Slope greening moisture regulation and control system and moisture regulation and control method

Abstract

The invention discloses a slope greening moisture regulation and control system and a moisture regulation and control method. The system comprises a photovoltaic power generation module which supplies power to the system; the sprinkling irrigation module is provided with an accurate drip irrigation system, a sprinkling irrigation system and a micro-irrigation system which are respectively suitable for the upper layer, the middle layer, the lower layer and the edge local area of the slope to realize accurate moisture compensation; the moisture detection module monitors related data in real time through soil humidity, weather and gradient sensors; and the intelligent control module receives the detection data, predicts the water demand by using a machine learning algorithm, and dynamically regulates and controls the sprinkling irrigation module according to the water demand. The regulation and control method comprises the steps of dividing a slope area, collecting preprocessed data, constructing an LSTM moisture demand prediction model, generating a self-adaptive regulation and control strategy based on Q-learning, performing real-time feedback optimization and the like. According to the system, accurate moisture regulation and control, intelligent management and energy self-sufficiency are achieved, the resource utilization efficiency is improved, the slope greening moisture management problem is effectively solved, and sustainable development of slope greening is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of slope greening, and in particular to a slope greening moisture control system and a moisture control method. Background Art

[0002] With the acceleration of urbanization, slope greening has gradually been widely used as an important ecological landscape and a means to prevent soil erosion. Slope greening plays an important role in improving the environment, reducing soil erosion, and enhancing landscape beautification. However, due to the complex slope terrain, uneven soil moisture distribution, and the great impact of climate change, the water management and irrigation regulation of slope greening face some challenges. At present, the water regulation system in the field of slope greening mainly relies on traditional automatic sprinkler systems. Although these systems can achieve timed and quantitative irrigation to a certain extent, they still cannot adjust the irrigation volume in real time according to soil moisture and meteorological conditions. Most existing irrigation systems only supply water according to set time intervals or fixed irrigation volumes, lacking real-time monitoring and intelligent regulation. The inability to fully utilize precipitation or natural wet conditions leads to a waste of water resources. Many traditional irrigation systems rely on power grid power supply, which may not be possible for slope areas far away from the power grid. Therefore, the above problems need to be solved urgently. Summary of the invention

[0003] Purpose of the invention: In order to overcome the shortcomings of the prior art, the present invention provides a slope greening moisture control system and method thereof, aiming to solve the problems of low moisture control accuracy, unstable energy supply, insufficient intelligence, etc. in the prior art. The present invention combines photovoltaic power generation, intelligent control module, precision irrigation technology and machine learning algorithm to achieve precise control and intelligent management of slope greening moisture, ensuring a balance between water saving, energy saving and plant growth needs.

[0004] The slope greening moisture control system includes: a photovoltaic power generation module, which is used to supply energy to the system; a sprinkler irrigation module, which is used to achieve moisture compensation in the upper slope layer, the middle slope layer and the lower slope layer of the slope; a moisture detection module, which is used to monitor the soil moisture, meteorological conditions and slope gradient of each area of ​​the slope in real time; an intelligent control module, which receives the detection data of the moisture detection module, analyzes the detection data and predicts the moisture demand, predicts the future moisture demand through a machine learning algorithm, and dynamically controls the sprinkler irrigation module according to the moisture demand;

