Visitas: 392
Based on Machine Learning Methods
By: Fudi Chen, Yishuai Du, Tianlong Qiu, Zhe Xu, Li Zhou, Jianping Xu, Ming Sun, Ye Li, and Jianming Sun *
An intelligent variable-flow RAS can rapidly remove suspended solids and reduce ammonia and nitrite generation from the source. The primary purpose of the present study was to develop the circulating pump- drum filter linkage working technique using machine learning methods.
Fishery productivity is facing a massive challenge of declining resources due to environmental pollution and over fishing. A recirculating aquaculture system (RAS) can offer a high degree of environmental control and uses various technologies to carry out physical filtration, biofiltration, and disinfection for water recycling.
The core of a RAS is the water treatment system, which mainly includes micro-screen drum filters, biofilters, oxidation devices, and disinfection devices. Suspended solid particles have been proven to be the leading cause of high turbidity in aquaculture water, which can cause stress reactions and endanger the health of aquatic animals.
As residence time increases, the suspended solids block the breeding facilities and increase chemical oxygen demand. Organic solid waste can be mineralized and decomposed to increase ammonia and nitrite concentrations and increase the load on the nitrification function of the biofilter.
The micro-screen drum filter, which is a physical filter device widely used in RASs, has the characteristics of strong adaptability, minimal floor space, and a high level of automation. In a drum filter, the screen is fixed on a rotating drum frame on the horizontal axis and partially submerged in water; water flows into the drum and radially through the straining cloth, which captures fine particles with a suitable mesh size.
“The micro-screen is the central working part of the drum filter, and the mesh number can directly affect filtration performance.”
Compared with the traditional fixed-flow RAS, the variable-flow RAS can increase the total water circulation to accelerate the water treatment process when organic particles increase, and the ammonia and nitrite then can be eliminated from the source. In addition, the variable-flow RAS consumes a low amount of electricity when the water is relatively clean.
However, manual operation is often used to adjust the circulation pump frequency to determine the appropriate total water circulation in the variable-flow RAS. The manual operation experience may cause the water treatment efficiency to not match the actual situation, resulting in insufficient water processing efficiency or waste of electricity.
For industrial control in recirculating aquaculture, in particular, there is an urgent need to apply machine learning models to improve instrument efficiency and promote the development of intelligent equipment applications.
The primary purpose of the present study was to develop the circulating pump- drum filter linkage working technique using machine learning methods. An intelligent variable-flow RAS can rapidly remove suspended solids and reduce ammonia and nitrite generation from the source.
Materials and Methods
Experimental RAS The experimental RAS used the recirculating aquaculture system of Dalian Huixin Titanium Equipment Development Co., Ltd. for breeding L. vannamei. The control tem collected the water quality indicators by connecting them with the sensors. Water system collected the water quality indicators by connecting them with the sensors.
Water quality changes can be monitored in real time, and the centrifugal pump was controlled by variable-frequency operation using a flow regulation model based on machine learning Variable-Flow Experiment Design The backwash frequency of the drum filter within a unit period (0.5 h) was used to represent overall RAS turbidity, and the variable-flow regulation model was constructed using the backwash frequency and various water quality data.
The intelligent variable-flow RAS technology is implemented by controlling the RAS circulation rate by changing the circulating pump flow rate. The primary purpose of the variableflow RAS is to implement a linkage control technology to model the relationship between the micro-screen drum filter backwash frequency and the circulation flow rate.
Turbidity sensors were placed at the main return pipeline to monitor and record overall RAS water turbidity. Water quality indicators, including water temperature (T), dissolved oxygen (DO), pH, and salinity, were measured by sensors in real time using YSI ProPlus portable sensors.
Establishing a variable-flow circulation strategy was the core task of the experiment, and therefore the circulation rate regulation model was constructed using the optimal classification model based on machine learning to control the variable-flow circulation rate in the RAS.
The application of machine learning methods in aquaculture-
related research is focused mainly on the prediction, classification, and evaluation of water quality indicators such as dissolved oxygen, salinity, pH, ammonia, and nitrite. In the present study, machine learning was used to model the variable-flow regulation strategy.
Research has shown that LSTM can indeed perform well in processing long time series sequences of data. The optimal classification model needs to be relatively simple in order to be applied in the embedded devices.
The variable-flow adjustment strategy in RAS also needs to respond quickly and satisfy the high standard of classification accuracy. All the evaluated indicators of the SVM models demonstrated better results compared with the LSTM model. The gene algorithm contributed the highest accuracy and F1- score among the four optimization.
A larger quantity of data from the running RAS can ensure higher availability and robustness for optimizing the intelligent variable-flow strategy. The continuous variableflow control technology prerequisite is required for the indicators (water quality, backwash frequency, and rearing cycle) to correspond to the ideal circulation volume. Furthermore, the interaction effects between various indicators need to be revealed through experiments and analysis.
Conclusions
Classification models based on machine learning methods between the explanatory variables and the regulation strategy were developed based on experimental data. The LSTM model had the highest accuracy and F1-score and was regarded as the best classification model among ANN methods. Results showed that SVM models required less training time and exhibited higher accuracy compared with ANN models. Finally, the optimal model was GA-SVM, with the highest classification accuracy (training 100%, test 98.95%) and F1-score.
* This is a summarized version developed by the editorial team of Aquaculture Magazine based on the review article titled “Design of an Intelligent Variable-Flow
