Investigation of Strawberry Irrigation Strategy Based on K-means Clustering Algorithm

Substrate cultivation is the development direction of greenhouse cultivation mode, and it is of great significance to study the modern precision irrigation method for substrate cultivation. On-demand irrigation is one of the main methods of modern precision irrigation. It is inseparable from real-time monitoring and feedback of water content. The water content is an important parameter in modern precision irrigation methods. Moisture sensor based on the principle of frequency domain reflection (FDR) was used in a substrate environment for monitoring water content. The moisture detection module of JZH-0 sensor adopted the principle of FDR, which had good performance on the soil moisture testing. However, it needed to be calibrated before using in different kinds of substrate. In order to further optimize the closed cultivation irrigation method in solar greenhouse and improve the water and fertilizer utilization rate, based on the theoretical research of on-demand irrigation, the existing integrated irrigation system of water and fertilizer, air temperature sensor and soil moisture sensor were mainly used for irrigation and real-time data monitoring and collection. The study on the irrigation strategy of closed greenhouse cultivation of strawberry in greenhouse was carried out by collecting the moisture content of the substrate under the experience of irrigation and the temperature in the greenhouse. The variation of moisture content of substrate in different time scales was firstly summarized. The result showed that the substrate water content of each irrigation rose rapidly, reaching a peak at the end of each irrigation time, with an average increase of about 21.5 percentage points; the second stage decreased rapidly, with an average decrease of 3.5 percentage points in 20 min; the third stage of the trendwas stabilized, with an average decrease of 1.2 percentage points in 30 min. During each irrigation cycle, the water content of the substrate showed a linear downward trend, and the slope of the substrate could better reflect the daily variation of the matrix moisture. During the whole fruiting period of strawberry, the slope was increased with the increase of daily average temperature, which was increased from 0.011 4 to 0.036 5, reflecting the positive significance of temperature increase on crop water demand. Then, the daily variation of the moisture content of substrate was approximated to the daily water requirement of the crop to study the water requirement of the crop. Among the many factors affecting the water demand of crop, temperature with convenient measurement and high accuracy was chosen as the independent variables, because its influence effect was obvious and had strong coupling relationship with light intensity and relative humidity of air. The data was classified by K-means clustering algorithm. According to the classification results, an interval quantitative automatic irrigation strategy based on daily average temperature was proposed in closed greenhouse cultivation: three temperature intervals were set (13, 18]℃, (18, 22.5]℃, (22.5, 27)℃ and in turn corresponded to the daily variation of water content of three substrates as 0.88 percentage points, 1.76 percentage points and 2.63 percentage points, respectively; calculating the daily average temperature of the day, and selecting the appropriate daily variation of moisture content of substrate according to the temperature range to which it belonged and irrigated the next day; it was revised on Monday in a weekly cycle. Theoretically only 4.51 L water was needed per strawberry plant in fruit stage, which can save 15.4% water. Under this strategy theoretically based on K-means clustering algorithm, the water and fertilizer utilization were effectively improved, and the closed greenhouse irrigation methods in solar greenhouses were further enhanced. © 2020, Chinese Society of Agricultural Machinery. All right reserved.