[0005] Furthermore, the sprinkler irrigation module described in the present invention includes: a precision drip irrigation system, a sprinkler irrigation system and a micro-irrigation system. The drip irrigation system is suitable for the upper and middle layers of the slope to accurately adjust the water supply; the dripper spacing and flow rate on the drip irrigation pipe can be customized according to factors such as the distribution density of plants in the upper and middle layers, root characteristics, and soil water retention capacity. For plants with shallow roots and dense distribution, drip irrigation pipes with smaller dripper spacing and lower flow rates can be selected to ensure that each plant can obtain an appropriate amount of water while avoiding excessive concentration of water causing waste or water accumulation at the roots. The sprinkler irrigation system is used in the lower layer of the slope to replenish water by uniform water spraying; the type of nozzle and the spraying angle need to be selected and adjusted according to the terrain characteristics and vegetation coverage of the lower layer of the slope. For the lower layer with gentle slopes and more uniform vegetation coverage, a rotary nozzle can be used, which has a wide spraying range and uniform coverage, and can quickly and effectively provide water for a large area of ​​vegetation; for areas with steep slopes or uneven vegetation distribution, it is necessary to select a scattering nozzle or an adjustable angle nozzle to ensure that water can be accurately sprayed to the area that needs irrigation to avoid water loss due to terrain reasons; the micro-irrigation system is used for the edge of the slope and local areas for refined irrigation. The plant growth environment in these areas is relatively special, close to rocks, poor soil, etc., and more sophisticated water management is required. The micro-irrigation system can use drop arrows and micro-sprinklers to accurately deliver water to the roots of each plant, ensuring that plants can obtain sufficient and appropriate water supply under special conditions, and promoting the healthy growth of plants.

[0006] Furthermore, the moisture detection module in the present invention includes: a soil moisture sensor, a meteorological sensor, and a slope sensor. The soil moisture sensor is set at different depths on the slope to collect soil moisture data in real time; the meteorological sensor includes a temperature sensor, an ambient humidity sensor, a precipitation sensor, and a wind speed sensor, which are used to obtain meteorological condition data; the slope sensor is used to monitor the slope of each area of ​​the slope, evaluate the risk of water loss, and transmit data to the intelligent control module for dynamically adjusting the water supply.

[0007] Furthermore, the intelligent control module in the present invention includes an analysis unit and an irrigation control unit, wherein the analysis unit is used to predict future water demand using a machine learning algorithm based on historical data and real-time data;

[0008] The irrigation control unit dynamically adjusts the working state of the sprinkler module according to the water demand prediction result, and automatically controls the water volume and spraying time.

[0009] Furthermore, the photovoltaic power generation module in the present invention includes a solar panel and an energy storage battery; the solar panel is used to collect solar energy and convert it into electrical energy; the energy storage battery is used to store the electrical energy converted by the solar panel, and the energy storage battery is used to store electrical energy, and provide continuous power supply through the energy storage battery in the absence of light or insufficient light.

[0010] The slope greening water control method is applied to the slope greening water control system, and the method comprises the following steps:

[0011] S1: Divide the slope area into upslope layer, midslope layer and downslope layer, start the real-time data acquisition equipment, obtain the soil moisture, meteorological conditions and slope information of each layer, and pre-process the collected data. The pre-processed data will be sent to the intelligent control module as input;

[0012] S2: In the intelligent control module, a long short-term memory network (LSTM) is used to build a water demand prediction model. The LSTM model is trained using preprocessed real-time data and historical data to capture the water variation patterns and periodic characteristics of different slope layers to predict future water demand;

[0013] S3: After obtaining the water demand prediction results of the LSTM model, the intelligent control module uses the reinforcement learning algorithm (Q-learning) to generate an adaptive control strategy. This strategy dynamically balances the water supply and photovoltaic power to ensure optimal resource utilization;

[0014] S4: The system obtains new real-time data at predetermined intervals, preprocesses the data and updates the LSTM prediction results. If it is detected that the soil moisture exceeds the set range or the photovoltaic power is lower than the threshold, the Q-learning strategy is re-run to adjust the irrigation parameters to ensure the continuous stability of water regulation;

[0015] S5: After each irrigation cycle, the soil moisture changes, irrigation water volume and energy consumption data of each layer of the slope are recorded and input into the LSTM and Q-learning models as a new round of training data to continuously optimize the water regulation strategy and achieve adaptive learning.

