Intelligent monitoring method of smart greenhouse based on multi-dimensional environmental perception and collaborative control

Abstract

The present invention relates to the technical field of agricultural Internet of Things, and specifically to a smart greenhouse intelligent monitoring method based on multi-dimensional environmental perception and collaborative control, which specifically comprises: real-time collection of environmental parameter data inside and outside the flower greenhouse to determine the environmental conditioning equipment that needs to be regulated; establishing an equipment priority matrix and a hybrid energy switching strategy for the environmental conditioning equipment, and realizing power supply to the environmental conditioning equipment through the hybrid energy switching strategy; constructing an equipment aging prediction model to obtain equipment aging prediction results, and combining an economy and health dual-objective optimization model constructed with a multi-objective optimization algorithm to optimize the start and stop timing of the environmental conditioning equipment and obtain an equipment operation log; generating energy-saving and consumption-reducing operation instructions based on the hybrid energy switching strategy and the Pareto optimal solution set and issuing them to a device control module, which controls the environmental conditioning equipment according to the received energy-saving and consumption-reducing operation instructions.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural Internet of Things, and in particular to a smart monitoring method for a smart greenhouse based on multi-dimensional environmental perception and collaborative control. Background Art

[0002] With the rapid development of global facility agriculture, intelligent greenhouses have become a key area of ​​transformation and upgrading for the floriculture industry. Traditional floriculture relies on extensive management and is significantly constrained by climatic conditions. Especially in the context of frequent extreme weather and rising labor costs, the development of environmentally controllable smart greenhouses has become an urgent need in the industry. While current mainstream smart greenhouse systems have implemented basic environmental parameter monitoring and mechanical control functions, they still face multiple technical bottlenecks in practical application, as follows:

[0003] First, existing solutions often focus on regulating a single environmental factor, such as independently controlling temperature, humidity, or light intensity. This ignores the synergistic effects of multiple parameters during plant growth, including soil moisture, gas composition, and light quality. Alternatively, fans may be activated only when the greenhouse temperature exceeds 30°C, ignoring the dynamic correlation between internal and external environmental parameters. When high outdoor temperatures coexist with high humidity inside the greenhouse, simply turning on the fans can exacerbate plant transpiration, leading to wilting. Furthermore, scientific control strategies fail to incorporate multi-dimensional data such as carbon dioxide concentration and light intensity. This one-sided monitoring system leads to a misalignment between control decisions and the crop's true physiological needs, making it difficult to meet the technical requirements of precision cultivation.

[0004] Second, the existing system has not yet established a scientific linkage protocol for equipment coordination. Key equipment such as water curtains, fans, and sunshades often operate independently, which can neither form a synergistic effect of cooling and dehumidification, nor may the service life be shortened due to frequent start-up and shutdown of the equipment.

[0005] Third, existing control systems generally rely on static preset thresholds, relying on manual experience to set fixed parameter thresholds. This lacks the ability to adapt to the dynamic demands of crops during different growth stages. When faced with sudden environmental fluctuations, system response delays can reach several minutes to tens of minutes, severely impacting flower quality and yield stability.

[0006] Fourth, current systems have significant shortcomings in energy efficiency. Statistics show that energy consumption for traditional greenhouse lighting and temperature control equipment accounts for 40%-60% of total operating costs. Existing solutions often rely on simple timed on / off strategies, failing to effectively integrate energy storage device scheduling with electricity price fluctuations. While some advanced systems incorporate energy management systems, their optimization algorithms are often based on linear programming models, making them incapable of handling complex operating conditions such as nonlinear attenuation of light intensity and aging equipment, leading to widespread energy waste.

[0007] In order to meet these challenges, it is urgent to develop an intelligent monitoring method for smart greenhouses based on multi-dimensional environmental perception and collaborative control to improve the growth efficiency of flowers while reducing operating costs. Summary of the Invention

[0008] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a smart greenhouse intelligent monitoring method based on multi-dimensional environmental perception and collaborative control. The smart greenhouse intelligent monitoring method can monitor the environmental parameter data inside and outside the flower greenhouse in real time, realize equipment linkage collaborative control and dynamic regulation of the operating status of the equipment, and optimize energy configuration.

[0009] The technical solution of the present invention to solve the above technical problems is:

[0010] A smart greenhouse intelligent monitoring method based on multi-dimensional environmental perception and collaborative control includes the following steps:

[0011] Step 1: Real-time collection of environmental parameter data inside and outside the flower greenhouse;

[0012] Step 2: Based on the collected environmental parameter data inside and outside the flower greenhouse, determine the environmental conditioning equipment that needs to be regulated;

[0013] Step 3: Based on the PV panel output power, the energy storage battery state of charge, and grid electricity price fluctuation data, establish a device priority matrix and hybrid energy switching strategy for the environmental conditioning equipment. This hybrid energy switching strategy enables dynamic coordinated power supply for the environmental conditioning equipment through PV discharge, energy storage discharge, and grid power supply.

