Electrically intelligent environment monitoring system

Abstract

The application discloses an electrical intelligent environment monitoring system, comprising a training set acquisition module, which is used for acquiring a historical parameter training set of a greenhouse environment; a model training module, which is used for inputting the historical parameter training set into a pre-training model to obtain a parameter prediction model through training; a first index module, which is used for acquiring a current environment parameter point vector of the greenhouse environment and inputting the current environment parameter point vector into the parameter prediction model to output a double-target benefit prediction index; a second index module, which is used for acquiring a plant growth speed observation value and an energy consumption observation value of the greenhouse environment and determining a double-target benefit current index according to the observation values; and an environment optimal value acquisition module, which is used for determining a plant growth speed optimal value and an energy consumption optimal value according to the double-target benefit prediction index and the double-target benefit current index. The application ensures that plants grow under optimal growth conditions and energy consumption is minimized.

Description

Technical Field

[0001] This application relates to the field of intelligent environmental monitoring, specifically to an electrical intelligent environmental monitoring system. Background Technology

[0002] With the development of modern agriculture, greenhouse cultivation technology has been widely applied around the world. Monitoring and controlling greenhouse environmental parameters is crucial for crop growth, and greenhouse electrical environmental control equipment, such as temperature and humidity regulators, lighting fixtures, and soil moisture controllers, can effectively regulate key environmental parameters such as temperature, humidity, light intensity, and carbon dioxide concentration within the greenhouse. By adjusting and monitoring these environmental parameters, crops can grow under optimal conditions, thereby improving production efficiency. However, current greenhouse environmental control technology faces several challenges, particularly in balancing plant growth rate with energy consumption.

[0003] For example, Chinese patent publication number CN110084417A discloses an intelligent monitoring system for strawberry greenhouse environmental parameters based on a GRNN neural network. This system effectively addresses the shortcomings of existing strawberry greenhouse environmental monitoring systems, which fail to predict yields based on the nonlinearity, large lag, and complex dynamic changes of strawberry greenhouse parameters, thus significantly impacting effective management of greenhouse production. However, current technologies lack control strategies that simultaneously maximize plant growth rate and minimize energy consumption, instead focusing on regulating only a single objective. For instance, adjusting temperature and humidity may significantly increase energy consumption, while reducing energy consumption may negatively affect plant growth rate. This imbalance often leads to resource waste or slow crop growth, hindering the achievement of intelligent monitoring of greenhouse cultivation. Summary of the Invention

[0004] This embodiment provides an intelligent electrical environment monitoring system to explore how to solve the problem of lacking a control strategy that simultaneously considers maximizing plant growth rate and minimizing energy consumption, and can only regulate a single objective.

[0005] This invention provides an intelligent electrical environment monitoring system, comprising:

[0006] The training set acquisition module is used to acquire a training set of historical parameters of the greenhouse environment;

[0007] The model training module is used to input the historical parameter training set into the pre-trained model to train the parameter prediction model;

[0008] The real-time monitoring module is used to monitor the electrical environmental parameters inside the greenhouse in real time.

[0009] The first index module is used to obtain the current environmental parameter point vector of the greenhouse environment, input the current environmental parameter point vector into the parameter prediction model, and output the dual-objective benefit prediction index.

[0010] The second index module is used to acquire observations of plant growth rate and energy consumption in the greenhouse environment, and to determine the current index of dual-objective benefits based on the observations.

[0011] The environmental optimal value acquisition module is used to determine the optimal values ​​for plant growth rate and energy consumption based on the dual-objective benefit prediction index and the current dual-objective benefit index.

[0012] In some embodiments, the acquisition step of the training set acquisition module includes:

[0013] S1-1, Obtain the environmental parameter point vector of each of the T time nodes in the historical database, wherein the historical database stores T time nodes and n environmental parameters corresponding to each time node;

[0014] S1-2, using each time node as an index, construct a dual-objective benefit historical index corresponding to that time node; the dual-objective benefit historical index is used to characterize maximizing plant growth rate while minimizing energy consumption in the historical database;

[0015] S1-3 uses environmental parameter point vectors as feature samples and bi-objective benefit historical index as target labels to construct binary training samples s. t ,

[0016] S t =(X t y h，t );

[0017] Among them, X t It is the vector of environmental parameter points at time node t, y h，t It is the historical index of dual-objective benefits at time node t;

[0018] S1-4, Collect the training samples to obtain the historical parameter training set EPS.