[0016] The slope greening moisture control method described in the present invention, in the step S1, the pre-processed data includes standardization and denoising of soil moisture data, meteorological data and slope data to improve the accuracy of model prediction; the soil moisture data at different depths are normalized and converted into a unified scale to eliminate the differences between different soil layers, the random noise in the meteorological sensor measurement is removed by using a filter to ensure the stability of the meteorological data, and the slope value measured by the slope sensor is standardized to avoid the influence of different slope ranges on model training.

[0017] In the slope greening water control method described in the present invention, in step S2, the data used by the LSTM model includes historical soil moisture data, historical meteorological data, and historical irrigation data. The LSTM model captures long-term dependencies in the time series through long short-term memory units (LSTM units) to predict future water demand; the training formula of the LSTM model is as follows:

[0018] h t =σ(W h ·[h t-1 ,x t ]+b h )

[0019]

[0020] Among them, h t is the hidden state of the current time step, C t is the cell state at the current time step, x t is the input of the current time step, f t For the forget gate, i t is the input gate, is the candidate cell state, σ represents the activation function, and * represents the element-wise multiplication operation.

[0021] In the slope greening moisture control method described in the present invention, the adaptive control strategy generated by the reinforcement learning algorithm (Q-learning) in step S3 is updated by the following formula:

[0022]

[0023] Among them, Q(s t ,a t ) is state s t Take action a t The quality value of α is the learning rate, r t is the current reward value, γ is the discount factor, s t+1 is the next state, a t+1 For the next action.

[0024] In the slope greening moisture control method described in the present invention, in step S4, if it is detected that the system energy is insufficient or the soil moisture exceeds the set range, the intelligent control module dynamically adjusts the irrigation parameters according to the Q-learning strategy and re-optimizes the training data of the LSTM model to ensure timely updating of the moisture control strategy and optimization of resource utilization.

[0025] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0026] The slope greening moisture control system and method of the present invention achieves the following advantages by integrating photovoltaic power generation, intelligent control, precision irrigation and machine learning technology:

[0027] 1. Precise water control: According to the water demand of different levels of the slope, precise drip irrigation, sprinkler irrigation and micro-irrigation are used to supply water, effectively avoiding waste of water resources.

[0028] 2. Intelligent management: Use the LSTM model to predict future water demand, and use the Q-learning algorithm to achieve adaptive regulation, which can dynamically adjust the irrigation strategy based on real-time data.

[0029] 3. Energy self-sufficiency: The combination of photovoltaic power generation modules and energy storage batteries enables the system to operate independently in slope areas without grid access, ensuring continuous and stable power supply to the system.

[0030] 4. Improve resource utilization efficiency: The system adjusts irrigation volume and power supply strategy based on real-time monitoring data to ensure optimal use of water and energy and support the sustainable development of slope greening BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is the architecture diagram of the slope greening moisture control system of the present invention;

[0032] Figure 2 This is a step diagram of the slope greening moisture control method of the present invention. DETAILED DESCRIPTION

[0033] Example 1

[0034] like Figure 1 , 2 Shown

[0035] 1. System configuration and device initialization

[0036] 1.1 Slope area division

[0037] The slope area is divided into the upper slope layer, the middle slope layer, the lower slope layer, the slope edge and local areas, and the different water requirements of plants in each area are clarified.

[0038] 1.2 Equipment installation and connection

[0039] Sprinkler Module

[0040] Install precision drip, sprinkler and micro irrigation systems:

[0041] Precision drip irrigation pipes are installed on the upslope and mid-slope layers to achieve precise water supply;

[0042] A sprinkler irrigation system is installed on the downslope layer to evenly replenish water;

[0043] Micro-irrigation systems are installed on the slope edges and local areas to achieve localized fine irrigation;

[0044] Photovoltaic power generation module

[0045] Solar panels (power 800W) and energy storage batteries (capacity 500Ah) are installed. The power of the solar panels is set to 800W based on the area estimation of the green area on the slope, the local sunshine duration and intensity statistics, and the calculation of the system energy consumption demand, in order to ensure that the solar energy can be fully collected to power the system during the day, while avoiding cost waste caused by over-configuration. The energy storage battery capacity of 500Ah is determined based on the time the system needs to maintain operation in the absence of light or insufficient light, the power consumption of each module, and the battery charging and discharging efficiency, so as to ensure the continuous and stable power supply of the system at night or on rainy days.