[0014] Step 4: Construct an equipment aging prediction model and obtain equipment aging prediction results based on the real-time operating parameters of the environmental control equipment. Based on the equipment aging prediction results, a multi-objective optimization algorithm is used to construct an economic and health dual-objective optimization model with the optimization objectives of minimizing energy consumption costs and minimizing equipment life loss, and with equipment operation logic constraints and grid capacity constraints as constraints. The start and stop timing of the environmental control equipment is optimized using this economic and health dual-objective optimization model to obtain a Pareto optimal solution set that takes into account both short-term energy consumption costs and equipment life loss, and this solution is used as the equipment operation log.

[0015] Step 5: Based on the hybrid energy switching strategy and the Pareto optimal solution set, energy-saving and consumption-reduction operation instructions are generated and sent to the equipment control module. The equipment control module controls the environmental conditioning equipment according to the received energy-saving and consumption-reduction operation instructions.

[0016] Preferably, in step 1, environmental parameter data inside the flower greenhouse are collected in real time through the monitoring unit inside the greenhouse; the environmental parameter data inside the flower greenhouse include soil temperature, soil moisture, air temperature, air humidity, carbon dioxide concentration and light intensity inside the flower greenhouse; environmental parameter data outside the flower greenhouse are collected in real time through the monitoring unit outside the greenhouse; the environmental parameter data outside the flower greenhouse include wind direction, wind speed, light intensity, air temperature, air humidity, rainfall and atmospheric pressure outside the flower greenhouse.

[0017] Preferably, in step 2, the difference between the corresponding environmental parameter data inside and outside the flower greenhouse is calculated, and the difference is compared with the corresponding threshold value, and the corresponding environmental adjustment device is selected for control based on the comparison result.

[0018] Preferably, the environmental conditioning equipment is one or more of fan equipment, air conditioning equipment, water curtain equipment, sunshade equipment, fill light equipment, and irrigation equipment.

[0019] Preferably, in step 3, the steps for establishing the device priority matrix are: obtaining the priority score of each environmental conditioning device by respectively calculating the sum of the product of the device power consumption weight, response delay tolerance, device importance and the corresponding weight coefficient of each environmental conditioning device, and ranking the priority of each environmental conditioning device according to the priority score of each environmental conditioning device, thereby obtaining the device priority matrix.

[0020] Preferably, in step 3, the hybrid energy switching strategy includes the following steps:

[0021] Step 301: Collecting in real time the output power of photovoltaic panels, the state of charge of energy storage batteries, grid electricity prices, device load requirements, and grid contract capacity, sorting the environmental conditioning devices in descending order according to the device priority matrix, and allocating energy to each environmental conditioning device in the order of arrangement;

[0022] Step 302: Prioritize the photovoltaic power generation direct supply phase, using photovoltaic power generation to power various environmental conditioning devices; if photovoltaic power generation is insufficient, calculate the first load gap of the environmental conditioning devices that have not yet been allocated energy;

[0023] Step 303: Entering the energy storage power supply phase, dynamically adjusting the energy storage intervention threshold for environmental conditioning devices that have not yet been allocated energy based on device priority, and filling the first load gap with energy storage power generation; if the energy storage power generation is insufficient to fill the first load gap, calculating the second load gap for the environmental conditioning devices that have not yet been allocated energy;

[0024] Step 304: Entering the grid power supply phase, within the constraints of the grid contract capacity, the second load gap is filled by the grid power supply. If the grid power supply exceeds the contract value, the environmental conditioning devices with lower device priorities are shut down in reverse order until the load is in compliance.

[0025] Step 305: After completing the energy distribution to all environmental conditioning devices, the power supply gap ratio between the actual power supply and the demand of all environmental conditioning devices is detected. If the power supply gap ratio exceeds the first preset threshold, the emergency protocol is triggered to forcibly increase the energy storage discharge power and suspend the environmental conditioning device with the lowest equipment priority until the power supply gap ratio is less than or equal to the first preset threshold.

[0026] Preferably, in step 305, if the power supply shortage rate is detected to exceed the first preset threshold for three consecutive times, it is determined that a power supply crisis has occurred and the system needs to automatically activate the emergency protocol; when the power supply shortage rate is detected to be less than the second preset threshold for three consecutive times, the emergency protocol is exited; the emergency protocol is triggered at most twice per hour.