[0019] EPS = { s 1, s2, ..., s T};

[0020] Among them, s T It is the Tth training sample.

[0021] In some embodiments, the environmental parameter point vectors for each of T time points in a historical database are obtained, including:

[0022] S1-1-1, collect n environmental parameters at time node t;

[0023] S1-1-2, Substitute time node t into the following formula to obtain the time dependency weight w of time node t in the historical time period T. t,k

[0024]

[0025] Among them, w t，k λ is the time dependency weight of time node t relative to time node k, λ is the decay coefficient used to control the time decay rate, and |tk| is the time distance between time node t and time node k. It involves summing and normalizing the weights of all historical time points to ensure that the sum of the weights of all historical time points equals 1.

[0026] S1-1-3, Preset the environmental attenuation coefficient λ for each environmental parameter i. i Substituting into the following formula, we obtain the time dependency weight w for each environmental parameter i at time node t. t，k，i ;

[0027]

[0028] Among them, w t，k，i λ is the weight of the i-th environmental parameter at time node t relative to time node k. i It is the attenuation coefficient of environmental parameters, used to characterize the time sensitivity of environmental parameters;

[0029] S1-1-4, Obtain the parameter value X of each environmental parameter i at time node k. k，i ;

[0030] S1-1-5, based on the time dependency weight w of each environmental parameter i at time node t. t，k，i And the parameter value X of each environmental parameter i at time node k. k，i Calculate the weighted environmental parameter value X′ at time node t. t，i

[0031]

[0032] The set of all weighted environmental parameter values ​​is represented as the environmental parameter point vector X at time node t. t ;

[0033] X t =[x′ t，1 , x′ t，2 ,...,x′ t，n ];

[0034] Among them, X t It is a vector of environmental parameters at time node t, containing n weighted environmental parameter values.

[0035] In some of these embodiments, a bi-objective benefit historical index is constructed corresponding to each time point, using each time point as an index, including:

[0036] S1-2-1, Using time node t as the index, extract the corresponding plant growth rate G from the historical database. h，t and energy consumption E h，t ;

[0037] S1-2-2, based on the plant growth rate G h，t and energy consumption E h，t Construct a dual-objective benefit historical index y at time node t. h，t ;

[0038]

[0039] 0≤α≤1;

[0040] Among them, G h，t E is the plant growth rate at time point t. h，t G represents the energy consumption at time point t. max It is the maximum plant growth rate in the historical database, E min It is the minimum energy consumption value in the historical database, and α is a weighting coefficient used to control the balance between the objectives of maximizing plant growth rate and minimizing energy consumption.

[0041] In some embodiments, the training steps of the model training module include:

[0042] S2-1, the environmental parameter point vector X from the historical parameter training set. t As input to the pre-trained model, the output is a bi-objective benefit prediction index.

[0043] S2-2, by minimizing the bi-objective benefit prediction index Historical index of dual-objective benefits y h，t The loss between them is used to define a first loss function L(θ), which is used to optimize the parameters of the pre-trained model and obtain the parameter set θ.

[0044] S2-3, Find the optimal parameter θ in the parameter set θ. * As a parameter prediction model M env The model parameters; the optimal parameter θ * The optimized expression is:

[0045]

[0046] The first loss function L(θ) is defined using the mean squared error:

[0047]

[0048] in, It is the bi-objective benefit prediction index output by the parametric prediction model, where T is the total number of time points.

[0049] In some embodiments, the step of the first index module outputting the bi-objective benefit prediction index includes:

[0050] S4-1, Collect environmental parameters at the current time node, and directly combine the environmental parameters at the current time node into a vector of current environmental parameter points;

[0051] S4-2, Receive current environmental parameters through the parameter prediction model, and output the current dual-objective benefit prediction index from the parameter prediction model.

[0052] In some of these embodiments, the dual-objective benefit is currently indices y. o，t The expression is:

[0053]

[0054] Among them, G o，t This is the observed plant growth rate in the current environment, E o，t It represents the current energy consumption figures for the environment.