[0046] Soil moisture sensor: buried at different depths (0cm, 10cm, 20cm) on the slope to obtain moisture information of different soil layers;

[0047] Meteorological sensors: including temperature, humidity, precipitation, wind speed measurement, and environmental condition data collection; Slope sensors: detect the slope of each area of ​​the slope to assess the risk of water loss, especially when the slope is greater than 30°

[0048] Areas, sensor spacing is increased to enhance monitoring of steep areas.

[0049] 2. Data Collection and Preprocessing

[0050] 2.1 Data Collection

[0051] Data is collected every 15 minutes to record the current soil moisture, meteorological conditions, and slope information, and time series data is established and stored in the database.

[0052] 2.2 Data Preprocessing

[0053] Soil moisture standardization: Normalize the moisture data at different depths so that the value range is unified between 0 and 1 to eliminate the impact of soil layer differences.

[0054] Meteorological data denoising: Use a low-pass filter to remove noise from sensor data to ensure the stability of data such as temperature, humidity, and wind speed.

[0055] Slope data normalization: Normalize the slope values ​​to a range of 0 to 1 so that differences in sensor locations do not affect model predictions.

[0056] 3. Water demand prediction model (based on LSTM)

[0057] LSTM model construction and training

[0058] Model configuration:

[0059] Input layer: soil moisture, meteorological conditions, and time series of historical irrigation data are used as input;

[0060] Hidden layer: contains 64 LSTM units to extract long-term dependencies in time series;

[0061] Output layer: predict the soil moisture demand of each layer in the next hour.

[0062] LSTM unit formula:

[0063] h t =σ(W h ·[h t-1 ,x t ]+b h )

[0064]

[0065] h t : The hidden state of the current time step;

[0066] C t : Cell state, used to memorize feature information for a long time;

[0067] x t : Current time step input, including real-time soil moisture, meteorological and other data;

[0068] f t 、i t : Forget gate and input gate control information transmission;

[0069] σ: activation function.

[0070] Model training: 50 epochs (the process of completely traversing the training data set. Each time the model performs forward and backward propagation on all training samples (each piece of data) and completes a parameter update, this is counted as an epoch) and a batch size of 32 are used to train the LSTM model to adapt the model to the changing laws of water demand in different slope layers.

[0071] 4. Adaptive irrigation strategy generation (based on Q-learning)

[0072] Q-learning algorithm configuration

[0073] Parameter settings:

[0074] The learning rate α=0.1, the discount factor γ=0.9, and the initial quality value Q table is empty.

[0075] Q-learning formula:

[0076]

[0077] Among them, Q(s t ,a t ) is state s t Take action a t The quality value of α is the learning rate, r t is the current reward value, γ is the discount factor, s t+1 is the next state, a t+1 For the next action, the optimal action is selected based on the moisture prediction results and the photovoltaic power.

[0078] Policy Update:

[0079] The Q-learning algorithm is used to adjust the water supply of sprinkler irrigation, micro-irrigation, and drip irrigation in real time to ensure the balance between water demand and photovoltaic power. Specific irrigation parameters for each layer are generated according to the different water demands of uphill, mid-slope, and downhill.

[0080] Strategy execution Based on the Q-learning strategy, the irrigation duration and amount of each area are controlled in real time:

[0081] The irrigation volume of the upslope layer is set at 1L / m 2 / Second-rate;

[0082] The mid-slope layer is set at 1.5L / m 2 / Second-rate;

[0083] The downhill layer is set to 1L / m 2 / times, but at a higher frequency to ensure adequate water.

