July 15, 2022

 

FAMSUN: New technology for quality control in the area of extruded aquafeed

 
 
Figure 1.  FAMSUN H Series Twin-Screw Extruder
 

Some processing parameters in extrusion, such as water addition amount, temperature, pressure and screw speed, have a direct and critical effect on the quality of extruded products.

 

These parameters determine the starch gelatinisation level of feed melt in the extruder, the shearing effects on it, the expansion ratio of floating pellets, the viscosity and cohesiveness of the sinking ones, and their bulk density and water durability.

With the awareness of productivity, cost, animal health and well-being, and the rise of environment sustainability, aquaculture farmers today pay more attention to feed quality factors as bulk density, oil/fat content, water durability and the leaching of oil and other nutrients in water. They also seek for feed with more precise or customised nutritional and physical quality to grow their products and aim for more economical and sustainable benefits.

 

This is a big challenge for the aquafeed industry. For most existing aquafeed mills, the quality control of products and the adjustment of processing parameters always depend on the skill, knowledge and experience of extruder operators.
 

Before launching a new product onto the market, skillful extruder operators go through much trial and error to develop the most suitable production solutions for the new formula, including the parameter adjustment practices of extruders. This means much time, labor and energy investment and much material and water wastage in the process.

 

Diversified feed demands and the precision farming trends call for more future-proof aquafeed production technologies.

 

At FAMSUN, researchers have a major assumption: If a group of reliable mathematical models is available to predict the product quality of an extrusion system, a costly trial production will not be a necessity anymore, and producers will be able to find the best processing parameters to produce their new products efficiently as soon as the new formulas created. The only thing to do is to define the quality parameters for the new products.

 

With the quality prediction models, producers will be able to accelerate the pace of new products onto the market, become more flexible with different, customised formulas, and achieve the best overall equipment effectiveness (OEE) and yield (YE) for their aquafeed mills.

 

In 2021, FAMSUN R&D experts successfully developed a group of mathematical regression models based on solid statistical theory and rich experience in the application of big data technology. The experts used the ANOVA multi-parameter coupling method to perform regression on the processing parameters of the FAMSUN twin-screw extruders. Combined with quality indicators such as bulk density, gelatinisation level and water durability, researchers obtained several mathematical models to predict the quality of extruded aquafeed produced by the twin-screw extrusion system. The prediction models are classified and stored in FAMSUN's quality prediction model database according to feed formula, screw configuration and the diameter of the die hole. As more data on the best aquafeed-production practices are used in model creation, the database grows and benefits more FAMSUN extruder users.

 

The regression models were proven to be able to predict extruded product quality precisely and reliably in the feed mills of FAMSUN partners. The quality parameters of the product produced by the extruder are the same as that predicted based on actual processing parameters.

 

On the other hand, by inputting the product quality requirements, a ‘backstage' database will calculate the recommended prediction models automatically, and provide the best proposal on start-up parameters for new productions. The quality prediction models allow feed manufacturers to respond to customer demands quickly and target excellence in operation.

 

Quality control practices on FAMSUN H Series Twin-Screw Extruder

 

Launched in 2018, FAMSUN H Series Twin-Screw Extruder (see Figure 1) is highlighted by excellent production stability, outstanding formula adaptability and simple operations.

 

So far, there have been over 100 sets of H Series extruders helping to produce successful outcomes at the feed mills of FAMSUN customers in Vietnam, Thailand, South America and other key aquaculture markets in the world.

 

The collaboration between FAMSUN and H series extruder users on the research, tests, trials and application of the aquafeed quality prediction and control technology now has achieved its first-phase goals.

 

Researchers worked out many regression models by using the ANOVA multi-parameter coupling method to test the quality data of extruded products and the corresponding processing parameters, and by taking an influencing factor of p<0.05 and its coefficient. With a MATLAB contour surface graph, the relationship between feed quality and extrusion processing parameters is described visually.

 

For critical quality indicators such as bulk density, gelatinisation level, water durability and water absorption capability, the regression process is performed separately, and the regression model only describes the relationship between an individual quality indicator and the processing parameters.

 

All prediction models are classified and stored in the database according to feed formula and die holes. For new productions, once the targeted product quality is defined and input, the ‘backstage' database will calculate the recommended prediction models and provide the best production control proposal on melt temperature in extruder and steam and water addition amount in conditioning, which ensure aquafeed production in a high-efficiency, low-consumption and less wasteful way. 

  

Table 1 shows a group of quality prediction models of the FAMSUN H series extruder that FAMSUN researched, tested and verified in cooperation with an aquafeed mill in Zhejiang, China.

 

The feed mill produces feed for four kinds of aquatic species. It is available in three or four quality prediction models for each specie that cover quality indicators of bulk density, gelatinisation level, water absorption capacity and water durability, taking an influencing factor of p<0.05 and guaranteeing the R2>0.95.

 

x, y respectively represents the melt temperature in extruder and moisture content of feed in the conditioner, the value is (-2: 0.05: 3), and the corresponding processing parameter of temperature and moisture content are shown in Table 2.
 

Table 1.  Prediction model - Extruded feed quality related to different aquatic species

 

 

Table 2.  Prediction model - Comparison for temperature and moisture content

 
 

With a MATLAB 3D contour surface graph, the visualised relationship between the bulk density of sea bass and the processing parameters of melt temperature and moisture content is shown in Figure 2.

 

Melt temperature and bulk density first show a positive correlation and then a negative one. The moisture content of conditioned feed mash shows a negative correlation with bulk density, followed by a positive correlation. For starch gelatinisation level, it is negatively correlated with melt temperature, followed by positively correlated, and its relationship with moisture content presents an opposite situation (see Figure 3). It is obvious the relationships of gelatinisation level with melt temperature and moisture content are contrary to that of bulk density with melt temperature and moisture content, which verifies the extrusion expertise that a high gelatinisation level always results in a high expansion ratio and low bulk density of the extruded product.  

 


Figure 2.  Bulk density of sea bass in relation to temperature and moisture content

 


Figure 3.  Gelatinisation level of sea bass in relation to temperature and moisture content

 

All prediction models are classified based on formula first before formulas for the same fish species are further classified according to the die holes.

 

With MATLAB calculation, the database will search for the recommended models automatically according to product quality parameters. After defining the targeted quality such as bulk density, gelatinisation level, water absorption capacity, water durability, etc., the ‘backstage' programme will calculate automatically according to the recommended prediction models and output proposal on start-up parameters for the extruder (see Figure 4).        

 


Figure 4.  Database searching and recommended start-up parameters

 

The quality prediction and control database is now available in FAMSUN Extruder Automatic Control System. With this function, operators can have the proper processing parameters to produce qualified products by the soonest time possible and cut loss of production time in trials and demanding quality tests such as for water absorption capacity and water durability.

 

Additionally, an intelligent function helps reduce the dependence on experienced operators and promotes scientific and economical production in feed mills.

 


Figure 5.  FAMSUN Extruder Automatic Control System

 

By combing intelligent quality prediction and control technology with high-level automatic control based on state-of-the-art extruders, FAMSUN aquafeed extrusion solutions allow quick start-up, full-automatic steam-water addition and high-performance quality control in aquafeed extrusion.

 

It also helps to simplify operations and provide the producer with powerful capability, flexibility, and efficiency to meet farmers' current and future demands.

 

- FAMSUN

Video >

Follow Us

FacebookTwitterLinkedIn