[0055] In some embodiments, the acquisition step of the environmental optimal value acquisition module includes:

[0056] S6-1, Define the dual-objective benefit prediction index and the current index y of dual-objective benefits o，t The second loss function L(G) between o，t E o，t );

[0057]

[0058] S6-2, Define the optimal index y for the dual-objective effect c，t The exponential formula, and setting G in the exponential formula c，t initial values ​​and E c，t The initial value;

[0059]

[0060] Among them, G c，tThe goal is to find the optimal value for plant growth rate, E. c，t The goal is to find the optimal value for current energy consumption, where G is... c，t initial values ​​and E c，t The initial value can be selected from the current observation value, the mean value in the historical database, or the variance value in the historical database;

[0061] S6-3, Update the optimal value of plant growth rate G using the gradient descent method. c，t And the optimal value of energy consumption E c，t ;

[0062] The optimal value of plant growth rate G c，t The expression for the partial derivative is:

[0063]

[0064] The optimal value of energy consumption E c，t The expression for the partial derivative is:

[0065]

[0066] in, This represents the optimal value G of the second loss function relative to the plant growth rate. c，t rate of change, This represents the optimal value E of the second loss function relative to energy consumption. c，t The rate of change;

[0067] S6-4, Update the optimal value of plant growth rate G using gradient descent. c，t And the optimal value of energy consumption E c，t This continues until the second loss function converges;

[0068] Optimal value of plant growth rate G c，t The update expression is:

[0069]

[0070] Optimal energy consumption E c，t The update expression is:

[0071]

[0072] Where η is the learning rate, and its value is in the range

[10] . -3 10 -2 This is used to control the duration of each update. and These are the optimal plant growth rate values ​​G during the i-th iteration. c，t And the optimal value of energy consumption E c，t ;

[0073] S6-5, when the convergence condition is met, the parameter prediction model stops iterating and outputs the optimal value of plant growth rate G. c，t And the optimal value of energy consumption E c，t .

[0074] Compared with existing technologies, the electrical intelligent environment monitoring system of the present invention, by combining historical data with real-time monitoring, can dynamically and accurately predict and adjust the electrical environment parameters in agricultural greenhouses, ensuring that plants grow under optimal growth conditions while minimizing energy consumption. Attached Figure Description

[0075] Figure 1 A flowchart of an electrical intelligent environment monitoring system provided in the embodiments of the present invention;

[0076] Figure 2 This is a structural block diagram of the intelligent electrical environment monitoring system provided in the embodiments of the present invention;

[0077] Figure 3 A logic diagram of the environmental optimal value acquisition module provided in the embodiments of the present invention; Detailed Implementation

[0078] To better understand the purpose, technical solution, and advantages of this application, the application is described and explained below in conjunction with the accompanying drawings and embodiments.

[0079] Unless otherwise defined, the technical or scientific terms used in this application shall have the general meaning as understood by one of ordinary skill in the art to which this application pertains. Words such as “a,” “an,” “an,” “the,” “the,” and “these,” used in this application, do not indicate quantitative limitation and may be singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or modules (units) is not limited to the listed steps or modules (units) but may include steps or modules (units) not listed, or may include other steps or modules (units) inherent to such processes, methods, products, or devices. The terms “connected,” “linked,” and “coupled,” used in this application, are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect. The term “multiple” used in this application refers to two or more. The "and / or" operator describes the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: A alone, A and B simultaneously, and B alone. Typically, the character " / " indicates that the objects before and after it are in an "or" relationship. The terms "first," "second," and "third," etc., used in this application are merely for distinguishing similar objects and do not represent a specific ordering of the objects.

[0080] Example 1: Reference Figures 1 to 3 , Figure 1 This is a flowchart of the electrical intelligent environment monitoring system of the present invention. The system flow includes:

[0081] Step S1: Obtain a training set of historical parameters for the greenhouse environment.

[0082] In this embodiment, step S1 specifically involves:

[0083] S1-1, obtain the environmental parameter point vector of each of the T time nodes in the historical database. The historical database stores T time nodes and n environmental parameters corresponding to each time node.

[0084] S1-2, using each time node as an index, construct a dual-objective benefit historical index corresponding to that time node; the dual-objective benefit historical index is used to characterize the maximization of plant growth rate while minimizing energy consumption in the historical database;

[0085] S1-3 uses environmental parameter point vectors as feature samples and bi-objective benefit historical indexes as target labels to construct binary training samples S1-3. t ,

[0086] S t =(Xt y h，t );

[0087] Among them, X t It is the vector of environmental parameter points at time node t, y h，t It is the historical index of dual-objective benefits at time node t;

[0088] S1-4, assemble the training samples to obtain the historical parameter training set EPS.