[0084] The downslope layer usually accumulates water loss from the upslope and midslope layers due to gravity. Although the water demand of the downslope layer is similar to that of the upslope layer (set to 1L / m 2 / times), but because water is more easily lost in this area, it is necessary to increase the irrigation frequency to ensure that the water is maintained at an ideal level. The water loss in the middle slope layer is moderate, and there may be water accumulation compared to the upper slope layer. It is also affected by gravity and loss, so it is set at a slightly higher 1.5L / m 2 / times to fully compensate for possible water consumption. The upslope layer loses water faster due to the high slope, and it is difficult for irrigation water to remain in the soil for a long time. By setting a lower 1L / m 2 / irrigation amount, reducing the pressure of excessive water leakage from the top of the slope on the lower areas.

[0085] 5. Real-time feedback and system adaptive optimization

[0086] Data Feedback and Model Update

[0087] Soil moisture, weather and energy consumption data are obtained every hour to update the LSTM prediction and Q-learning strategy. If the soil moisture deviates from the target value, the irrigation strategy is regenerated.

[0088] Adaptive Optimization

[0089] After each irrigation, the system uses real-time data as training data for LSTM and Q-learning to iteratively optimize the water control strategy.

[0090] 6. Summary of operation process

[0091] (1) Start the sensor to collect and pre-process the real-time slope data;

[0092] (2) Use the LSTM model to predict water demand in the next hour;

[0093] (3) Generate adaptive irrigation strategies through Q-learning and set sprinkler irrigation parameters;

[0094] (4) Implement irrigation, update data every 15 minutes, and adjust strategies if humidity or energy deviates;

[0095] (5) Record data after the irrigation cycle to optimize the LSTM and Q-learning models and improve the intelligence of the system. The above implementation methods are exemplary, and their purpose is to illustrate the technical concept and features of the present invention so that people familiar with the technology in this field can understand the content of the present invention and implement it accordingly, and they cannot be used to limit the scope of protection of the present invention. Any changes or modifications made according to the spirit of the present invention should be included in the scope of protection of the present invention.

Claims

1. The slope greening water control system is characterized by: include: A photovoltaic power generation module, which is used to supply energy to the system; A sprinkler irrigation module, which is used to achieve moisture compensation for the upslope layer, midslope layer and downslope layer of the slope; a moisture detection module, which is used to monitor the soil moisture, meteorological conditions and slope gradient of each area of ​​the slope in real time; an intelligent control module, which receives the detection data of the moisture detection module, analyzes the detection data and predicts the moisture demand, predicts the future moisture demand through a machine learning algorithm, and dynamically adjusts the sprinkler irrigation module according to the moisture demand.

2. The slope greening moisture control system according to claim 1 is characterized by: The sprinkler irrigation module includes: a precise drip irrigation system, a sprinkler irrigation system and a micro-irrigation system. The drip irrigation system is suitable for the upper and middle layers of the slope to accurately adjust the water supply; the sprinkler irrigation system is used for the lower layer of the slope to replenish water by uniformly spraying water; the micro-irrigation system is used for the edge and local areas of the slope for refined irrigation.

3. The slope greening moisture control system according to claim 2 is characterized by: The moisture detection module includes: a soil moisture sensor, a meteorological sensor, and a slope sensor. The soil moisture sensor is set at different depths on the slope to collect soil moisture data in real time; the meteorological sensor includes a temperature sensor, an ambient humidity sensor, a precipitation sensor, and a wind speed sensor, which are used to obtain meteorological condition data; the slope sensor is used to monitor the slope of each area of ​​the slope, evaluate the risk of water loss, and transmit data to the intelligent control module for dynamically adjusting the water supply.

