0012] As a further solution of the present invention, the oxygen consumption gradient trend state includes the gradient direction vector, the change frequency characteristics, and the amplitude fluctuation range; the dynamic offset of the osmotic pressure change is the maximum positive deviation value, the maximum negative deviation value, and the offset duration; the metabolic abnormality state judgment result includes the abnormal condition category, the cell risk rating, and the root temperature at the corresponding moment; the root temperature control instruction is the target temperature range, the control mode code, and the instruction effective time; the control result of adjusting the circulation pump speed includes the set speed value, the expected flow rate target, and the PID output parameter.

[0013] As a further solution of the present invention, the root oxygen consumption monitoring module includes:

[0014] Oxygen concentration acquisition submodule: This module monitors the original oxygen concentration signal in the root zone through dissolved oxygen sensors, smoothes the original oxygen concentration signal using low-pass filtering, and automatically adjusts the filter window length based on the pore connectivity. The lower the porosity, the longer the window length. Furthermore, by constraining the boundary value of the filter window length and superimposing the Gaussian white noise characteristic map, differential noise reduction is performed. The data points are calibrated according to the three-dimensional Cartesian coordinate system to generate an oxygen concentration time series set.

[0015] The pore connectivity index is obtained by performing CT scanning on soil samples and analyzing the geometry, size, distribution and connection mode of the pores using Avizo;

[0016] Gradient calculation submodule: calling the oxygen concentration time series set, calculating the concentration difference between adjacent points in a continuous time window, performing a specified time normalization process on the concentration difference, constructing a spatial gradient field using a piecewise interpolation method, introducing a radial basis function to compensate for missing boundary data, and generating a gradient distribution rate;

[0017] Trend judgment submodule: The gradient distribution rate is analyzed through a multivariate nonlinear coupling model, the root depth data is obtained through the root profile sampling method, and the total porosity is measured by the ring knife method. The soil porosity parameters are obtained by combining the mercury injection method. The root depth data and soil porosity parameters are integrated to calculate the diffusion rate of the root oxygen consumption rate gradient field in the direction of the pore connectivity index. The volume water content is measured in situ by a time domain reflectometer and the diffusion resistance coefficient within the nonlinear coupling model is adjusted to generate the oxygen consumption gradient trend state and transmit it to the osmotic pressure state monitoring module.

[0018] As a further solution of the present invention, the original oxygen concentration signal is smoothed using the formula:

[0019] ;

[0020] Among them, CLPF(m) is the output after filtering, which is the smoothing result after taking the average of the data in the window near index m in the original signal. N represents the window length, which is the total number of data points included in the filter window involved in the averaging calculation during filtering. It is dynamically adjusted according to the pore connectivity index. Represents the original data sequence, which is the input signal sequence without filtering. m represents the current index, which is the target data position where the filter value needs to be calculated. It represents the signal point currently being processed. j is the offset for the current index m, which is used to traverse all data points in the window. represents the mean coefficient, represents the unilateral offset, which is the number of data points on both sides of the window center point. N-1 represents the window symmetric offset, which is the total window width minus 1.

[0021] As a further solution of the present invention, the osmotic pressure status monitoring module includes:

[0022] Osmotic pressure acquisition submodule: Configure digital temperature sensors to monitor root temperature, deploy sensor nodes at depth intervals, synchronously collect temperature data, record osmotic pressure at fixed time intervals, and generate an osmotic pressure time series set including temperature compensation parameters;

[0023] Gradient difference submodule: calls the osmotic pressure time series set, combines it with the root temperature data collected by the digital temperature sensor, calculates the temperature-compensated osmotic pressure, extracts the osmotic pressure difference within adjacent time windows, and superimposes the oxygen consumption gradient trend state to generate the osmotic pressure gradient difference;

[0024] The change amplitude submodule: calls the peak-valley extreme values ​​in the osmotic pressure gradient difference, calculates the sliding window mean of the absolute value of the peak-valley difference, and the window length is 5 minutes. The sliding window mean is adjusted by the temperature compensation parameter to generate the dynamic offset of the osmotic pressure change and transmit it to the root metabolic state analysis module.

[0025] As a further embodiment of the present invention, the root metabolic state analysis module includes:

[0026] The damage risk submodule calls the dynamic offset of the osmotic pressure change and the trend state of the oxygen consumption gradient, calculates the covariance between the two in a continuous time window, extracts the main diagonal elements of the covariance, and compares them item by item with the preset damage risk threshold to generate a damage risk coefficient; the damage risk threshold is determined by the extreme value of the covariance when the cell damage probability is greater than 90% in the previous data; the state integration submodule calls the damage risk coefficient, superimposes the real-time temperature data, performs linear interpolation on the damage risk coefficient according to the temperature segment interval, and generates a metabolic state vector;

[0027] Model analysis submodule: Input the metabolic state vector into the multi-condition state machine model, define the damage risk coefficient and temperature value as the state transition condition, assign weights based on pore connectivity and root depth, input the multi-condition state machine model to determine the damage risk level, calculate the Manhattan distance between the current state and the metabolic abnormality pattern, generate the metabolic abnormality state determination result, and pass it to the root temperature parameter reconstruction module;

[0028] The weight factor of the Manhattan distance is dynamically adjusted according to the cell risk rating; the multi-condition state machine model is constructed by setting the discrete metabolic state of the root system physiology, setting the damage risk coefficient and logical rules, and determining the target metabolic abnormality pattern.