[0089] EPS = {s1, s2, ..., s} T};

[0090] Among them, s T It is the Tth training sample.

[0091] In this embodiment, step S1-1 specifically involves:

[0092] S1-1-1, collect n environmental parameters at time node t;

[0093] S1-1-2, Substitute time node t into the following formula to obtain the time dependency weight of time node t in the historical time period T;

[0094]

[0095] Among them, w t，k λ is the time dependency weight of time node t relative to time node k, λ is the decay coefficient used to control the time decay rate, and |tk| is the time distance between time node t and time node k. It involves summing and normalizing the weights of all historical time points to ensure that the sum of the weights of all historical time points equals 1.

[0096] S1-1-3, Preset the environmental attenuation coefficient λ for each environmental parameter i. i Substituting into the following formula, we obtain the time dependency weight w for each environmental parameter i at time node t. t，k，i ;

[0097]

[0098] Among them, w t，k，i λ is the weight of the i-th environmental parameter at time node t relative to time node k. i It is the attenuation coefficient of environmental parameters, used to characterize the time sensitivity of environmental parameters;

[0099] S1-1-4, Obtain the parameter value x of each environmental parameter i at time node k. k，i ;

[0100] S1-1-5, based on the time dependency weight w of each environmental parameter i at time node t. t，k，i And the parameter value x of each environmental parameter i at time node k. k，i Calculate the weighted environmental parameter value x′ at time node t. t，i ;

[0101]

[0102] The set of all weighted environmental parameter values ​​is represented as the environmental parameter point vector X at time node t. t ;

[0103] X t =[x′ t，1 , x′ t，2 ,...,x′ t，n ];

[0104] Among them, X t It is a vector of environmental parameters at time node t, containing n weighted environmental parameter values.

[0105] In this embodiment, step S1-2 specifically involves:

[0106] S1-2-1, Using time node t as the index, extract the corresponding plant growth rate G from the historical database. h，t and energy consumption E h，t ;

[0107] S1-2-2, based on the plant growth rate G h，t and energy consumption E h，t Construct a dual-objective benefit historical index y at time node t. h，t ;

[0108]

[0109] 0≤α≤1;

[0110] Among them, G h，t E is the plant growth rate at time point t. h，t G represents the energy consumption at time point t. max It is the maximum plant growth rate in the historical database, E min It is the minimum energy consumption value in the historical database, and α is a weighting coefficient used to control the balance between the objectives of maximizing plant growth rate and minimizing energy consumption.

[0111] For example, the relationship between plant growth rate and energy consumption is generally non-linear. For instance, an extreme increase in temperature or light may cause the optimal value of energy consumption to rise sharply, but the improvement in plant growth rate gradually becomes insignificant. The dual-objective benefit history index in this embodiment is a type of weighted geometric mean model, which can better handle this non-linear objective relationship compared to the conventional linear weighted model.

[0112] Furthermore, This section amplifies the effect of plant growth rate, making the plant growth rate as close as possible to the maximum value G. max At that time, the larger the historical index of the dual-objective benefit, the greater the plant growth rate. Similarly, This is used to reduce the impact of the optimal energy consumption value, so that the energy consumption is as close as possible to the minimum value E. min When the historical index of dual-objective benefits is smaller, energy consumption is minimized.

[0113] Furthermore, the weighted geometric mean model is more sensitive to extreme input values. Therefore, when energy consumption is very high, it will have a significant impact on the bi-objective benefit index, thus reflecting the problem of very high energy consumption and making it easier to find a more suitable combination of parameters.

[0114] The process of the electrical intelligent environmental monitoring system in this invention introduces time-dependent weights to construct a dual-objective effect historical index, thereby dynamically measuring the impact of historical data at different time points on the current time point. Compared to simply averaging historical data, the time-dependent weights assign different weights to different time points based on time differences. More recent time points, being closer to the current environmental conditions, have a greater impact on the current plant growth rate and energy consumption, and are therefore given higher weights. Conversely, more distant time points have a relatively smaller impact and are given lower weights. This weight design can effectively capture the time decay characteristics of historical data, enabling the system to more accurately construct a dual-objective benefit historical index.

