KR-102962759-B1 - AI-Based Nutrient Solution Supply Apparatus Using a Transpiration Model

Abstract

The present invention relates to a device that generates an optimal nutrient solution supply schedule through an AI transpiration model that triangulates weather forecasts, past learning data, and real-time actual measurement data. It is a technology that predicts future environmental conditions to preemptively prevent condensation, dynamically adjusts nutrient solution concentration inversely proportional to transpiration to fundamentally block salt accumulation and physiological disorders within the growing medium, ensures growth stability by inducing forced transpiration through artificial light in the event of insufficient sunlight, and self-evolves into a dedicated model optimized for the facility environment as cultivation cycles are repeated.

Inventors

• 강소영
• 신동철

Assignees

• 신강 주식회사

Dates

Publication Date

20260508

Application Date

20260305

Claims (6)

1. A sensor unit (110) for collecting accumulated actual measurement data from the time of planting of crops to the present; A storage unit (120) that stores life cycle-specific learning data collected in advance for the above crop; A communication unit (130) that receives forecast data from an environment server (500) and It includes a control unit (190) equipped with an artificial intelligence-based production model, A nutrient solution supply device based on an AI increase model, characterized by the above-mentioned increase model generating information on the growth status of a crop according to its current life cycle by first comparing training data and actual data, generating predicted increase information by second comparing training data with weather patterns similar to forecast data, and generating a nutrient solution supply schedule including reference times for sunrise and sunset and reference intervals for supplying and not supplying nutrient solution by third comparing the growth status information and the predicted increase information, and providing analysis through the first to third comparisons.
2. In paragraph 1, The sensor unit includes a weight sensor (111) that detects the weight of the badge and an illuminance sensor (112) that detects real-time light intensity, and It further includes a supply unit (140) that supplies nutrient solution based on the above nutrient solution supply schedule, An AI-based nutrient supply device characterized by the above-mentioned nutrient supply model comparing weight information and light information obtained through a sensor unit with a predicted nutrient supply pattern on a nutrient supply schedule to determine whether there is an abnormality in the system, including errors in forecast data, environmental changes in the cultivation facility, data errors in the sensor unit, and operational abnormalities in the supply unit, and determining whether to correct the nutrient supply schedule based on the determination result.
3. In paragraph 1, The above-mentioned nutrient solution supply device based on an AI nutrient solution supply device is characterized by updating training data by matching the entire actual measurement data collected at the time the crop's life cycle is completed with the regional weather characteristics or the structure of the cultivation facility, and automatically applying a nutrient solution supply correction value specialized for the region or cultivation facility when generating a nutrient solution supply schedule for the next life cycle based on the updated training data.
4. In paragraph 1, The above-mentioned production model calculates the cumulative light amount during a set reference period based on forecast data and actual measurement data, and determines the operating time of the artificial light source (155) by identifying a continuous low light period in which the cumulative light amount falls below the energy threshold for minimum growth of the crop. The above-described control unit controls the operation of an artificial light source to ensure forced transpiration and nutrient absorption of crops according to a nutrient supply schedule, even during periods of insufficient solar radiation caused by the rainy season or winter snowfall.
5. In paragraph 1, The sensor unit above includes a temperature and humidity sensor (113) that generates actual measurement data regarding temperature and humidity information inside the cultivation facility, and The above-mentioned transpiration model generates condensation prediction information by mutually analyzing temperature and humidity trends in forecast data with real-time temperature and humidity information, and A nutrient solution supply device based on an AI production model, characterized in that the above-mentioned control unit controls the operation of an environmental control unit installed in a cultivation facility based on condensation prediction information, and reflects actual measurement data resulting from the control of the environmental control unit into growth status information.
6. In paragraph 1, The sensor unit above includes a nutrient solution sensor (115) that generates actual measurement data regarding nutrient solution supply information, including the electrical conductivity (EC) or acidity (pH) of the supplied nutrient solution. The above-mentioned transpiration model generates concentration prediction information by mutually analyzing transpiration prediction information and nutrient solution supply information, and An AI-based nutrient supply device characterized by the above-mentioned control unit determining whether to correct the target concentration set in the nutrient supply schedule based on concentration prediction information.