[0027] Preferably, in step 4, the dual-objective optimization model of economy and health is:

[0028] ;

[0029] ;

[0030] Where: is the operating power of device i in time period t (unit: kW); is the time interval; is the time-of-use electricity price in period t; is a binary variable that represents the operating status of device i in time period t, where 1 indicates that the device is in the operating state and 0 indicates that the device is in the shutdown state; is the single startup loss cost of equipment j; is a binary variable representing the startup action identifier of device i in time period t, where 1 indicates that the device is triggered to start in time period t, and 0 indicates that the device does not perform the startup action; is a binary variable representing the startup action identifier of device j in time period t, where 1 indicates that the device is triggered to start in time period t, and 0 indicates that the device does not perform the startup action; is the start-stop loss coefficient, that is, the loss coefficient of equipment i when it is started once; is the operating loss coefficient, that is, the loss coefficient of equipment i when it runs for one hour; is the minimum continuous running time of device i; k is the time period index, which is used for sliding calculation of the time window; M is the total number of environmental control devices.

[0031] Preferably, the method further comprises step 6, wherein the step 6 is:

[0032] A three-layer long short-term memory network is used to construct an energy consumption prediction model. Based on historical energy consumption data, real-time environmental parameter data and equipment operation logs, the energy consumption prediction model is used to predict the energy consumption curve for the next 24 hours.

[0033] Preferably, in step 6, the energy consumption prediction model is:

[0034] ;

[0035] Where: is the energy consumption value from the predicted period t+1 to t+H, where H=288; is the LSTM network function; is the historical energy consumption value at time k; is the ambient temperature in the greenhouse at time k; is the ambient humidity in the greenhouse at time k; is the carbon dioxide concentration in the greenhouse at time k; is the light intensity in the greenhouse at time k; Encode the device status at time k; is the time-of-use electricity price coefficient at time k; is the equipment aging index at time k.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. The intelligent monitoring method for smart flower sheds based on multi-dimensional environmental perception and collaborative control of the present invention selects the corresponding environmental conditioning equipment for regulation according to the real-time collected environmental parameter data inside and outside the flower shed; by establishing an equipment priority matrix and a hybrid energy switching strategy for the environmental conditioning equipment, energy distribution is realized for the environmental conditioning equipment, thereby realizing the coordinated power supply of photovoltaic power supply, energy storage power supply and power grid to the environmental conditioning equipment to be regulated; combined with the prediction results of the equipment aging prediction module, a multi-objective optimization algorithm is used to solve the constructed economic and health dual-objective optimization model to obtain the Pareto optimal solution set (equipment operation log), and according to the hybrid energy switching strategy and the Pareto optimal solution set, energy-saving and consumption-reducing operation instructions are generated and issued to the equipment control module, and the equipment control module regulates the environmental conditioning equipment according to the received energy-saving and consumption-reducing operation instructions; experiments show that the intelligent monitoring method for smart flower sheds of the present invention can reduce energy consumption costs by more than 40%, improve the stability of the flower growth environment, and is suitable for modern facility agriculture scenarios.

[0038] 2. The intelligent monitoring method for smart greenhouses based on multi-dimensional environmental perception and collaborative control of the present invention can monitor the environmental parameter data inside and outside the flower greenhouse in real time, realize the coordinated control of equipment linkage and dynamically adjust the operating status of equipment, and optimize energy configuration, thereby achieving the goal of reducing energy consumption costs while greatly extending the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a structural block diagram of the smart greenhouse intelligent monitoring system based on multi-dimensional environmental perception and collaborative control of the present invention.

[0040] Figure 2 Flowchart of the hybrid energy switching strategy.

[0041] Figure 3 This is the network structure diagram of the equipment aging prediction model. DETAILED DESCRIPTION

[0042] The present invention will be described in further detail below with reference to the embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0043] See also Figure 1 and Figure 2 The intelligent monitoring system of the smart greenhouse based on multi-dimensional environmental perception and collaborative control of the present invention includes a multi-source environmental monitoring module, an equipment control module and an energy optimization management module.

[0044] See also Figure 1 and Figure 2 The multi-source environmental monitoring module is used to collect real-time environmental parameter data inside and outside the flower greenhouse, and regulate one or more of the fan equipment, air conditioning equipment, water curtain equipment, sunshade equipment, fill light equipment, and irrigation equipment according to the collected environmental parameter data inside and outside the flower greenhouse; the multi-source environmental monitoring module includes an indoor monitoring unit and an outdoor monitoring unit, wherein,

[0045] The greenhouse monitoring unit is used to collect environmental parameter data inside the flower greenhouse in real time. The greenhouse monitoring unit adopts a distributed architecture and includes four slave devices, which are respectively located in four areas of the flower greenhouse and are used to comprehensively monitor the environmental parameter data inside the flower greenhouse. Each slave device in the greenhouse monitoring unit includes a soil temperature and humidity sensor, an air temperature and humidity sensor, a carbon dioxide concentration sensor, and a light intensity sensor. The collected environmental parameter data inside the flower greenhouse include soil temperature, soil humidity, air temperature, air humidity, carbon dioxide concentration, and light intensity in the flower greenhouse.