[0115] Example 2: Reference Figures 1 to 3 The technical solution of this embodiment 2 differs from that of embodiment 1 in that the system process further includes:

[0116] Step S2: Input the historical parameter training set into the pre-trained model to train the parameter prediction model.

[0117] In this embodiment, step S2 specifically involves:

[0118] S2-1, the environmental parameter point vector X from the historical parameter training set. t As input to the pre-trained model, the output is a bi-objective benefit prediction index.

[0119] S2-2, by minimizing the bi-objective benefit prediction index Historical index of dual-objective benefits y h，t The loss between them is used to define a first loss function L(θ), which is used to optimize the parameters of the pre-trained model and obtain the parameter set θ.

[0120] S2-3, Find the optimal parameter θ in the parameter set θ. * As a parameter prediction model M env The model parameters; the optimal parameter θ * The optimized expression is:

[0121]

[0122] The first loss function L(θ) is defined using the mean squared error:

[0123]

[0124] in, It is the bi-objective benefit prediction index output by the parametric prediction model, where T is the total number of time points.

[0125] Step S3: Real-time monitoring of electrical environmental parameters within the greenhouse; In this embodiment, environmental parameters include temperature, humidity, light intensity, soil moisture, and carbon dioxide concentration within the greenhouse; electrical environmental parameters represent parameters within the greenhouse. Since real-time environmental parameters are all related environmental parameters regulated after electrical equipment is used within the greenhouse, in this embodiment, environmental parameters are the parameters reflected after the electrical parameters are executed by the relevant equipment within the greenhouse; electrical equipment may include temperature and humidity regulators, controllable lighting lamps, soil humidifiers, soil dryers, etc.

[0126] Step S4: Obtain the current environmental parameter point vector of the greenhouse environment, input the current environmental parameter point vector into the parameter prediction model, and output the current dual-objective benefit prediction index; wherein, the current dual-objective benefit prediction index is used to characterize the predicted value of maximizing plant growth rate while minimizing energy consumption in the greenhouse environment.

[0127] In this embodiment, step S4 specifically involves:

[0128] S4-1, collect the environmental parameters at the current time node and directly combine the environmental parameters at the current time node into the current environmental parameter point vector; because the parameter prediction model only needs the environmental parameters at the current time node for prediction, and the model objective is to directly output the bi-objective benefit prediction index, the feature vector at the current time node can directly use the actual environmental parameter values, i.e., the current environmental parameters.

[0129] S4-2, receives current environmental parameters through a parameter prediction model, and outputs the current dual-objective benefit prediction index.

[0130] Step S5: Obtain observed values ​​of plant growth rate and energy consumption in the greenhouse environment. These are the directly observable plant growth rate and energy consumption values ​​in the current greenhouse environment, not values ​​calculated from environmental electrical parameters. Based on these observations, determine the current dual-objective benefit index. The current dual-objective benefit index characterizes the actual observed values ​​in the greenhouse environment that maximize plant growth rate while minimizing energy consumption. The current dual-objective benefit index y o，t The expression is:

[0131]

[0132] Among them, G o，t This is the observed plant growth rate in the current environment, E o，t It represents the current energy consumption figures for the environment.

[0133] Step S6: Based on the dual-objective benefit prediction index and the dual-objective benefit current index, determine the optimal value of plant growth rate (i.e., the optimal value of plant growth rate based on environmental parameters at the current time point, which is a predicted value) and the optimal value of energy consumption (i.e., the optimal value of energy consumption based on environmental parameters at the current time point, which is a predicted value).

[0134] refer to Figure 3 In this embodiment, step S6 specifically involves:

[0135] S6-1, Define the dual-objective benefit prediction index and the current index y of dual-objective benefits o，t The second loss function L(G) between o，t E o，t );

[0136]

[0137] S6-2, Define the optimal index y for the dual-objective effect c，t The exponential formula, and setting G in the exponential formula c，t initial values ​​and E c，t The initial value;

[0138]

[0139] Among them, G c，t The goal is to find the optimal value for plant growth rate, E. c，t The goal is to find the optimal value for current energy consumption, where G is... c，t initial values ​​and Ec，t The initial value can be selected from the current observation value, the mean value in the historical database, or the variance value in the historical database;

[0140] S6-3, Update the optimal value of plant growth rate G using the gradient descent method. c，t And the optimal value of energy consumption E c，t ;

[0141] The optimal value of plant growth rate G c，t The expression for the partial derivative is:

[0142]

[0143] The optimal value of energy consumption E c，t The expression for the partial derivative is:

[0144]

[0145] in, This represents the optimal value G of the second loss function relative to the plant growth rate. c，t rate of change, This represents the optimal value E of the second loss function relative to energy consumption. c，t The rate of change. That is, the optimal value of the current plant growth rate G. c，t And the optimal value of energy consumption E c，t Change G c，t or E c，t How will this affect the change in the second loss function?