[0029] As a further solution of the present invention, the root temperature parameter reconstruction module includes:

[0030] Temperature data submodule: obtains the steady-state root temperature data in the metabolic abnormality determination result, performs standardization processing, and generates a temperature data set including a timestamp and a temperature feature mark;

[0031] Correlation analysis submodule: calling the temperature data set, traversing the corresponding data points of the metabolic abnormality trigger temperature series and the steady-state temperature series, calculating the absolute difference of each data point, extracting the minimum and maximum values ​​of all the differences, calling the resolution coefficient ρ=0.5 based on the grey correlation analysis algorithm, calculating the correlation coefficient of each data point, taking the arithmetic mean of all the correlation coefficients, and generating the correlation coefficient within the time window;

[0032] The control instruction submodule: calls the correlation coefficient, compares the coefficient sequence with the correlation threshold item by item, filters out nodes whose coefficients are lower than the correlation threshold, extracts the temperature adjustment amplitude and direction data of the node corresponding to the time stamp, and generates the root temperature control instruction; the correlation threshold determines the optimal classification performance value by the point corresponding to the maximum value of the Youden index on the ROC curve.

[0033] As a further solution of the present invention, the nutrient solution circulation linkage module includes:

[0034] Instruction parsing submodule: receives the root temperature control instruction, parses the temperature adjustment range and timestamp parameters in the instruction, filters out timed instructions based on timestamp alignment rules, and generates time control parameters that match the current clock;

[0035] Flow rate dynamic submodule: calls the time window of the time-effect control parameter, collects the original flow rate data of the electromagnetic flowmeter, sets the number of sample points of the sliding average filter according to the length of the time window, eliminates noise interference, and generates a flow rate dynamic baseline synchronized with the temperature control timing;

[0036] Speed ​​decision submodule: call the temperature adjustment amplitude in the time-effect control parameter and the flow rate dynamic baseline, map the temperature amplitude to the target flow rate increment based on the PID control algorithm, calculate the deviation value between the current flow rate baseline and the target increment, and adjust the circulation pump speed by superposition of proportional terms, integral terms, and differential terms; the PID control algorithm adopts the Ziegler-Nichols critical proportion method through the Kp, Ki, and Kd parameter tuning method. The Kp range is obtained by optimizing the parameter range through 50 flow rate step response experiments. The obtained Kp range is 0.5-5.0, Ki0.01-0.5, and Kd0.1-2.0. For every 10L / min increase in flow rate, Kp is increased by 10%-30% and Ki is decreased by 5%-15%. The adjustment range is self-matched and adjusted according to the flow rate fluctuation amplitude.

[0037] As a further solution of the present invention, the deviation between the current flow rate baseline and the target increment is calculated using the formula:

[0038] ;

[0039] Among them, u(t) is the real-time control signal output by the PID control algorithm, is the proportional gain, e(t) is the current error, Represents the integral gain, which eliminates the steady-state error according to the accumulated value of the error. represents the differential gain, represents the error change rate, t represents the current time, Represents the integral time variable.

[0040] In another aspect, a method for implementing intelligent root temperature control of hydroponic crops is provided. The method is applied to an intelligent root temperature control system for hydroponic crops. The method comprises:

[0041] S1: Real-time collection of root oxygen concentration data, calculation of root oxygen consumption rate gradient, and determination of oxygen consumption gradient trend based on a multivariable nonlinear coupling model;

[0042] S2: obtaining real-time data of cell osmotic pressure in the hydroponic nutrient solution through dissolved oxygen sensing, obtaining cell osmotic pressure change data, performing difference calculation based on the oxygen consumption gradient trend state, and generating a dynamic offset of osmotic pressure change;

[0043] S3: Based on the dynamic offset of the osmotic pressure change and the oxygen consumption rate gradient trend state, the root cell damage risk and the real-time root temperature of the risk state are determined, and the metabolic abnormality state determination result is obtained through a multi-condition state machine model;

[0044] S4: using a grey correlation analysis algorithm to compare the metabolic abnormality trigger temperature with the stable root temperature before the abnormal state occurs in the metabolic abnormality state determination result, and generate a root temperature control instruction;

[0045] S5: Call the root temperature control instruction, obtain the circulation flow rate data in real time, apply the PID control algorithm, and adjust the circulation pump speed through proportional, integral, and differential combined operations.

[0046] The technical solutions provided by the embodiments of the present invention offer at least the following beneficial effects: obtaining a root oxygen consumption rate gradient through periodic interpolation of dissolved oxygen concentrations, determining trend status using a univariate linear regression model, and dynamically quantifying the root metabolic activity rhythm. This trend is then used as a precondition for determining abnormal cell osmotic pressure, enhancing the accuracy of damage identification responses; combining the amplitude of cell osmotic pressure changes with the oxygen consumption gradient trend to determine root damage risk, improving the collaborative identification of abnormal cell function; dynamically selecting a target temperature setpoint using a gray correlation analysis algorithm by comparing the abnormal state determination temperature with the monitored historical stable state temperature, overcoming the limitations of a fixed threshold control approach; and after generating a root temperature control instruction, incorporating real-time circulation rate data to call a PID control algorithm to adjust the pump speed, achieving dynamic closed-loop matching between the root temperature setpoint and flow rate behavior. In this processing logic, a coupled metabolic trend and cell state judgment mechanism replaces a single-parameter threshold triggering approach, making the root temperature response event-driven and automatically generating a target setpoint with the minimum difference based on risk point data. This effectively enhances the targeted and real-time nature of root temperature control, further improving root stability and metabolic coordination in a hydroponic environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] 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.

[0048] Figure 1 is a system flow chart of the present invention;

[0049] Figure 2 is a system block diagram of the present invention;

[0050] Figure 3 Schematic diagram of the steps of the method of the present invention. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0052] 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.

[0053] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0054] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0055] 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.