[0046] The outside-shed monitoring unit is used to collect environmental parameter data outside the flower greenhouse in real time. The outside-shed monitoring unit includes an air temperature and humidity sensor, a wind direction and speed sensor, a light intensity sensor, a carbon dioxide concentration sensor, a rainfall sensor and an atmospheric pressure sensor; the collected environmental parameter data outside the flower greenhouse include wind direction, wind speed, light intensity, air temperature, air humidity, rainfall and atmospheric pressure outside the flower greenhouse.

[0047] See also Figure 1 and Figure 2 The equipment control module is used to control the environmental control equipment according to the environmental parameter data obtained by the multi-source environmental monitoring module, the growth status of the flowers and the feedback information obtained by the energy optimization management module, and includes a light intensity control submodule, a soil temperature and humidity control submodule, an air temperature and humidity control submodule and a carbon dioxide concentration control submodule, wherein the light intensity control submodule is used to control the light intensity in the flower greenhouse; the soil temperature and humidity control submodule is used to control the soil temperature and humidity in the flower greenhouse; the air temperature and humidity control submodule is used to control the air temperature and humidity in the flower greenhouse; the carbon dioxide concentration control submodule is used to control the carbon dioxide concentration in the flower greenhouse;

[0048] In this embodiment, the environmental control equipment includes fan equipment, air-conditioning equipment, water curtain equipment, sunshade equipment, fill light equipment and irrigation equipment. The difference between the corresponding environmental parameter data inside and outside the flower greenhouse can be calculated, and the difference can be compared with the corresponding threshold value. The corresponding environmental control equipment is selected for control based on the comparison result. For example, when the temperature difference between the inside and outside of the flower greenhouse is greater than 5°C, the air-conditioning equipment is turned on; when the temperature difference between the inside and outside of the flower greenhouse is between 0-5°C, the fan equipment and water curtain equipment are turned on.

[0049] See also Figure 1 and Figure 2 The energy optimization management module includes a dynamic power allocation unit, an intelligent scheduling unit and an energy efficiency evaluation unit.

[0050] See also Figure 1 and Figure 2 The dynamic power allocation unit is used to establish a device priority matrix and a hybrid energy switching strategy based on the output power of the photovoltaic panels, the charge state of the energy storage battery, and the grid electricity price fluctuation data, so as to realize the dynamic coordinated power supply of photovoltaic discharge, energy storage discharge and grid power supply to the environmental conditioning equipment, so as to ensure that key equipment (i.e., environmental conditioning equipment with high equipment priority) can still receive priority protection when the power supply fluctuates.

[0051] In this embodiment, the device priority matrix is ​​obtained by calculating the device priority ranking of each environmental control device through three dimensions: device power consumption weight (energy consumption ratio), response delay tolerance (maximum power interruption time allowed for the device), and device importance (quantification of the importance of the device to regulating the flower greenhouse environment).

[0052] ;

[0053] Where: is the weight coefficient, where ; The normalized weight of the device power consumption to the total system load, ranging from 0 to 1; The maximum power interruption time allowed for the device (unit: s); Quantify the importance of equipment in regulating the flower greenhouse environment;

[0054] in, Based on the three technical dimensions of irreplaceability of environmental control, severity of failure consequences, and collaborative dependence, a scientific grading is achieved through weighted scoring. The specific method is as follows:

[0055] =0.5×irreplaceability of environmental control score+0.3×severity of failure consequence score+0.2×cooperation dependency score;

[0056] The quantitative standard for the environmental control irreplaceability score is the device's ability to independently maintain the target environmental parameters, with time coverage and parameter accuracy as indicators:

[0057] ;

[0058] Where: T is the statistical period (unit: hours); is the indicator function of environmental parameter compliance when the equipment is independently controlled at time t (compliance = 1, noncompliance = 0); The actual deviation of the parameter from the target value when the equipment is not running; The parameter is allowed to deviate from the threshold (such as the temperature threshold ±2°C);

[0059] The quantitative standard for the severity score of the fault consequences is the correlation between the duration of environmental parameters exceeding the limit caused by equipment failure and the physiological damage function of flowers;

[0060] ;

[0061] Where: is the yield loss rate of the kth type of flowers due to parameter exceeding the limit (based on agricultural experimental data); It is the benchmark value for normal flower production; the statistics of over-limit duration are derived from the historical fault database.

[0062] The quantitative standard for the collaborative dependency score is the additional energy consumption ratio when the equipment is operating independently (the increase in energy consumption caused by the need for complementary regulation of the linked equipment);

[0063] ;

[0064] Where: Energy consumption required for the equipment to operate independently and achieve environmental goals; It is the total energy consumption of the device and its associated devices when operating together.