4. The slope greening moisture control system according to claim 3 is characterized by: The intelligent control module includes an analysis unit and an irrigation control unit. The analysis unit is used to predict future water demand based on historical data and real-time data using a machine learning algorithm; the irrigation control unit dynamically adjusts the working state of the sprinkler module based on the water demand prediction result and automatically controls the water volume and spraying time.

5. The slope greening moisture control system according to claim 4 is characterized in that: The photovoltaic power generation module includes a solar panel and an energy storage battery; the solar panel is used to collect solar energy and convert it into electrical energy; the energy storage battery is used to store the electrical energy converted by the solar panel, and the energy storage battery is used to store electrical energy, and provide continuous power supply through the energy storage battery when there is no light or insufficient light.

6. The method for regulating water content in slope greening is characterized by: Applied to the slope greening moisture control system according to any one of claims 1 to 5, the method comprises the following steps: S1: Divide the slope area into upslope layer, midslope layer and downslope layer, start the real-time data acquisition equipment, obtain the soil moisture, meteorological conditions and slope information of each layer, and pre-process the collected data. The pre-processed data will be sent to the intelligent control module as input; S2: In the intelligent control module, a long short-term memory network (LSTM) is used to build a water demand prediction model. The LSTM model is trained using preprocessed real-time data and historical data to capture the water variation patterns and periodic characteristics of different slope layers to predict future water demand. S3: After obtaining the water demand prediction results of the LSTM model, the intelligent control module uses the reinforcement learning algorithm (Q-learning) to generate an adaptive control strategy, which dynamically balances the water supply and photovoltaic power to ensure optimal resource utilization; S4: The system obtains new real-time data at predetermined time intervals, preprocesses the data and updates the LSTM prediction results. If it detects that the soil moisture exceeds the set range or the photovoltaic power is lower than the threshold, the Q-learning strategy is re-run to adjust the irrigation parameters to ensure the continuous stability of water regulation; S5: After each irrigation cycle, the soil moisture changes, irrigation water volume and energy consumption data of each layer of the slope are recorded and input into the LSTM and Q-learning models as a new round of training data to continuously optimize the water regulation strategy and achieve adaptive learning.

7. The slope greening moisture control method according to claim 6, characterized in that: In S1, the preprocessed data includes standardization and denoising of soil moisture data, meteorological data and slope data to improve the accuracy of model prediction; soil moisture data at different depths are normalized and converted into a unified scale to eliminate the differences between different soil layers; random noise in the measurement of meteorological sensors is removed by using filters to ensure the stability of meteorological data; the slope value measured by the slope sensor is standardized to avoid the influence of different slope ranges on model training.

8. The slope greening moisture control method according to claim 7, characterized in that: In step S2, the data used by the LSTM model include historical soil moisture data, historical meteorological data, and historical irrigation data. The LSTM model captures long-term dependencies in time series through long short-term memory units (LSTM units) to predict future water demand. The training formula of the LSTM model is as follows: h t =σ(W h ·[h t-1 ,x t ]+b h ) Among them, h t is the hidden state of the current time step, C t is the cell state at the current time step, x t is the input of the current time step, f t For the forget gate, i t is the input gate, C ～ t is the candidate cell state, σ represents the activation function, and * represents the element-wise multiplication operation.

9. The slope greening moisture control method according to claim 6, characterized in that: The adaptive control strategy generated by the reinforcement learning algorithm (Q-learning) in step S3 is updated by the following formula: Among them, Q(s t ,a t ) is state s t Take action a t The quality value of α is the learning rate, r t is the current reward value, γ is the discount factor, s t+1 is the next state, a t+1 For the next action.

10. The slope greening moisture control method according to claim 6, characterized in that: In step S4, if it is detected that the system energy is insufficient or the soil moisture exceeds the set range, the intelligent control module dynamically adjusts the irrigation parameters according to the Q-learning strategy and re-optimizes the training data of the LSTM model to ensure timely updating of the water control strategy and optimization of resource utilization.

Images