[0056] The embodiment of the present invention provides a root temperature intelligent control system for hydroponic crops, please refer to Figures 1 to 2 The present invention provides a technical solution, a root temperature intelligent control system for hydroponic crops, comprising:

[0057] The root oxygen consumption monitoring module collects root oxygen concentration data in real time, calculates the root oxygen consumption rate gradient, determines the trend state of the oxygen consumption gradient based on a multivariable nonlinear coupling model, and transmits the data to the osmotic pressure status monitoring module;

[0058] The osmotic pressure status monitoring module obtains the cell osmotic pressure change data, performs difference calculation based on the oxygen consumption gradient trend state, obtains the dynamic offset of the osmotic pressure change, and transmits it to the root metabolism status analysis module;

[0059] The root metabolic state analysis module, based on the oxygen consumption gradient trend and the dynamic offset of osmotic pressure changes, assigns weights based on pore connectivity and root depth, inputs a multi-conditional state machine model to determine the damage risk level, obtains the metabolic abnormality determination result, and transmits it to the root temperature parameter reconstruction module;

[0060] The root temperature parameter reconstruction module compares the metabolic abnormality trigger temperature with the root temperature before the abnormal state occurs in the metabolic abnormality state judgment result, generates a root temperature control instruction, and transmits it to the nutrient solution circulation linkage module;

[0061] The nutrient solution circulation linkage module receives root temperature control instructions, obtains circulation flow rate data in real time, and calls the PID control algorithm to adjust the circulation pump speed;

[0062] The oxygen consumption gradient trend status includes the gradient direction vector, change frequency characteristics, and amplitude fluctuation range. The dynamic offset of osmotic pressure change specifically refers to the maximum positive deviation value, the maximum negative deviation value, and the duration of the offset. The metabolic abnormality status judgment result includes the abnormal condition category, cell risk rating, and root temperature at the corresponding moment. The root temperature control instruction specifically includes the target temperature range, control mode code, and the instruction effective time. The control result of adjusting the circulation pump speed includes the set speed value, expected flow rate target, and PID output parameters.

[0063] See also Figure 2 , the root oxygen consumption monitoring module includes:

[0064] Oxygen concentration acquisition submodule: This module monitors the original oxygen concentration signal in the root zone through dissolved oxygen sensors, smoothes the original oxygen concentration signal using low-pass filtering, and automatically adjusts the filter window length based on the pore connectivity. The lower the porosity, the longer the window length. Furthermore, by constraining the boundary value of the filter window length and superimposing the Gaussian white noise characteristic map, differential noise reduction is performed. The data points are calibrated according to the three-dimensional Cartesian coordinate system to generate an oxygen concentration time series set.

[0065] The pore connectivity index is obtained by performing CT scanning on soil samples and analyzing the geometry, size, distribution and connection of pores using Avizo;

[0066] The dissolved oxygen sensor collects the original oxygen concentration signal in a cycle of 0.5 seconds , where k is the discrete time sequence number, and the moving average low-pass filter formula is used , the original oxygen concentration signal is smoothed, where CLPF(m) is the filtered output, the filtering result at index m, which represents the smoothed data value, and N represents the total number of data points included in the filter window participating in the average calculation during filtering, which is dynamically adjusted according to the pore connectivity index. Represents the original data sequence, which is the unprocessed input signal, usually containing noise, m represents the current index, j represents the offset relative to the current index m, and is used to traverse adjacent data points in the window. Represents the mean coefficient, N-1 represents the window symmetry offset, where N is dynamically adjusted according to the pore connectivity index CI (reflecting pore connectivity). When the pore connectivity is low (CI < 0.4), N is increased (such as N = 7) to enhance noise reduction. When the pore connectivity is high (CI > 0.6), N is reduced to retain high-frequency details. At the same time, the N range is constrained to avoid filtering distortion caused by extreme values. When executing, set N = 5 (window time span 2.5 seconds). When k = 3, the filtering result is calculated by to If the original signal sequence is [8.4, 8.5, 8.3, 8.6, 8.2] mg / L, then The window length N is set based on the following: According to the sampling period of 0.5 seconds and the upper limit of the temperature fluctuation frequency of 0.1Hz (period of 10 seconds), it is necessary to cover at least 5 temperature fluctuation cycles to fully suppress high-frequency noise. Therefore, N=5 corresponds to a 2.5-second time window, which can filter out frequencies higher than =0.4Hz noise, while retaining the temperature component below 0.1Hz. After filtering, superimposed Gaussian white noise with a mean of 0 and a variance of 0.05, the noise value of -0.1mg / L was superimposed on the filtered value of 8.4mg / L to generate a characteristic spectrum value of 8.3mg / L. When performing differential denoising, the variance difference of adjacent periodic spectra was calculated to be 0.02, which was corrected to 8.35mg / L after triggering the denoising threshold. Combined with the three-dimensional coordinates (2.3m, 5.1m, -1.2m), a time series set of oxygen concentration was generated.

[0067] Table 1 Oxygen concentration time series data table

[0068]

[0069] As shown in Table 1, the filtering process smoothes the signal by arithmetic averaging the data within the window, and the window movement step is 1 sampling point (0.5 seconds) to generate an oxygen concentration time series set.

[0070] Gradient calculation submodule: Calls the oxygen concentration time series set, calculates the concentration difference between adjacent points in a continuous time window, performs specified time normalization on the concentration difference, uses piecewise interpolation to construct the spatial gradient field, introduces radial basis functions to compensate for missing boundary data, and generates the gradient distribution rate;

[0071] The oxygen concentration data of five adjacent monitoring points from 10:00 to 10:05 in Table 1 were used to calculate the concentration difference between adjacent points. When the difference ΔC between point A (8.4 mg / L) and point B (8.1 mg / L) was 0.3 mg / L, ΔC was normalized using a 1-minute time window. The 0.3 mg / L difference was divided by the time window length of 60 seconds to obtain a change rate of 0.005 mg / (L·s). The spatial gradient field was constructed using linear interpolation. When the gradient values ​​of the four adjacent points were [0.005, 0.004, 0.006, 0.003] mg / (L·s), a radial basis function with ε = 0.5 was introduced to compensate for the missing boundary points. When the theoretical gradient of the boundary point was 0.004 mg / (L·s), the gradient distribution rate after compensation was corrected to 0.004 × (1 + 0.5) = 0.006 mg / (L·s).