[0146] S6-4, Update the optimal value of plant growth rate G using gradient descent. c，t And the optimal value of energy consumption E c，t This continues until the second loss function converges;

[0147] Optimal value of plant growth rate G c，t The update expression is:

[0148]

[0149] Optimal energy consumption E c，t The update expression is:

[0150]

[0151] Where η is the learning rate, and its value is in the range

[10] . -3 10 -2 This is used to control the duration of each update. and These are the optimal plant growth rate values ​​G during the i-th iteration. c，tAnd the optimal value of energy consumption E c，t ;

[0152] S6-5, when the convergence condition is met, the parameter prediction model stops iterating and outputs the optimal value of plant growth rate G. c，t And the optimal value of energy consumption E c，t .

[0153] The convergence condition is that the second loss function L(G) is... o，t E o，t The change ΔL is lower than the preset value by 10. -6 Or, the number of iterations in the parameter prediction model has reached the maximum number of iterations n. max n max ∈[1000, 5000].

[0154] Step S7: Adjust the actuators of the electrical equipment according to the optimal values ​​of plant growth rate and energy consumption.

[0155] For example, temperature and humidity regulators are widely used in greenhouses. By controlling electrical actuators such as heating and cooling, they ensure that the temperature and humidity in the greenhouse environment are maintained within a range suitable for plant growth.

[0156] Once the optimal plant growth rate and energy consumption are calculated, the temperature and humidity regulator automatically adjusts the operating mode and time of the heater, air conditioner, humidifier, or dehumidifier based on these target values ​​to ensure that the environmental conditions meet the optimal needs of plant growth.

[0157] Furthermore, in greenhouse cultivation, LED lights or other forms of artificial light sources are often used to supplement natural light and control the photoperiod of plants to improve their growth rate.

[0158] When light parameters are not up to standard, the system adjusts the brightness and operating time of the lamps based on the optimal values ​​for plant growth rate and energy consumption. If the system determines that the plant is not receiving enough light but energy consumption is high, it can automatically adjust the brightness of the lamps to the lowest acceptable range to maintain the lowest energy consumption.

[0159] Of course, the automatic control of electrical equipment actuators based on the optimal input value is existing technology, so it will not be elaborated further.

[0160] The process of the intelligent electrical environment monitoring system in this embodiment combines historical data with real-time monitoring to dynamically and accurately predict and adjust the electrical environment parameters in agricultural greenhouses, ensuring that plants grow under optimal conditions while minimizing energy consumption.

[0161] Its specific effects are divided into:

[0162] By introducing a parametric prediction model and utilizing historical environmental data and real-time collected environmental parameters, a dual-objective benefit index can be accurately predicted: the index that maximizes plant growth rate and minimizes energy consumption under the current environment. Furthermore, by combining actual observed plant growth rate and energy consumption, the optimal values ​​for both are further optimized, ensuring consistency between prediction and practical application, thus improving the accuracy and scientific rigor of the regulation.

[0163] The system optimizes the dual-objective benefit prediction index through gradient descent, which not only finds the optimal balance between plant growth rate and energy consumption, but also dynamically adjusts the actuators of electrical equipment such as temperature and humidity regulators and lighting lamps based on the prediction results, ensuring that the greenhouse environment is always in the optimal state, greatly improving the plant growth rate while minimizing energy consumption.

[0164] The intelligent electrical environment monitoring system provided in this embodiment of the invention is used to execute the above-described system process, and details already described will not be repeated. The terms "module," "unit," and "subunit" used below refer to combinations of software and / or hardware that achieve a predetermined function. Although the system described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0165] like Figure 2 As shown, Figure 2 This is a structural block diagram of the electrical intelligent environment monitoring system of the present invention, which includes:

[0166] The training set acquisition module is used to acquire a training set of historical parameters of the greenhouse environment;

[0167] The model training module is used to input the historical parameter training set into the pre-trained model to train the parameter prediction model;

[0168] The real-time monitoring module is used to monitor the electrical environmental parameters inside the greenhouse in real time.

