KR20240052346A - Optimum control system of plant factory based on digital twin - Google Patents

Abstract

The present invention relates to a digital twin-based plant factory optimal control system, which is based on artificial intelligence (AI) based on big data for plant factory operation obtained from the spatial information of the plant factory and the results of actual cultivation of target plants in the plant factory. : Targeting the plant factory digital twin model created to enable plant cultivation simulation using the Artificial Intelligence method, target plants to be cultivated in the actual plant factory, their yield and harvest period are set as goals, and simulation is performed to determine the harvest period between planting and planting periods. When divided into harvest stages of the growth and harvest stages, the optimal values of the environmental variables required for cultivation of the target plant are calculated for each harvest stage, and the work plan to be carried out for the cultivation of the target plant is developed according to the environmental variables of the calculated optimal values. It is established and transmitted to the controller of the actual plant factory.
According to the present invention, by establishing a work plan to be carried out for cultivation of the target plant according to the environmental variables optimized in the actual plant factory and transmitting it to the controller of the actual plant factory, optimal control and management of the target plant at each stage of growth is possible. As a result, it is possible to maximize the yield and quality improvement and increase in profitability of the target plant. In particular, the optimized environmental variables are calculated to minimize the resources consumed when cultivating the target plant, thereby minimizing the production costs required for plant factory operation. You can.

Description

Optimum control system of plant factory based on digital twin}

The present invention relates to a plant factory control system, and more specifically to a digital twin-based plant factory optimal control system.

Plant factories utilize the Internet of Things (IoT) and big data in plant cultivation technology to automatically manage the plant growth environment in greenhouses using artificial light sources (e.g. LED) remotely using computers or smartphones, without limitations of time and space. This can increase not only production efficiency but also convenience.

The plant factory provides precise management and prediction for each stage of growth of the target plant based on data on environmental information such as temperature, humidity, carbon dioxide (CO ₂ ), rainfall, light intensity, wind direction, wind speed, and soil, as well as growth information of the target plant. It is possible to increase profitability by improving the yield and quality of target plants, and to reduce production costs by efficiently managing the labor and energy (e.g., electric energy, heat energy, etc.) required for plant factory operation.

However, as published in Patent Document 1, existing plant factories collect growth statistics and literature information on target plants, and based on the collected statistics and information, provide environmental information and target plants necessary for cultivating target plants in plant factories. Because it extracts and utilizes growth information, it has the disadvantage of not being able to apply optimized environmental variables in actual plant factories.

KR 10-2018-0003791 A

The present invention is to solve the above-described conventional problems, and the purpose of the present invention is to use big data for plant factory operation obtained from the spatial information of the plant factory and the results of actual cultivation of target plants in the plant factory. This is a plant factory digital twin model created to enable plant cultivation simulation using a machine learning method. Target plants to be cultivated in an actual plant factory, their yield and harvest period are set as targets, and simulation is performed to adjust the harvest period to the sowing, growing and harvesting periods. When divided into harvest stages, the optimal values of the environmental variables required for cultivation of the target plant are calculated for each harvest stage, and a work plan to be carried out for the cultivation of the target plant is established according to the environmental variables of the calculated optimal value, so that the actual plant can be grown. It provides a digital twin-based optimal control system for plant factories that is transmitted to the factory controller.

In order to achieve the purpose of the present invention as described above, the digital twin-based plant factory optimal control system according to the present invention includes a sensor that reads the physical quantity that needs to be controlled for plant cultivation in a greenhouse, a driver that changes the physical quantity, and the A plant factory including a sensor and a controller that controls the operation of the actuator; And a plant factory digital twin model created to enable plant cultivation simulation using a machine learning method based on big data for plant factory operation obtained from the spatial information of the plant factory and the results of actual cultivation of target plants in the plant factory. When the target plants to be cultivated in an actual plant factory, their yield, and harvest period are set as goals and a simulation is performed and the harvest period is divided into the sowing, growth, and harvest stages, the environment required for cultivating the target plants at each harvest stage is It is characterized by being composed of an optimizer that calculates the optimal values of variables, establishes a work plan to be carried out for cultivation of the target plant according to the environmental variables of the calculated optimal values, and transmits it to the controller of the actual plant factory.

In the digital twin-based plant factory optimal control system according to the present invention, the optimizer obtains the physical quantity information of the sensor and the sensor in a determined sampling unit from the results of the plant factory controller controlling the operation of the sensor and the actuator according to the work plan. It is characterized by adding status information, driver status information, and controller status information to big data for plant factory operation.