[0072] Trend judgment submodule: The gradient distribution rate is analyzed through a multivariate nonlinear coupling model. Root depth data is obtained through root profile sampling. The total porosity is measured using the ring knife method. Combined with the mercury intrusion method, soil porosity parameters are obtained. The root depth data and soil porosity parameters are integrated to calculate the diffusion rate of the root oxygen consumption rate gradient field in the direction of the pore connectivity index. The volumetric water content is measured in situ by time-domain reflectometry and the diffusion resistance coefficient within the nonlinear coupling model is adjusted to generate the oxygen consumption gradient trend state, which is then transmitted to the osmotic pressure state monitoring module.

[0073] The gradient distribution ratio was input into the multivariate model. When the root profile sampling depth was measured at 0.8 m, the ring knife method was used to obtain a soil sample with a wet weight of 125 g and a dry weight of 110 g. The total porosity was calculated as (125-110) / 125 × 100% = 12%. The mercury intrusion method measured the proportion of pores with a diameter greater than 50 μm as 35%, and the pore connectivity index was set as 12% × 35% = 4.2%. When the diffusion rate of the oxygen consumption rate gradient field in the X direction was 0.008 mg / (L·s), the volumetric water content of 28% measured by time-domain reflectometry was substituted into the model. If the water content exceeded 25%, the diffusion resistance coefficient was adjusted by 0.8. The original coefficient of 1.2 was adjusted to 1.2 × 0.8 = 0.96, thus generating the oxygen consumption gradient trend state.

[0074] See also Figure 2 , the osmotic pressure status monitoring module includes:

[0075] Osmotic pressure acquisition submodule: Configure digital temperature sensors to monitor root temperature, deploy sensor nodes at depth intervals, synchronously collect temperature data, record osmotic pressure at fixed time intervals, and generate an osmotic pressure time series set including temperature compensation parameters;

[0076] DS18B20 digital temperature sensors are deployed in the root profile at depths of 0.1m, 0.3m, and 0.5m. The sensor sampling frequency is 1Hz. When synchronously collecting temperature data, the original osmotic pressure value is recorded every 5 seconds. When the sensor at a depth of 0.3m measures a temperature of 25.6℃, the original osmotic pressure value is 1.2MPa. According to the temperature compensation parameters Correction for osmotic pressure is made, and the calculation formula is: ,in, It is the corrected osmotic pressure calculated by combining the temperature measurement value and the original osmotic pressure value through the compensation formula. Represents the raw measurement of osmotic pressure, reference temperature =25℃, when T=25.6℃, the osmotic pressure after compensation is 1.2+0.02×(25.6-25)=1.212MPa. After continuously recording 10 sets of data, the osmotic pressure time series set is generated.

[0077] Table 2 Osmotic pressure data table

[0078]

[0079] As shown in Table 2, the temperature compensation parameter α is set according to laboratory calibration data. The calibration method is as follows: the temperature is controlled to rise from 20°C to 30°C in a constant temperature chamber, the change in osmotic pressure is measured every time the temperature rises by 1°C, and the slope of the linear regression curve is fitted to obtain α = 0.02 MPa / °C. The time series set storage format is (timestamp, depth, compensated osmotic pressure).

[0080] Gradient difference submodule: This module calls the osmotic pressure time series set, combines it with the root temperature data collected by the DS18B20 digital temperature sensor, calculates the temperature-compensated osmotic pressure, extracts the osmotic pressure difference within adjacent time windows, and superimposes the oxygen consumption gradient trend state to generate the osmotic pressure gradient difference.

[0081] Call the compensated osmotic pressure data from 11:00 to 11:10 in Table 2 and calculate the difference between adjacent time windows (5-minute intervals). =11:00 =1.212MPa, =11:05 =1.228MPa, difference =1.228-1.212=0.016MPa, when the oxygen consumption gradient trend state parameter G=0.008mg / (L·s) is superimposed, w=0.5 is calculated to generate the osmotic pressure gradient difference =0.016+0.5×0.008=0.020MPa, the weight w is set according to the correlation coefficient between osmotic pressure and oxygen consumption rate. When the correlation coefficient is 0.6 calculated through data, it is mapped to w=0.6 / 1.2=0.5.

[0082] The change amplitude submodule calls the peak-valley extremes in the osmotic pressure gradient difference, calculates the sliding window mean of the absolute value of the peak-valley difference, and adjusts the sliding window mean using the temperature compensation parameter to generate the dynamic offset of the osmotic pressure change and pass it to the root metabolic state analysis module.

[0083] Extract the peak-valley extreme values ​​of the osmotic pressure gradient difference from 11:00 to 11:30 in Table 2. When the peak sequence is [0.020, 0.018, 0.022] MPa and the valley sequence is [0.015, 0.012, 0.014] MPa, calculate the sliding window mean (window length 3 groups) of the absolute value of the peak-valley difference. The peak-valley difference of window 1 is 、 、 , the mean is (0.005+0.006+0.008) / 3=0.0063MPa, and the mean is adjusted according to the temperature compensation parameter α. =0.5℃, adjustment coefficient ,in represents the temperature fluctuation amplitude, α represents the temperature compensation coefficient, and the dynamic offset of osmotic pressure change =0.0063×1.01=0.00636MPa.

[0084] See also Figure 2 , the root metabolic status analysis module includes:

[0085] The damage risk submodule uses the dynamic offset of osmotic pressure changes and the trend state of oxygen consumption gradient to calculate the covariance between the two within a continuous time window. The main diagonal elements of the covariance are extracted and compared item by item with the preset damage risk threshold to generate a damage risk coefficient. The damage risk threshold is determined by the extreme value of the covariance when the probability of cell damage in previous data is greater than 90%.

[0086] Calling the dynamic offset of osmotic pressure change and oxygen consumption gradient trend Time series data, intercept the data sequence of three consecutive time windows (window 1 to window 3), and calculate the covariance matrix , extract the main diagonal elements , calculate the mean for the data in window 1 , calculate the square of the deviation of each data point from the mean, sum it up and divide it by the degrees of freedom n-1=2, and we get , call the preset damage risk threshold , item by item comparison and , when window 1 When the risk factor is marked =0, and similarly calculate the values ​​of window 2 and window 3. They are 0.00000003 and 0.00000002 respectively, both less than the threshold, generating a risk coefficient sequence The threshold value is set based on the following: 100 groups of samples were counted through laboratory cell damage experiments, and the damage group mean , mean of the non-injured group , take the middle value As a dividing point.