[0169] The first index module is used to obtain the current environmental parameter point vector of the greenhouse environment, input the current environmental parameter point vector into the parameter prediction model, and output the dual-objective benefit prediction index; wherein, the dual-objective benefit prediction index is used to characterize the predicted value of maximizing plant growth rate while minimizing energy consumption in the greenhouse environment;

[0170] The second index module is used to acquire the observed values ​​of plant growth rate and energy consumption in the greenhouse environment, and to determine the current index of dual objective benefit based on the observed values; wherein, the current index of dual objective benefit is used to characterize the actual observed values ​​of maximizing plant growth rate and minimizing energy consumption in the greenhouse environment.

[0171] The environmental optimal value acquisition module is used to determine the optimal values ​​for plant growth rate and energy consumption based on the dual-objective benefit prediction index and the current dual-objective benefit index.

[0172] The electrical equipment control module is used to adjust the actuators of electrical equipment according to the optimal values ​​of plant growth rate and energy consumption.

[0173] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the 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 the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0174] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent electrical environment monitoring system, characterized in that, include: The training set acquisition module is used to acquire a training set of historical parameters of the greenhouse environment; The model training module is used to input the historical parameter training set into the pre-trained model to train the parameter prediction model; The real-time monitoring module is used to monitor the electrical environmental parameters inside the greenhouse in real time. The first index module is used to obtain the current environmental parameter point vector of the greenhouse environment, input the current environmental parameter point vector into the parameter prediction model, and output the dual-objective benefit prediction index. The second index module is used to acquire observations of plant growth rate and energy consumption in the greenhouse environment, and to determine the current index of dual-objective benefits based on the observations. The environmental optimal value acquisition module is used to determine the optimal values ​​for plant growth rate and energy consumption based on the dual-objective benefit prediction index and the current dual-objective benefit index.

2. The intelligent electrical environment monitoring system according to claim 1, characterized in that, The acquisition steps of the training set acquisition module include: S1-1, Obtain the environmental parameter point vector of each of the T time nodes in the historical database, wherein the historical database stores T time nodes and n environmental parameters corresponding to each time node; S1-2, using each time node as an index, construct a dual-objective benefit historical index corresponding to that time node; the dual-objective benefit historical index is used to characterize maximizing plant growth rate while minimizing energy consumption in the historical database; S1-3 uses environmental parameter point vectors as feature samples and bi-objective benefit historical indexes as target labels to construct binary training samples S1-3. t , S t =(X t ,y h，t ); Among them, X t It is the vector of environmental parameter points at time node t, y h，t It is the historical index of dual-objective benefits at time node t; S1-4, The training samples are combined to obtain the historical parameter training set EPS. EPS={S1,S2,...,S T }; Among them, S T It is the Tth training sample.

3. The intelligent electrical environment monitoring system according to claim 2, characterized in that, Retrieve the environmental parameter point vectors for each of the T time points in the historical database, including: S1-1-1, collect n environmental parameters at time node t; S1-1-2, Substitute time node t into the following formula to obtain the time dependency weight of time node t in the historical time period T; Among them, w t，k λ is the time dependency weight of time node t relative to time node k, λ is the decay coefficient used to control the time decay rate, and |tk| is the time distance between time node t and time node k. It involves summing and normalizing the weights of all historical time points to ensure that the sum of the weights of all historical time points equals 1. S1-1-3, Preset the environmental attenuation coefficient λ for each environmental parameter i. i Substituting into the following formula, we obtain the time dependency weight w for each environmental parameter i at time node t. t，k,i ; Among them, w t，k，i λ is the weight of the i-th environmental parameter at time node t relative to time node k. i It is the attenuation coefficient of environmental parameters, used to characterize the time sensitivity of environmental parameters; S1-1-4, Obtain the parameter value X of each environmental parameter i at time node k. k，i ; S1-1-5, based on the time dependency weight w of each environmental parameter i at time node t. t，k，i And the parameter value X of each environmental parameter i at time node k. k，i Calculate the weighted environmental parameter value X′ at time node t. t，j ; The set of all weighted environmental parameter values ​​is represented as the environmental parameter point vector X at time node t. t ; X t =[X′ t，1 ,X t，2 ,...,X′ t，n ]; Among them, X t It is a vector of environmental parameters at time node t, containing n weighted environmental parameter values.