In the digital twin-based plant factory optimal control system according to the present invention, when the big data for plant factory operation changes, the optimizer performs simulation according to the environmental variables of the optimal values calculated for each harvest stage of the target plant to obtain a new optimal value. It is characterized by calculating environmental variables.

In the digital twin-based plant factory optimal control system according to the present invention, the environmental variable is characterized as a variable for controlling environmental factors inside and outside the plant factory that affect the yield and harvest time when cultivating the target plant.

In the digital twin-based plant factory optimal control system according to the present invention, the environmental variables are calculated to minimize resources consumed when cultivating target plants.

In the digital twin-based plant factory optimal control system according to the present invention, the optimizer determines the environmental variables according to the control location and control time for cultivation of the target plant according to the environmental variables of optimal values calculated for each harvest stage of the target plant. A work plan that applies the optimal value of is established and transmitted to the controller.

According to the present invention, by establishing a work plan to be carried out for the cultivation of the target plant according to the environmental variables optimized in the actual plant factory and transmitting it to the controller of the actual plant factory, optimal control and management of the target plant at each stage of growth is possible. And as a result, it is possible to maximize the yield and quality improvement of target plants and increase profitability.

In particular, the optimized environmental variables include the resources consumed when cultivating the target plant (e.g., human and material resources required for maintenance, management, and replacement of sensors, actuators, and controllers, or the energy required to create a plant cultivation environment inside and outside the plant factory (e.g., , electric energy, heat energy, etc.), it is calculated to minimize the human and material resources required for the supply, so the production costs required for plant factory operation can be minimized.

Figure 1 is an embodiment showing the configuration of a digital twin-based plant factory optimal control system according to the present invention.
Figure 2 is an embodiment showing a work plan of a simulation-based controller.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the attached drawings.

The digital twin-based plant factory optimal control system according to the present invention described below is not limited to the following examples, and can be used by anyone with ordinary knowledge in the relevant technical field without departing from the gist of the technology claimed in the claims. The technical spirit exists to the extent that anyone can change and implement it.

Referring to FIG. 1, the digital twin-based plant factory optimal control system according to the present invention includes a plant factory 100 and an optimizer 200.

The plant factory 100 includes a sensor 110 that reads physical quantities that need to be controlled for plant cultivation in a greenhouse, a driver 120 that changes the physical quantities, and the operation of the sensor 110 and the driver 120. Includes a controller 130 for control.

The sensor 110 reads physical quantities representing environmental factors such as temperature, humidity, carbon dioxide (CO ₂ ), rainfall, light amount, wind direction, wind speed, and soil (e.g., soil acidity, etc.) and converts them into digital signals through analog/digital conversion. It outputs as a temperature sensor, humidity sensor, carbon dioxide sensor, rainfall sensor, light sensor, wind direction sensor, wind speed sensor, soil acidity sensor, etc.

The actuator 120 is a device for controlling the above-described environmental factors, and for example, a heater, cooler, blowing fan, exhaust fan, waterer, culture medium supply, etc. may be used.

The controller 130 operates the sensor 110 and the driver 120, and while operating the sensor 110 and the driver 120, the controller 130 provides physical quantity information, sensor status information, driver status information, and Obtain controller status information and record it in internal memory.

The optimizer 200 performs machine learning based on spatial information of the plant factory 100 and big data for operation of the plant factory 100 obtained from the results of actual cultivation of target plants in the plant factory 100. With the plant factory digital twin model created to enable plant cultivation simulation, the target plants to be cultivated in the actual plant factory (100), their yield and harvest period are set as targets, and the simulation is performed.

The optimizer 200 performs a simulation to calculate the optimal values of the environmental variables necessary for cultivating the target plant at each harvest stage when the harvest period is divided into the sowing stage, the growth stage, and the harvest stage, and the environment of the calculated optimal value is calculated. A work plan to be carried out for cultivation of the target plant is established according to the variables and transmitted to the controller 130 of the actual plant factory 100.

Preferably, the optimizer 200 applies the optimal value of the environmental variable according to the control location and control time for cultivation of the target plant according to the environmental variable with the optimal value calculated for each harvest stage of the target plant. A plan is established and transmitted to the controller 130.

The optimizer 200 may be installed inside or near the plant factory 100, or may be installed and used in a cloud that optimally controls and manages the plant factory 100 remotely.

When the optimizer 200 is installed inside or near the plant factory 100, the controller ( 130).

When the optimizer 200 is mounted on a cloud that optimally controls and manages the plant factory 100 remotely, it can communicate with the controller 130 of the plant factory 100 through a cloud communication network.