[0087] State integration submodule: calls the damage risk coefficient, superimposes the real-time temperature data, performs linear interpolation on the damage risk coefficient according to the temperature segment interval, and generates a metabolic state vector;

[0088] Calling the damage risk factor and real-time temperature data , divided into temperature ranges: low temperature section T<25℃, medium temperature section 25≤T<28℃, high temperature section T≥28℃, when the temperature of window 2 =26.1℃ belongs to the medium temperature range, according to the interpolation rule , is the parameter value after temperature adjustment, calculated ,in It is the parameter value after the temperature adjustment in window 2. When the temperature in window 3 is =25.8℃, also belongs to the medium temperature range, calculate The temperature segmentation is based on the experimental data of plant root metabolic rate: metabolic activity decreases by 50% below 25℃, and enzyme activity decreases by 30% above 28℃. The interpolation coefficient is obtained by linearly fitting the slope of the experimental data. Determine and generate metabolic state vector .

[0089] Model analysis submodule: Input the metabolic state vector into the multi-conditional state machine model, define the damage risk coefficient and temperature value as the state transition conditions, and assign weights based on pore connectivity and root depth. Input the multi-conditional state machine model to determine the damage risk level, calculate the Manhattan distance between the current state and the metabolic abnormality pattern, generate the metabolic abnormality state judgment result, and pass it to the root temperature parameter reconstruction module;

[0090] The Manhattan distance weight factor is dynamically adjusted based on the cell risk rating. The multi-conditional state machine model is constructed by setting discrete metabolic states based on root physiology, setting damage risk coefficients and logical rules, and clarifying the target metabolic abnormality pattern.

[0091] The metabolic state vector is input into the multi-condition state machine model, and four discrete states are defined: normal (S0), mild abnormality (S1), moderate abnormality (S2), and severe abnormality (S3). Based on the real-time data obtained, a vector is constructed. , oxygen consumption gradient = -28%, based on porosity =0.3, connectivity CI=0.5, depth 0.8m, calculate oxygen consumption rate ; Osmotic pressure offset (Dynamic prediction model outputs expected value bar); root temperature =26.5℃, calculate Manhattan distance , contrast threshold , satisfying d≤2.4, calculate the damage risk coefficient ,in represents the intensity of oxygen consumption trend, represents the osmotic pressure bias weight, represents the porosity, represents the normalized value of root depth, PCI represents the soil pore connectivity index, and the state machine determines: d = 1.949 ≤ 2.4, but r = 0.8 > 0.5 (severe anomaly threshold); it is determined to be a severe anomaly (S3), and a metabolic abnormality state determination result is generated.

[0092] Table 3 Metabolic state determination data table

[0093]

[0094] See also Figure 2 , the root temperature parameter reconstruction module includes:

[0095] Temperature data submodule: obtains the steady-state root temperature data from the metabolic abnormality determination results, performs standardization processing, and generates a temperature data set including timestamps and temperature feature markers;

[0096] The temperature data submodule extracts the steady-state root temperature data from the metabolic abnormality judgment results and obtains the timestamp sequence Corresponding temperature data , the standardization process adopts the normalization method, and the formula is: ,in , , is the maximum temperature in the current time window 1, is the minimum temperature in the current time window 1, calculate the standardized temperature of window 1: , generate the temperature feature flag field is-anomaly, and assign a value based on the metabolic abnormality judgment result (if window 2 is judged to be abnormal, it is marked as 1). The temperature dataset format is: .

[0097] Table 4 Example of temperature data set table

[0098]

[0099] As shown in Table 4, the standardized temperature is calculated by normalization, and the abnormal flag field is bound to the metabolic judgment result.

[0100] Correlation analysis submodule: Call the temperature dataset, traverse the corresponding data points of the metabolic abnormality trigger temperature series and the steady-state temperature series, calculate the absolute difference of each data point, extract the minimum and maximum values ​​of all differences, call the resolution coefficient ρ = 0.5 based on the grey correlation analysis algorithm, calculate the correlation coefficient of each data point, take the arithmetic mean of all correlation coefficients, and generate the correlation coefficient within the time window;

[0101] The correlation analysis submodule calls the abnormal trigger sequence in the temperature dataset and steady-state series , traverse the data points and calculate the absolute difference, the difference of window 2: ,in, The temperature value corresponding to the first trigger event, is the temperature reference value in the second stable state, and the minimum value of all differences is extracted and maximum value , call the grey correlation analysis formula to calculate the correlation coefficient: , where the resolution coefficient ρ = 0.5, the correlation coefficient of window 2 is calculated as: , take the arithmetic mean of the correlation coefficients of all data points, the three windows , then the correlation coefficient is: ,

[0102] Control instruction submodule: Call the correlation coefficient, compare the coefficient sequence with the correlation threshold item by item, filter out nodes with coefficients lower than the correlation threshold, extract the temperature adjustment amplitude and direction data of the node's corresponding timestamp, and generate root temperature control instructions; the correlation threshold determines the optimal classification performance value by the point corresponding to the maximum value of the Youden index on the ROC curve;

[0103] Control instruction submodule calls correlation coefficient sequence The correlation threshold is determined by the Youden index of the ROC curve, and the true positive rate in the experimental data is , false positive rate , calculate the Youden index: ,get , the maximum value corresponds to the threshold , filter out nodes with coefficients lower than 0.4 (including 0.572>0.4 in window 3, which is not filtered out), extract the timestamps of nodes that meet the conditions (including 0.623>0.4 in window 1, which is not filtered out), and if no node meets the conditions, output "no regulation required", otherwise generate root temperature regulation instructions according to the temperature adjustment rules.