Description

AI-Based Nutrient Solution Supply Apparatus Using a Transpiration Model The present invention relates to a nutrient solution supply device based on an AI production model, and more specifically, to a technology that generates an optimal nutrient solution supply schedule through an AI production model that fuses weather forecast data and real-time actual measurement data, and supplies nutrient solution to crops based on the nutrient solution supply schedule. Recently, in the agricultural sector, Smart Farm technology, which intelligently manages crop growth environments by integrating Information and Communication Technology (ICT), is rapidly spreading to address labor shortages caused by an aging population and climate change. The core of a smart farm is to maintain optimal growth conditions by collecting environmental information such as temperature, humidity, and light intensity inside the greenhouse in real time, analyzing it, and remotely or automatically controlling heating and cooling systems, ventilation fans, window opening and closing, and nutrient supply devices. In particular, irrigation and fertilization technologies, which supply nutrients to crops, are decisive factors determining yield and quality. Early nutrient solution supply systems simply provided a fixed amount of water at set times, but they have gradually evolved into methods that determine the timing of supply by detecting solar radiation or the moisture status of the growing medium. In this regard, various patent documents for precise control of facility horticulture have been presented. These include facility control technology based on environmental data (Patent Documents 1, 2), supply amount control technology based on light intensity and temperature and humidity (Patent Documents 3, 5), substrate moisture management technology using a weight sensor (Patent Document 4), and technology for precisely adjusting the concentration (EC) and acidity (pH) of a nutrient solution using artificial intelligence or machine learning (Patent Documents 6, 7). However, despite these technological advancements, existing systems show limitations in accurately predicting transpiration—a physiological indicator of crops—and organically linking environmental control with nutrient supply. In particular, preemptive prevention of condensation caused by sudden weather changes and precise nutrient supply control considering the increase in biomass due to crop growth remain challenges that need to be addressed. FIG. 1 is a diagram illustrating the problems of a simple reactive control structure in conventional smart farms. Patent documents 1 and 2 describe a method in which environmental information such as temperature, humidity, and illuminance is measured in real time through a sensor unit placed inside a greenhouse, transmitted to a server, and actuated by an actuator such as an irrigation pump or ventilation window when the value exceeds or falls short of a pre-set fixed threshold. However, this simple reactive control structure causes the following serious chronic problems. First, there is the issue of nutrient solution supply instability and time lag. Conventional technology relies on post-feedback control, which simply activates the device when currently measured environmental values reach a critical threshold, without considering changes in transpiration—an ecological characteristic of crops—or external weather forecast data. This results in a time lag between when the crops actually require nutrient solution and when the device is activated, leading to an unstable supply condition that fails to provide adequate water and nutrients to the crops in a timely manner. Second, there is the issue of increased mechanical load and reduced durability of drive units. Control methods based on fixed thresholds cause drive units, such as nutrient solution pumps and electric valves, to repeatedly switch on and off at short intervals when environmental values fluctuate slightly near the reference value. This unnecessary repetitive operation places excessive mechanical stress on core hardware, increasing the frequency of failures and drastically shortening the lifespan of the equipment. Consequently, this not only impairs the precision of growth management but also leads to economic losses by increasing system maintenance costs. FIG. 2 is a diagram illustrating the problem of rapid fluctuations in environmental values and the resulting inhibition of growth occurring in a fixed threshold and static database-based control method. Patent documents 1 and 2 employ a structure in which a driving device is activated when the temperature and humidity sensor values inside the greenhouse deviate from a pre-entered fixed threshold or reach a reference value for a growth stage stored in a database. This conventional technology exhibits the following serious technical defects. First, there is the non-uniformity of environmental control and the occurrence of hunting phenomena. A