4. The intelligent electrical environment monitoring system according to claim 3, characterized in that, Using each time point as an index, construct a historical index of dual-objective benefits corresponding to that time point, including: S1-2-1, Using time node t as the index, extract the corresponding plant growth rate G from the historical database. h，t and energy consumption E h，t ; S1-2-2, based on the plant growth rate G h，t and energy consumption E h，t Construct a dual-objective benefit historical index y at time node t. h，t ; O≤α≤1； Among them, G h，t E is the plant growth rate at time point t. h，t G represents the energy consumption at time point t. max It is the maximum plant growth rate in the historical database, E min It is the minimum energy consumption value in the historical database, and α is a weighting coefficient used to control the balance between the objectives of maximizing plant growth rate and minimizing energy consumption.

5. The intelligent electrical environment monitoring system according to claim 4, characterized in that, The training steps of the model training module include: S2-1, the environmental parameter point vector X from the historical parameter training set. t As input to the pre-trained model, the output is a bi-objective benefit prediction index. S2-2, by minimizing the bi-objective benefit prediction index Historical index of dual-objective benefits y h，t The loss between them is used to define a first loss function L(θ), which is used to optimize the parameters of the pre-trained model and obtain the parameter set θ. S2-3, Find the optimal parameter θ in the parameter set θ. * As a parameter prediction model M env The model parameters; the optimal parameter θ * The optimized expression is: The first loss function L(θ) is defined using the mean squared error: in, It is the bi-objective benefit prediction index output by the parametric prediction model, where T is the total number of time points.

6. The intelligent electrical environment monitoring system according to claim 5, characterized in that, The steps for the first index module to output the bi-objective benefit prediction index include: S4-1: Collect environmental parameters at the current time point and directly combine the environmental parameters at the current time point into a vector of current environmental parameter points; S4-2, Receive current environmental parameters through the parameter prediction model, and output the current dual-objective benefit prediction index from the parameter prediction model.

7. The intelligent electrical environment monitoring system according to claim 6, characterized in that, The current index y of the dual-objective benefit o，t The expression is: Among them, G o，t This is the observed plant growth rate in the current environment, E o，t It represents the current energy consumption figures for the environment.

8. The intelligent electrical environment monitoring system according to claim 7, characterized in that, The acquisition steps of the environmental optimal value acquisition module include: S6-1, Define the dual-objective benefit prediction index and the current index y of dual-objective benefits o，t The second loss function L(G) between o，t E o，t ); S6-2, Define the optimal index y for the dual-objective effect c，t The exponential formula, and setting G in the exponential formula c，t initial values ​​and E c，t The initial value; Among them, G c，t The goal is to find the optimal value for plant growth rate, E. c，t The goal is to find the optimal value for current energy consumption, where G is... c，t initial values ​​and E c，t The initial value can be selected from the current observation value, the mean value in the historical database, or the variance value in the historical database; S6-3, Update the optimal value of plant growth rate G using the gradient descent method. c，t And the optimal value of energy consumption E c，t ; The optimal value of plant growth rate G c，t The expression for the partial derivative is: The optimal value of energy consumption E c，t The expression for the partial derivative is: in, This represents the optimal value G of the second loss function relative to the plant growth rate. c，t rate of change, This represents the optimal value E of the second loss function relative to energy consumption. c，t The rate of change; S6-4, Update the optimal value of plant growth rate G using gradient descent. c，t And the optimal value of energy consumption E c，t , until the second loss function converges; Optimal value of plant growth rate G c，t The update expression is: Optimal energy consumption E c，t The update expression is: Where η is the learning rate, and its value is in the range η = [10]. -3 10 -2 This is used to control the duration of each update. and These are the optimal plant growth rate values ​​G during the i-th iteration. c，t And the optimal value of energy consumption E c，t ; S6-5, when the convergence condition is met, the parameter prediction model stops iterating and outputs the optimal value of plant growth rate G. c，t And the optimal value of energy consumption E c，t .

Images