[0104] See also Figure 2 , the nutrient solution circulation linkage module includes:

[0105] The instruction parsing submodule receives root temperature control instructions, parses the temperature adjustment range and timestamp parameters in the instructions, filters out timed instructions based on timestamp alignment rules, and generates time-sensitive control parameters that match the current clock.

[0106] The instruction parsing submodule receives the root temperature control instruction from the control instruction submodule. For example, the instruction format is [timestamp: 10:25, adjustment range: +2.5°C]. The current system clock is 10:30. The timestamp alignment rule is called, the timeout threshold is defined as 5 minutes, and the difference between the instruction timestamp and the current time is calculated: =10:30-10:25=5 minutes, judgment ≤5 minutes is a valid instruction, keep the instruction, if If the time difference is greater than 5 minutes, it will be filtered out and the aging control parameters will be generated [effective time window: 10:25-10:30, adjustment range: +2.5℃]. In the example, the current time difference is 5 minutes, which just meets the threshold, so the instruction is retained.

[0107] Table 5 Timeout judgment rules

[0108]

[0109] As shown in Table 5, the time difference is calculated by the clock difference, and the timeout threshold is set according to the system response delay experiment.

[0110] Flow rate dynamic submodule: Call the time window of the time-effect control parameter, collect the original flow rate data of the electromagnetic flowmeter, set the number of sample points of the sliding average filter according to the length of the time window, eliminate noise interference, and generate a flow rate dynamic baseline synchronized with the temperature control timing;

[0111] The flow rate dynamic submodule calls the time window 10:25-10:30 of the aging control parameter and collects the original flow rate data from the electromagnetic flowmeter. For example, the original data is a sequence sampled once per second: , according to the time window length of 5 minutes (300 seconds), set the number of sample points of the sliding average filter to 300, and the calculation formula is ,in The Q value of the physical quantity after filtering, is the i-th original data point, Is the normalization coefficient for mean calculation, calculated using the first five sample points: , generating a dynamic baseline of flow velocity [102.48], which is synchronized with the temperature control timing.

[0112] Speed ​​decision submodule: Call the temperature adjustment amplitude and flow rate dynamic baseline in the time-effect control parameters, map the temperature amplitude to the target flow rate increment based on the PID control algorithm, calculate the deviation between the current flow rate baseline and the target increment, and adjust the circulation pump speed by superposition of proportional terms, integral terms, and differential terms; the PID control algorithm adopts the Ziegler-Nichols critical proportion method through the Kp proportional gain, Ki integral gain, and Kd differential gain parameter tuning method. The Kp range is 0.5-5.0, Ki 0.01-0.5, and Kd 0.1-2.0. For every 10L / min increase in flow rate, Kp is increased by 10%-30% and Ki is decreased by 5%-15%. The adjustment range is self-matched and adjusted according to the flow rate fluctuation amplitude;

[0113] The speed decision submodule calls the temperature adjustment range in the time control parameter =+2.5℃, mapped to the target flow rate increment, the mapping rule is: , Indicates the value of the physical quantity Q after filtering, the current flow rate baseline , target flow rate , collect real-time flow rate , calculate the deviation: =103.48-103.1=0.38L / min, e(t) represents the deviation over time, that is, the instantaneous difference between the target value and the current actual value. Based on the Ziegler-Nichols critical proportion method, the initial PID parameters are 、 、 , dynamically adjust the rules according to the flow rate increment: when the flow rate increases by 10L / min, Increase by 20%, Reduce by 10%, The current flow rate baseline is 102.48L / min (base value 100L / min), the increment is , does not reach the 10L / min threshold, the parameter remains at the initial value. Calculate the error signal , where u(t) is the real-time control signal output by the PID control algorithm, is the proportional gain, e(t) is the current error, Represents the integral gain, which eliminates the steady-state error according to the accumulated value of the error. represents the differential gain, represents the error change rate, t represents the current time, Represents the integral time variable, assuming that the integral term accumulates the error , differential term , u(t)=2.0×0.38+0.1×0.15+0.5×0.02=0.76+0.015+0.01=0.785, and introduce porosity ( , which is the ratio of pore volume to total soil volume) and oxygen consumption rate ( , where k represents the intrinsic oxygen consumption rate constant, represents the porosity attenuation coefficient, represents the connectivity gain coefficient, CI represents the pore connectivity index, Represents real-time oxygen concentration, dynamically corrects target flow rate increment and adjusts PID parameters based on oxygen consumption rate Base increment for temperature control (For example, +5L / min) to perform feedforward compensation, the formula is: ,in is the oxygen consumption flow rate coupling coefficient, is the reference oxygen consumption rate. In this example, when the porosity =0.3, connectivity CI=0.6, oxygen concentration When the target increment is corrected from +5L / min to +5.38L / min. PID parameter adaptation: low porosity ( =0.3) trigger proportional gain (Initial value ) According to the formula Increased to 3.36 to enhance response speed; when the flow rate increment threshold (10L / min) is not reached, the initial integral gain is retained and differential gain , real-time flow rate deviation Output speed adjustment value calculated by PID , to achieve dynamic coordination between soil pore structure characteristics and fluid control, improve accuracy, and adjust the circulation pump speed according to u(t)=0.785.

[0114] See also Figure 3 , methods include:

[0115] S1: Real-time collection of root oxygen concentration data, calculation of root oxygen consumption rate gradient, and determination of oxygen consumption gradient trend based on a multivariable nonlinear coupling model;

[0116] S2: Obtain real-time data on cell osmotic pressure in the hydroponic nutrient solution through dissolved oxygen sensing, obtain cell osmotic pressure change data, and perform difference calculation based on the oxygen consumption gradient trend state to generate a dynamic offset of osmotic pressure change;

[0117] S3: Based on the dynamic offset of osmotic pressure changes and the trend of oxygen consumption rate gradient, the root cell damage risk and real-time root temperature of the risk state are determined, and the abnormal metabolic state judgment result is obtained through a multi-condition state machine model;

[0118] S4: using a grey correlation analysis algorithm to compare the metabolic abnormality trigger temperature with the stable root temperature before the abnormal state occurs in the metabolic abnormality state determination result, and generate a root temperature control instruction;

[0119] S5: Call the root temperature control instruction, obtain the circulation flow rate data in real time, apply the PID control algorithm, and adjust the circulation pump speed through the proportional, integral, and differential combined operations.