The plant factory digital twin model used by the optimizer 200 for simulation is the plant factory 100 obtained from the spatial information of the plant factory 100 and the results of actual cultivation of the target plant in the plant factory 100. It was created to enable plant cultivation simulation using machine learning based on big data for operation.

The plant factory digital twin model was produced using Digital Twin technology, and includes a computer processor, memory unit (or memory device), input device, output device, etc., and is a simulation model (or program), and the computing device is preferably capable of wired or wireless communication with an external device.

The above-mentioned Digital Twin technology performs simulations using a digital twin model that maintains identity and consistency with an actual physical model capable of real-time dynamic analysis or an actual sensor information model, etc. It is known as a technology that can predict future phenomena and control actual physical models or actual sensor information models.

The optimizer 200 determines a sampling unit (e.g., , 1 minute, 5 minutes, 30 minutes, 1 hour, etc.), the physical quantity information of the sensor 110, sensor status information, driver status information, and controller status information are added to the big data for plant factory operation.

The optimizer 200 determines the unit (e.g., 1 minute, 5 minutes, The physical quantity information, sensor status information, driver status information, and controller status information of the sensor 110 acquired over time (30 minutes, 1 hour, etc.) are added to the big data for operating the plant factory 100.

When the big data for the operation of the plant factory 100 changes, the optimizer 200 performs a simulation according to the environmental variables with optimal values calculated for each harvest stage of the target plant and calculates environmental variables with new optimal values.

The environmental variables include environmental factors inside and outside the plant factory 100 (e.g., temperature, humidity, carbon dioxide (CO ₂ ), rainfall, light amount, wind direction, wind speed, soil, etc.) that affect the yield and harvest time when cultivating the target plant. It is desirable that it is a variable to control.

The environmental variables are preferably calculated to minimize the resources consumed when cultivating the target plant.

Resources consumed when cultivating the target plant include, for example, human resources (e.g., labor) and physical resources (e.g., maintenance and management) required for maintenance, management, and replacement of the sensor 110, the actuator 120, and the controller 130. , parts, equipment, etc. required for replacement), or human resources (e.g., labor) and material resources (e.g., labor) and material resources (e.g., labor) required to supply energy (e.g., electric energy, heat energy, etc.) required to create a plant cultivation environment inside and outside the plant factory 100. For example, parts and equipment necessary for energy supply), etc.

The digital twin-based plant factory optimal control system according to the present invention configured as described above operates as follows.

As described above, the optimizer 200 may be installed inside or near the plant factory 100, or may be mounted on a cloud that optimally controls and manages the plant factory 100 remotely.

First, the optimizer 200 sets targets for the target plants to be grown in the actual plant factory 100, their yield and harvest period, and simulates them with a plant factory digital twin model created to enable plant cultivation simulation using a machine learning method. When the harvest period of the target plant is divided into the sowing, growth, and harvest stages, the optimal values of the environmental variables necessary for cultivating the target plant are calculated for each harvest stage.

Once the optimal values of the environmental variables required for cultivation of the target plant are calculated as described above, the optimizer 130 must then proceed to cultivate the target plant according to the environmental variables of the optimal values calculated for each harvest stage of the target plant. A work plan is established and transmitted to the controller 130.

In a preferred embodiment of the present invention, the optimizer 200 optimizes the environmental variables according to the control location and control time for cultivation of the target plant according to the environmental variable with the optimal value calculated for each harvest stage of the target plant. A work plan for applying the values is established and transmitted to the controller 130.

For example, as shown in FIG. 2, the optimizer 200 cultivates tomatoes at a temperature of 10°C at night during the growth period of tomatoes being grown in greenhouse 1 included in the plant factory 100 and during the day. A work plan for growing tomatoes at a temperature of 25°C can be established and transmitted to the controller 130.

In addition, as shown in FIG. 2, the optimizer 200 cultivates cucumbers with 50% humidity at night during the growing season of cucumbers being grown in greenhouse 2 included in the plant factory 100 and during the day. A work plan for growing cucumbers at 90% humidity can be established and transmitted to the controller 130.

In addition, as shown in FIG. 2, the optimizer 200 produces strawberries with a carbon dioxide (CO ₂ ) concentration of 1000 ppm at night during the growing season of strawberries being grown in greenhouse 3 included in the plant factory 100. A work plan for cultivating strawberries with a carbon dioxide (CO ₂ ) concentration of 2000 ppm during the week can be established and transmitted to the controller 130.