[0065] Score The higher the value, the higher the device priority. The smaller it is; ultimately, the device priority matrix is:

[0066]

[0067] The hybrid energy switching strategy realizes intelligent decision-making through a multi-level energy buffer mechanism. Photovoltaic power generation combines energy storage discharge with grid power supply to form a three-level energy security system. Among them, the multi-level energy real-time buffer mechanism strategy is as follows: Figure 2 As shown, the entire process is started in a minute-level cycle, specifically:

[0068] Real-time collection of photovoltaic panel output power, energy storage battery state of charge (SOC), grid electricity prices, equipment load requirements, and grid contract capacity. The system then sorts the environmental control equipment in descending order based on the equipment priority matrix (e.g., fill light equipment > air conditioning equipment > water curtain equipment, etc.), and allocates energy to each environmental control equipment in descending order.

[0069] Prioritize the direct photovoltaic power generation phase, using photovoltaic power generation to power various environmental conditioning devices. If photovoltaic power generation is insufficient, calculate the first load gap of environmental conditioning devices that have not yet been allocated energy.

[0070] Specifically, in the photovoltaic power generation direct supply stage, photovoltaic power generation is allocated to environmental conditioning equipment with high equipment priority first. The photovoltaic power allocation of equipment i is :

[0071] ;

[0072] Where: is the power requirement of device i (unit: kW); is the current remaining available photovoltaic output (unit: kW), , where k is the number of devices that have been allocated PV power;

[0073] During the energy storage power supply phase, the energy storage intervention threshold for environmental conditioning devices that have not yet been allocated energy is dynamically adjusted based on device priority (priority 1 devices are allowed to have an SOC as low as 15%, while priority 6 devices require an SOC of ≥25%). Energy storage power generation is used to fill the first load gap. If energy storage power generation is insufficient to fill the first load gap, a second load gap is calculated for the environmental conditioning devices that have not yet been allocated energy.

[0074] Specifically, during the energy storage discharge phase, the system will determine whether to enable the energy storage battery to power the corresponding environmental conditioning device based on the dynamic SOC threshold, where the energy storage power supply of device i is :

[0075] ;

[0076] Where: is the current available power of the energy storage system (considering 10% safety redundancy), ,in, is the total capacity of the energy storage system; is the unmet power demand of device i, i.e., the remaining value after PV power supply allocation; is the energy storage power supply intervention SOC threshold of device i (unit: kW), ,in, is the priority of device i.

[0077] Finally, the grid power supply phase begins. Under the grid contract capacity constraint (i.e., real-time accumulation of grid load and verification of whether it exceeds the contract value), the second load gap is filled by grid power supply. If the grid power supply exceeds the contract value, environmental conditioning devices with lower priority (such as irrigation equipment and wind turbines) are shut down in reverse order of device priority until the load is brought into compliance.

[0078] Specifically, during the grid power supply phase, the grid power supply of device i is :

[0079] ;

[0080] Where: The grid contract capacity is the maximum permitted power consumption specified in the contract; is the unmet power demand of device i, i.e., the remaining value after the allocation of photovoltaic power generation and energy storage power generation; The total amount of power currently allocated from the grid.

[0081] After completing the energy distribution for all environmental conditioning devices, the power supply gap ratio between the actual power supply and the demand of all environmental conditioning devices is detected. If the power supply gap ratio exceeds the first preset threshold (5%), the emergency protocol is triggered, the energy storage discharge power is forcibly increased, and the environmental conditioning device with the lowest equipment priority is suspended.

[0082] The formula for power supply gap rate is:

[0083] ;

[0084] Where, is the total power requirement of all devices in the system.

[0085] In this embodiment, the triggering conditions of the emergency protocol are:

[0086] When the system detects that the power shortage rate exceeds the first preset threshold (5%) for three consecutive times, it determines that a power supply crisis has occurred and the system needs to automatically activate the emergency protocol;

[0087] When the system detects that the power supply shortage rate is less than the second preset threshold (2%) for three consecutive times, it exits the emergency protocol;

[0088] The content of the emergency agreement is: starting from the environmental conditioning equipment with the lowest equipment priority, shut down step by step until the overall power supply shortage rate is less than or equal to 5%; in order to reduce damage to the environmental conditioning equipment, the emergency agreement will be triggered at most twice per hour to prevent the environmental conditioning equipment from being frequently started and stopped.

[0089] Finally, the system status is updated and the energy consumption data is fed back into the "Energy Consumption Prediction Model" below to form a closed-loop control of "monitoring-decision-execution-optimization", ensuring that high-priority environmental conditioning equipment continues to receive optimal energy supply under complex operating conditions such as energy fluctuations and grid capacity limitations.