[0120] It should be understood that the term "and / or" as used herein simply describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the related objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0121] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0122] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0123] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0124] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0125] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, and can be electrical, mechanical, or other forms.

[0126] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0127] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0128] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0129] The above are merely specific embodiments 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 the present invention should be included within 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.

Claims

1. An intelligent root temperature control system for hydroponic crops, characterized in that: The system comprises: The root oxygen consumption monitoring module collects root oxygen concentration data in real time, calculates the root oxygen consumption rate gradient, determines the trend state of the oxygen consumption gradient based on a multivariable nonlinear coupling model, and transmits the data to the osmotic pressure status monitoring module; The osmotic pressure status monitoring module obtains the cell osmotic pressure change data, performs difference calculation based on the oxygen consumption gradient trend state, obtains the dynamic offset of the osmotic pressure change, and transmits it to the root metabolism status analysis module; The root metabolic state analysis module, based on the oxygen consumption gradient trend state and the osmotic pressure change dynamic offset, assigns weights by pore connectivity and root depth, inputs a multi-condition state machine model to determine the damage risk level, obtains a metabolic abnormality determination result, and transmits it to the root temperature parameter reconstruction module; The root temperature parameter reconstruction module compares the metabolic abnormality trigger temperature with the root system temperature before the abnormal state occurs in the metabolic abnormality state determination result, generates a root system temperature control instruction, and transmits it to the nutrient solution circulation linkage module; The nutrient solution circulation linkage module receives the root temperature control instruction, obtains the circulation flow rate data in real time, and calls the PID control algorithm to adjust the circulation pump speed.

2. The root temperature intelligent control system for hydroponic crops according to claim 1, characterized in that: The oxygen consumption gradient trend state includes the gradient direction vector, the change frequency characteristics, and the amplitude fluctuation range. The dynamic offset of the osmotic pressure change is the maximum positive deviation value, the maximum negative deviation value, and the offset duration. The metabolic abnormality state judgment result includes the abnormal condition category, the cell risk rating, and the root temperature at the corresponding moment. The root temperature control instruction is the target temperature range, the control mode code, and the instruction effective time. The control result of adjusting the circulation pump speed includes the set speed value, the expected flow rate target, and the PID output parameter.

3. The root temperature intelligent control system for hydroponic crops according to claim 1, characterized in that: The root oxygen consumption monitoring module includes: Oxygen concentration acquisition submodule: This module monitors the original oxygen concentration signal in the root zone through dissolved oxygen sensors, smoothes the original oxygen concentration signal using low-pass filtering, and automatically adjusts the filter window length based on the pore connectivity. The lower the porosity, the longer the window length. Furthermore, by constraining the boundary value of the filter window length and superimposing the Gaussian white noise characteristic map, differential noise reduction is performed. The data points are calibrated according to the three-dimensional Cartesian coordinate system to generate an oxygen concentration time series set. The pore connectivity index is obtained by performing CT scanning on soil samples and analyzing the geometry, size, distribution and connection mode of the pores using Avizo; Gradient calculation submodule: calling the oxygen concentration time series set, calculating the concentration difference between adjacent points in a continuous time window, performing a specified time normalization process on the concentration difference, constructing a spatial gradient field using a piecewise interpolation method, introducing a radial basis function to compensate for missing boundary data, and generating a gradient distribution rate; Trend judgment submodule: The gradient distribution rate is analyzed through a multivariate nonlinear coupling model, the root depth data is obtained through the root profile sampling method, and the total porosity is measured by the ring knife method. The soil porosity parameters are obtained by combining the mercury injection method. The root depth data and soil porosity parameters are integrated to calculate the diffusion rate of the root oxygen consumption rate gradient field in the direction of the pore connectivity index. The volume water content is measured in situ by a time domain reflectometer and the diffusion resistance coefficient within the nonlinear coupling model is adjusted to generate the oxygen consumption gradient trend state and transmit it to the osmotic pressure state monitoring module.

4. The intelligent root temperature control system for hydroponic crops according to claim 3, characterized in that: The original oxygen concentration signal is smoothed using the formula: ； Among them, CLPF(m) is the output after filtering, which is the smoothing result after taking the average of the data in the window near index m in the original signal. N represents the window length, which is the total number of data points included in the filter window involved in the averaging calculation during filtering. It is dynamically adjusted according to the pore connectivity index. Represents the original data sequence, which is the input signal sequence without filtering. m represents the current index, which is the target data position where the filter value needs to be calculated. It represents the signal point currently being processed. j is the offset for the current index m, which is used to traverse all data points in the window. represents the mean coefficient, represents the unilateral offset, which is the number of data points on both sides of the window center point. N-1 represents the window symmetric offset, which is the total window width minus 1.

5. The intelligent root temperature control system for hydroponic crops according to claim 1, characterized in that: The osmotic pressure status monitoring module includes: Osmotic pressure acquisition submodule: Configure digital temperature sensors to monitor root temperature, deploy sensor nodes at depth intervals, synchronously collect temperature data, record osmotic pressure at fixed time intervals, and generate an osmotic pressure time series set including temperature compensation parameters; Gradient difference submodule: calls the osmotic pressure time series set, combines it with the root temperature data collected by the digital temperature sensor, calculates the temperature-compensated osmotic pressure, extracts the osmotic pressure difference within adjacent time windows, and superimposes the oxygen consumption gradient trend state to generate the osmotic pressure gradient difference; The change amplitude submodule: calls the peak-valley extreme values ​​in the osmotic pressure gradient difference, calculates the sliding window mean of the absolute value of the peak-valley difference, and the window length is 5 minutes. The sliding window mean is adjusted by the temperature compensation parameter to generate the dynamic offset of the osmotic pressure change and transmit it to the root metabolic state analysis module.