When the work plan of the optimizer 120 as described above is transmitted to the controller 130 of the plant factory 100, the controller 130 of the plant factory 100 operates the sensor 110 and the driver 120. The target plant is cultivated by controlling and executing the work plan established by the optimizer 130 according to environmental variables with optimal values calculated for each harvest stage of the target plant.

Meanwhile, the optimizer 200 uses a sampling unit determined from the results of the controller 130 of the plant factory 100 controlling the operation of the sensor 110 and the driver 120 according to the work plan of the optimizer 130. The physical quantity information, sensor status information, driver status information, and controller status information of the sensor 110 acquired (e.g., 1 minute, 5 minutes, 30 minutes, 1 hour, etc.) are stored in big data for operating the plant factory 100. Add.

In this way, when the big data for operating the plant factory 100 changes, the optimizer 200 performs a simulation according to the environmental variables with optimal values calculated for each harvest stage of the target plant to calculate environmental variables with new optimal values. do.

In addition, when a new optimal environmental variable is calculated as described above, the optimizer 200 optimizes the environmental variable according to the control location and control time for cultivation of the target plant according to the environmental variable with the newly calculated optimal value. A new work plan for applying the value is established and transmitted to the controller 130.

While the plant factory 100 operates, the optimizer 200 repeatedly performs the operation of calculating environmental variables with optimal values and establishing and transmitting a work plan that applies the optimal values of the environmental variables.

In this way, according to an embodiment of the present invention, a work plan to be carried out for cultivation of the target plant according to the optimized environmental variables in the actual plant factory 100 is established and sent to the controller 130 of the actual plant factory 100. By transmitting, optimal control and management of the target plant at each stage of growth is possible, and as a result, the yield and quality of the target plant can be improved, as well as the increase in profitability can be maximized.

In particular, the optimized environmental variables include resources consumed when cultivating target plants, such as human resources (e.g., labor) and physical resources required for maintenance, management, and replacement of the sensor 110, actuator 120, and controller 130. Human resources (e.g., labor force) required to supply resources (e.g., parts and equipment required for maintenance, management, and replacement) or energy (e.g., electric energy, heat energy, etc.) required to create a plant cultivation environment inside and outside the plant factory 100. ) and physical resources (e.g., parts and equipment required for energy supply, etc.) are calculated to minimize the production costs required to operate the plant factory 100.

100: plant factory 110: sensor
120: driver 130: controller
200: Optimizer

Claims (6)

A sensor 110 that reads a physical quantity that needs to be controlled for plant cultivation in a greenhouse, a driver 120 that changes the physical quantity, and a controller 130 that controls the operation of the sensor 110 and the driver 120. Plant factory including (100); and
Plant cultivation simulation is possible using a machine learning method based on the spatial information of the plant factory 100 and big data for the operation of the plant factory 100 obtained from the results of actual cultivation of target plants in the plant factory 100. When the target plants to be cultivated in the actual plant factory (100) and their yield and harvest period were set as goals and simulations were performed using the plant factory digital twin model created to distinguish the harvest period into the planting phase, growth phase, and harvest phase. The controller of the actual plant factory 100 calculates the optimal values of the environmental variables necessary for cultivating the target plant at each harvest stage and establishes a work plan to proceed with the cultivation of the target plant according to the environmental variables of the calculated optimal values. Optimizer 200 sending to 130;
A digital twin-based plant factory optimal control system characterized by consisting of. The method of claim 1, wherein the optimizer 200 obtains data in predetermined sampling units from the results of the controller 130 of the plant factory 100 controlling the operation of the sensor 110 and the driver 120 according to the work plan. A digital twin-based plant factory optimal control system characterized by adding physical quantity information, sensor status information, driver status information, and controller status information of one sensor (110) to big data for operation of the plant factory (100). According to claim 1, when the big data for operating the plant factory 100 changes, the optimizer 200 performs a simulation according to the environmental variables of the optimal values calculated for each harvest stage of the target plant to create an environment with a new optimal value. A digital twin-based plant factory optimal control system characterized by calculating variables. The optimal control system for a digital twin-based plant factory according to claim 1, wherein the environmental variables are variables for controlling environmental factors inside and outside the plant factory (100) that affect the yield and harvest time when cultivating the target plant. The optimal control system for a digital twin-based plant factory according to claim 1, wherein the environmental variables are calculated to minimize resources consumed when cultivating the target plant. According to claim 1, the optimizer 200 applies the optimal value of the environmental variable according to the control location and control time for cultivation of the target plant according to the environmental variable with the optimal value calculated for each harvest stage of the target plant. A digital twin-based plant factory optimal control system, characterized in that a work plan is established and transmitted to the controller (130).

Images