[0090] See also Figure 1 and Figure 2 The intelligent scheduling unit is used to optimize the start and stop timing of energy-consuming equipment by combining the time-of-use electricity price billing strategy with the equipment aging prediction model; and to achieve intelligent arrangement of equipment operation strategies by adopting a multi-objective optimization model and an adaptive decision-making mechanism, specifically:

[0091] The LSTM-GRU hybrid network is used to build a device aging prediction model. The structure of the LSTM-GRU hybrid network is as follows: Figure 3As shown in the figure, the real-time operating parameters of the environmental control equipment (such as power, operating time, etc.) are used as input and sent to the LSTM branch and GRU branch respectively. The LSTM branch captures the long-term degradation characteristics of the environmental control equipment; the GRU branch captures the impact of short-term working conditions on the environmental control equipment. The outputs of the LSTM branch and the GRU branch (LSTM is 256 dimensions, GRU is 64 dimensions) are then sent as input to the spatiotemporal feature fusion layer for splicing to obtain a fused feature vector. The fused feature vector is then sent to the attention weighted layer and the fully connected output layer in sequence to finally obtain the equipment aging prediction result. Based on the equipment aging prediction result, the NSGA-II non-dominated sorting genetic algorithm is used to construct an economic and health dual-objective optimization model with the optimization objectives of minimizing energy consumption cost and minimizing equipment life loss, and the constraints of equipment operation logic constraints and grid capacity constraints:

[0092] ;

[0093] ;

[0094] Where: is the operating power of device i in time period t (unit: kW); is the time interval; is the time-of-use electricity price in period t; is a binary variable that represents the operating status of device i in time period t, where 1 indicates that the device is in the operating state and 0 indicates that the device is in the shutdown state; is the single startup loss cost of equipment j; is a binary variable representing the startup action identifier of device i in time period t, where 1 indicates that the device is triggered to start in time period t, and 0 indicates that the device does not perform the startup action; is a binary variable representing the startup action identifier of device j in time period t, where 1 indicates that the device is triggered to start in time period t, and 0 indicates that the device does not perform the startup action; is the start-stop loss coefficient, that is, the loss coefficient of equipment i when it is started once; is the operating loss coefficient, that is, the loss coefficient of equipment i when it runs for one hour; is the minimum continuous running time of device i; k is the time period index, which is used for sliding calculation of the time window; M is the total number of environmental control devices.

[0095] By using SBX crossover and dynamic mutation operators to evolve the population, a Pareto optimal solution set that takes into account both short-term energy consumption costs and equipment life loss is obtained, and this is used as the equipment operation log; wherein, the Pareto optimal solution set includes the equipment operation status matrix, economic target value and health target value.

[0096] According to the hybrid energy switching strategy and the Pareto optimal solution set, energy-saving and consumption-reduction operation instructions are generated and sent to the equipment control module. The equipment control module controls the environmental conditioning equipment according to the received energy-saving and consumption-reduction operation instructions.

[0097] By adopting this hierarchical decision-making mechanism, the system can comprehensively consider multiple constraints to dynamically adjust equipment operating parameters to reduce ineffective energy consumption and extend equipment life, thereby reducing operating costs.

[0098] See also Figure 1 and Figure 2 The energy efficiency evaluation unit uses a prediction model to predict the energy consumption curve for the next 24 hours. In this embodiment, the energy efficiency evaluation unit provides a data-driven feedback loop for continuous system optimization. The energy consumption prediction model, based on a long short-term memory (LSTM) network, integrates historical energy consumption data, real-time environmental parameters, and equipment operation logs to generate a 24-hour energy consumption curve forecast, ensuring a mean absolute percentage error (MAPE) of less than 8%. The prediction results of the energy consumption prediction model can be used to guide dynamic power allocation and intelligent scheduling decisions. They can also trigger the policy verification engine to perform counterfactual simulations. For example, by adjusting the threshold for the on-state of fill lights, the relationship between energy consumption and production changes can be predicted, providing a decision-making basis for long-term optimization. The system outputs short-term instructions such as shutting down idle equipment, while long-term recommendations include specific measures such as equipment replacement or parameter tuning.