6. The intelligent root temperature control system for hydroponic crops according to claim 1, characterized in that: The root metabolic state analysis module includes: The damage risk submodule calls the dynamic offset of the osmotic pressure change and the trend state of the oxygen consumption gradient, calculates the covariance between the two in a continuous time window, extracts the main diagonal elements of the covariance, and compares them item by item with the preset damage risk threshold to generate a damage risk coefficient; the damage risk threshold is determined by the extreme value of the covariance when the cell damage probability is greater than 90% in the previous data; the state integration submodule calls the damage risk coefficient, superimposes the real-time temperature data, performs linear interpolation on the damage risk coefficient according to the temperature segment interval, and generates a metabolic state vector; Model analysis submodule: Input the metabolic state vector into the multi-condition state machine model, define the damage risk coefficient and temperature value as the state transition condition, assign weights based on pore connectivity and root depth, input the multi-condition state machine model to determine the damage risk level, calculate the Manhattan distance between the current state and the metabolic abnormality pattern, generate the metabolic abnormality state determination result, and pass it to the root temperature parameter reconstruction module; The weight factor of the Manhattan distance is dynamically adjusted according to the cell risk rating; the multi-condition state machine model is constructed by setting the discrete metabolic state of the root system physiology, setting the damage risk coefficient and logical rules, and determining the target metabolic abnormality pattern.

7. The intelligent root temperature control system for hydroponic crops according to claim 1, characterized in that: The root temperature parameter reconstruction module includes: Temperature data submodule: obtains the steady-state root temperature data in the metabolic abnormality determination result, performs standardization processing, and generates a temperature data set including a timestamp and a temperature feature mark; Correlation analysis submodule: calling the temperature data set, traversing the corresponding data points of the metabolic abnormality trigger temperature series and the steady-state temperature series, calculating the absolute difference of each data point, extracting the minimum and maximum values ​​of all the differences, calling the resolution coefficient ρ=0.5 based on the grey correlation analysis algorithm, calculating the correlation coefficient of each data point, taking the arithmetic mean of all the correlation coefficients, and generating the correlation coefficient within the time window; The control instruction submodule: calls the correlation coefficient, compares the coefficient sequence with the correlation threshold item by item, filters out nodes whose coefficients are lower than the correlation threshold, extracts the temperature adjustment amplitude and direction data of the node corresponding to the time stamp, and generates the root temperature control instruction; the correlation threshold determines the optimal classification performance value by the point corresponding to the maximum value of the Youden index on the ROC curve.

8. The intelligent root temperature control system for hydroponic crops according to claim 1, characterized in that: The nutrient solution circulation linkage module includes: Instruction parsing submodule: receives the root temperature control instruction, parses the temperature adjustment range and timestamp parameters in the instruction, filters out timed instructions based on timestamp alignment rules, and generates time control parameters that match the current clock; Flow rate dynamic submodule: calls the time window of the time-effect control parameter, collects the original flow rate data of the electromagnetic flowmeter, sets the number of sample points of the sliding average filter according to the length of the time window, eliminates noise interference, and generates a flow rate dynamic baseline synchronized with the temperature control timing; Speed ​​decision submodule: call the temperature adjustment amplitude in the time-effect control parameter and the flow rate dynamic baseline, map the temperature amplitude to the target flow rate increment based on the PID control algorithm, calculate the deviation value between the current flow rate baseline and the target increment, and adjust the circulation pump speed by superposition of proportional terms, integral terms, and differential terms; the PID control algorithm adopts the Ziegler-Nichols critical proportion method through the Kp, Ki, and Kd parameter tuning method. The Kp range is obtained by optimizing the parameter range through 50 flow rate step response experiments. The obtained Kp range is 0.5-5.0, Ki0.01-0.5, and Kd0.1-2.

0. For every 10L / min increase in flow rate, Kp is increased by 10%-30% and Ki is decreased by 5%-15%. The adjustment range is self-matched and adjusted according to the flow rate fluctuation amplitude.

9. The intelligent root temperature control system for hydroponic crops according to claim 8, characterized in that: The deviation between the current velocity baseline and the target increment is calculated using the formula: ； Among them, u(t) is the real-time control signal output by the PID control algorithm, is the proportional gain, e(t) is the current error, Represents the integral gain, which eliminates the steady-state error according to the accumulated value of the error. represents the differential gain, represents the error change rate, t represents the current time, Represents the integral time variable.

10. A method for realizing intelligent control of root temperature of hydroponic crops, characterized in that: The method is used to implement the root temperature intelligent control system of hydroponic crops according to any one of claims 1 to 9, and the method comprises: S1: Real-time collection of root oxygen concentration data, calculation of root oxygen consumption rate gradient, and determination of oxygen consumption gradient trend based on a multivariable nonlinear coupling model; S2: obtaining real-time data of cell osmotic pressure in the hydroponic nutrient solution through dissolved oxygen sensing, obtaining cell osmotic pressure change data, performing difference calculation based on the oxygen consumption gradient trend state, and generating a dynamic offset of osmotic pressure change; S3: Based on the dynamic offset of the osmotic pressure change and the oxygen consumption rate gradient trend state, the root cell damage risk and the real-time root temperature of the risk state are determined, and the metabolic abnormality state determination result is obtained through a multi-condition state machine model; S4: using a grey correlation analysis algorithm to compare the metabolic abnormality trigger temperature with the stable root temperature before the abnormal state occurs in the metabolic abnormality state determination result, and generate a root temperature control instruction; S5: Call the root temperature control instruction, obtain the circulation flow rate data in real time, apply the PID control algorithm, and adjust the circulation pump speed through proportional, integral, and differential combined operations.

Images