[0099] In this embodiment, the energy efficiency evaluation unit uses a three-layer long short-term memory network LSTM to construct an energy consumption prediction model. The three-layer structure includes: output layer, hidden layer, and output layer. Its mathematical expression is:

[0100] ;

[0101] Where: is the energy consumption value from the predicted period t+1 to t+H, where H=288; is the LSTM network function; the expression is:

[0102] ;

[0103] Where: are the activation values ​​of the forget gate, input gate, and output gate, respectively, and their value range is [0,1]; is the weight matrix (dimension: hidden layer size × input dimension); is the bias vector; Represents Hadamard product (element-wise multiplication); Represents the hidden state at the previous moment; represents the input vector at the current moment; X is the input feature matrix with a dimension of T×D, where T is the time step and D is the feature dimension, and D=8; is the historical energy consumption value at time k (unit: kWh); is the ambient temperature in the greenhouse at time k (unit: °C); is the ambient humidity in the greenhouse at time k (unit: %RH); is the carbon dioxide concentration in the greenhouse at time k (unit: ppm); is the light intensity in the greenhouse at time k (unit: Lux); The device status code at time k (binary bit combination); is the time-of-use electricity price coefficient at time k; is the equipment aging index at time k (unit: kWh);

[0104] In the above process, all historical energy consumption data need to be preprocessed first. The data preprocessing is divided into missing value filling and data normalization.

[0105] The missing value filling is performed using triple exponential smoothing, and the mathematical expression is:

[0106] ;

[0107] Where: Indicates the actual energy consumption value at time t (if any); Indicates the actual energy consumption value at time t-1; represents the smoothed level value at time t (the estimated energy consumption after filling); represents the smoothing level value at time t-1; represents the smoothing coefficient (controls the weight of historical data); represents the trend term update coefficient; represents the trend estimate at time t; represents the trend estimate at time t-1;

[0108] Data normalization uses improved Robust Scaling to normalize each data point in the input feature matrix X. This is to uniformly map features of different dimensions (such as temperature °C, humidity %RH, energy consumption kW, etc.) to a dimensionless numerical range, avoiding interference from dimensional differences on model training and improving the LSTM network's ability to capture feature relationships. The mathematical expression for normalization is:

[0109] ;

[0110] Where: x represents a specific value in the input feature matrix X; Represents the 25% quantile of the data set; Represents the 75% quantile of the data set; Indicates the normalized value.

[0111] Finally, the intelligent monitoring method of the smart greenhouse based on multi-dimensional environmental perception and collaborative control of the present invention includes the following steps:

[0112] Step 1: Real-time collection of environmental parameter data inside and outside the flower greenhouse;

[0113] Step 2: Based on the collected environmental parameter data inside and outside the flower greenhouse, determine the environmental conditioning equipment that needs to be regulated;

[0114] Step 3: Based on the PV panel output power, the energy storage battery state of charge, and grid electricity price fluctuation data, establish a device priority matrix and hybrid energy switching strategy for the environmental conditioning equipment. This hybrid energy switching strategy enables dynamic coordinated power supply for the environmental conditioning equipment through PV discharge, energy storage discharge, and grid power supply.

[0115] Step 4: Construct an equipment aging prediction model and obtain equipment aging prediction results based on the real-time operating parameters of the environmental control equipment. Based on the equipment aging prediction results, a multi-objective optimization algorithm is used to construct an economic and health dual-objective optimization model with the optimization objectives of minimizing energy consumption costs and minimizing equipment life loss, and with equipment operation logic constraints and grid capacity constraints as constraints. The start and stop timing of the environmental control equipment is optimized using this economic and health dual-objective optimization model to obtain a Pareto optimal solution set that takes into account both short-term energy consumption costs and equipment life loss, and this solution is used as the equipment operation log.

[0116] Step 5: Based on the hybrid energy switching strategy and the Pareto optimal solution set, energy-saving and consumption-reduction operation instructions are generated and sent to the equipment control module. The equipment control module controls the environmental conditioning equipment according to the received energy-saving and consumption-reduction operation instructions.

[0117] The above is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited to the above content. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A smart greenhouse intelligent monitoring method based on multi-dimensional environmental perception and collaborative control, characterized by: The following steps are involved: Step 1: Real-time collection of environmental parameter data inside and outside the flower greenhouse; Step 2: Based on the collected environmental parameter data inside and outside the flower greenhouse, determine the environmental conditioning equipment that needs to be regulated; Step 3: Based on the output power of the photovoltaic panels, the state of charge of the energy storage batteries, and grid electricity price fluctuation data, a device priority matrix and a hybrid energy switching strategy are established for the environmental conditioning devices. The hybrid energy switching strategy is used to achieve dynamic coordinated power supply for the environmental conditioning devices through photovoltaic discharge, energy storage discharge, and grid power supply. The device priority matrix is ​​established by calculating the sum of the product of the device power consumption weight, response delay tolerance, device importance, and the corresponding weight coefficient for each environmental conditioning device to obtain the priority score for each environmental conditioning device. The priority scores of each environmental conditioning device are then used to prioritize the environmental conditioning devices, thereby obtaining the device priority matrix. Step 4: Construct an equipment aging prediction model and obtain equipment aging prediction results based on the real-time operating parameters of the environmental control equipment. Based on the equipment aging prediction results, a multi-objective optimization algorithm is used to construct an economic and health dual-objective optimization model with the optimization objectives of minimizing energy consumption costs and minimizing equipment life loss, and with equipment operation logic constraints and grid capacity constraints as constraints. The start and stop timing of the environmental control equipment is optimized using this economic and health dual-objective optimization model to obtain a Pareto optimal solution set that takes into account both short-term energy consumption costs and equipment life loss, and this solution is used as the equipment operation log. Step 5: Based on the hybrid energy switching strategy and the Pareto optimal solution set, energy-saving and consumption-reduction operation instructions are generated and sent to the equipment control module. The equipment control module controls the environmental conditioning equipment according to the received energy-saving and consumption-reduction operation instructions.

2. The intelligent monitoring method for a smart greenhouse based on multi-dimensional environmental perception and collaborative control according to claim 1 is characterized in that: In step 1, the environmental parameter data inside the flower greenhouse are collected in real time through the monitoring unit inside the greenhouse; the environmental parameter data inside the flower greenhouse include soil temperature, soil moisture, air temperature, air humidity, carbon dioxide concentration and light intensity inside the flower greenhouse; the environmental parameter data outside the flower greenhouse are collected in real time through the monitoring unit outside the greenhouse; the environmental parameter data outside the flower greenhouse include wind direction, wind speed, light intensity, air temperature, air humidity, rainfall and atmospheric pressure outside the flower greenhouse.

3. The intelligent monitoring method for a smart greenhouse based on multi-dimensional environmental perception and collaborative control according to claim 2 is characterized in that: In step 2, the difference between the corresponding environmental parameter data inside and outside the flower greenhouse is calculated, and the difference is compared with the corresponding threshold value. The corresponding environmental adjustment device is selected for control based on the comparison result.

4. The intelligent monitoring method for a smart greenhouse based on multi-dimensional environmental perception and collaborative control according to claim 3 is characterized in that: The environmental conditioning equipment is one or more of fan equipment, air conditioning equipment, water curtain equipment, sunshade equipment, fill light equipment, and irrigation equipment.

5. The intelligent monitoring method for a smart greenhouse based on multi-dimensional environmental perception and collaborative control according to claim 4 is characterized in that: In step 3, the hybrid energy switching strategy includes the following steps: Step 301: Collecting in real time the output power of photovoltaic panels, the state of charge of energy storage batteries, grid electricity prices, device load requirements, and grid contract capacity, sorting the environmental conditioning devices in descending order according to the device priority matrix, and allocating energy to each environmental conditioning device in the order of arrangement; Step 302: Prioritize the photovoltaic power generation direct supply phase, using photovoltaic power generation to power various environmental conditioning devices; if photovoltaic power generation is insufficient, calculate the first load gap of the environmental conditioning devices that have not yet been allocated energy; Step 303: Entering the energy storage power supply phase, dynamically adjusting the energy storage intervention threshold for environmental conditioning devices that have not yet been allocated energy based on device priority, and filling the first load gap with energy storage power generation; if the energy storage power generation is insufficient to fill the first load gap, calculating the second load gap for the environmental conditioning devices that have not yet been allocated energy; Step 304: Entering the grid power supply phase, within the constraints of the grid contract capacity, the second load gap is filled by the grid power supply. If the grid power supply exceeds the contract value, the environmental conditioning devices with lower device priorities are shut down in reverse order until the load is in compliance. Step 305: After completing the energy distribution to all environmental conditioning devices, the power supply gap ratio between the actual power supply and the demand of all environmental conditioning devices is detected. If the power supply gap ratio exceeds the first preset threshold, the emergency protocol is triggered to forcibly increase the energy storage discharge power and suspend the environmental conditioning device with the lowest equipment priority until the power supply gap ratio is less than or equal to the first preset threshold.

6. The intelligent monitoring method for smart greenhouses based on multi-dimensional environmental perception and collaborative control according to claim 5 is characterized in that: In step 305, if the power supply shortage rate is detected to exceed the first preset threshold for three consecutive times, it is determined that a power supply crisis has occurred and the system needs to automatically activate the emergency protocol; when the power supply shortage rate is detected to be less than the second preset threshold for three consecutive times, the emergency protocol is exited; the emergency protocol is triggered at most twice per hour.

7. The intelligent monitoring method for a smart greenhouse based on multi-dimensional environmental perception and coordinated control according to claim 6 is characterized in that: The method further comprises step 6, wherein the step 6 is: A three-layer long short-term memory network is used to construct an energy consumption prediction model. Based on historical energy consumption data, real-time environmental parameter data and equipment operation logs, the energy consumption prediction model is used to predict the energy consumption curve for the next 24 hours.

Images