Effects of control sequence optimisation on the performance of bivariate fertiliser applicator

https://doi.org/10.1016/j.compag.2021.106594Get rights and content

Highlights

  • The control sequence optimization (CSO) method was expanded.

  • The effects of CSO were evaluated on a bivariate fertilizer applicator.

  • The optimal control sequence provides accurate and uniform fertiliser application.

  • The impact elements of the transition rate adjustment time were explored.

Abstract

For a bivariate controller of a fertiliser applicator, the optimal control sequence should be determined because it substantially affects the variable-rate fertilisation performance. To evaluate the effects of control sequences on the variable-rate fertilisation performance of a newly developed bivariate fertiliser applicator, we developed control sequence optimisation (CSO) approach and conducted comparative experiment to validate its performance. The test results revealed that the optimised control sequence can improve the fertilisation accuracy and uniformity. The average of the mean relative error and coefficient of variation using CSO were 3.75% and 3.54%, respectively, improving the fertilisation accuracy by 2.77% and the uniformity by 1.51%. The rate transitions of univariate-rate control (without CSO) were much shorter than that of bivariate-rate control (with CSO). Specifically, the rate transitions without CSO ranged from 1.50 to 2.07 s, while those with CSO ranged from 2.53 to 3.37 s in three increment processes. However, the rate transition variation fluctuated as the transition gap increased without CSO, whereas it remained stable around 3.5 s with CSO. The results and analysis show that CSO could efficiently improve the fertilisation performance.

Introduction

Variable-rate fertilisation (VRF) refers to application of fertilisers according to soil status with a number of benefits, which include, but are not limited to, reducing excessive fertilisation, improving spatial differences of soil nutrients, reducing greenhouse gas emissions, and promoting high crop yield (Balafoutis et al., 2017, Hou et al., 2017). The adoption of VRF in the USA reached 63.3% by 2015 (Say et al., 2017). It has also been adopted in Australia, Canada, and some European Union countries given its social, economic, and environmental benefits (Bramley et al., 2019, Robertson et al., 2012, Vecchio et al., 2020). VRF can be implemented through map- or sensor-based approaches (Kim et al., 2008, Maleki et al., 2008). Map-based VRF leverages analyses of crop yield, soil properties, climate change, and other factors over long periods (e.g., multi-year) to achieve scientific fertilisation decisions. The sensor-based approach is a relatively new method, which has made significant progress recently. However, the sensor-based method faces challenges in accurately, reliably, and timely collecting/processing necessary information required for VRF, such as soil status, and crop nutrient conditions. Therefore, the map-based means is the main VRF approach used in practical application.

A typical map-based VRF applicator consists of a global positioning system (GPS) receiver, a controller, and a VRF actuator (Fulton et al., 2005). In working mode, the GPS receiver identifies the current VRF location, and then, the controller drives the actuator through a motor (hydraulic or DC) to adjust the application rate of granular fertiliser according to prescription map associated with the location information VRF (Jia et al., 2014, Su et al., 2015a). Current fertiliser distributors can be categorized by different types of rollers, such as centrifugal, rotary, spiral, and fluted types (Shi et al., 2018, Song et al., 2020). Among them, the fluted roller is widely used given its simple structure and reliable control, as it adjusts the fertiliser rate by changing the rotational speed of the driving shaft. However, the fluted roller has a limited rate adjustment range and presents pulsed application, resulting in fertilisation inaccuracy and non-uniformity (Yuan et al., 2010).

To improve VRF, the bivariate fertiliser applicator (BFA) has been proposed, and extensive research has been conducted on its mechanical structure design, sensing, and system integration. Liu et al. (2010) developed a BFA that adjusts both rotational speed of the driving shaft (n) and the active feed-roll length (L). To improve the BFA performance, Yuan et al. (2010) applied Gaussian process regression to identify the VRF process and optimised the combination of control parameters (L and n) using a genetic algorithm. Field test results have indicated that an optimised control sequence involving L and n can decrease the average error by 4% (Yuan et al., 2011). Su et al. (2015b) designed a device to adjust L of the fluted roller on the driving shaft and applied the device to a seed drill. They verified experimentally that the average coefficient of variation (CV) at five L values was 8.4%, and the transition from 204 to 319 kg/ha took 4.2 s. Chen et al. (2015) developed a BFA based on crop spectral reflection and divided eight fertiliser distributors into two groups. In each group, L was adjusted by a stepper motor, and n was adjusted by driving a DC motor. Subsequent field test results showed a fertilisation control accuracy above 90% (Chen et al., 2018). Alameen et al. (2019) developed a VRF controller by attaching a pneumatic cylinder to a lever handle for fertiliser rate adjustment on a seed drill and conducted laboratory tests to evaluate its performance. Zhang et al. (2019) proposed a control sequence optimisation (CSO) method to improve the fertilisation performance. Laboratory test results on an experimental platform confirmed that the optimised control sequence can increase the control accuracy, improve uniformity, and shorten the response time (∼2 s).

Existing studies mainly focus on the BFA with fertiliser distributor using one straight fluted roller to drive all distributors spaced at a fixed working width. When working on cross-area or -prescription raster in irregular fields, severe overlap usually occurs at the edges (Fulton et al., 2013). The control sequence is essential for proper BFA operation because it influences the fertilisation accuracy and uniformity. Nevertheless, few tests have been conducted to evaluate its influence on the VRF performance.

We developed a BFA with fertiliser distributor using a spiral fluted roller and independent control of each distributor and devised a CSO method using a multi-objective evolutionary algorithm based on decomposition with differential evolution (MOEA/D-DE). In this study, we aimed to evaluate the effect of an optimised control sequence on the VRF performance of the BFA. Specifically, we aimed to 1) expand CSO to spiral fluted roller distributors and 2) evaluate CSO in terms of fertilisation accuracy, uniformity, and fertiliser rate transition (adjustment) time.

Section snippets

System components

The flow diagram of the proposed BFA controller is shown in Fig. 1. To perform experiment on a road, the BFA was mounted on a 5E-904 tractor (rated power of 90 HPM; John Deere, Moline, IL, USA). The BFA was developed at China Agricultural University and can control each fertiliser distributor independently. In addition, the fertiliser rate can be automatically adjusted considering two operational parameters (active feed-roll length L and rotational speed n of the driving shaft). This BFA

Results and discussion

A photograph of the test area is shown in Fig. 6. The fertilisers deposited on the plastic sheet were collected in numbered of 100 boxes and weighted for analysis. According to the test area division (Fig. 5), 10 samples were collected per zone.

Conclusions

Comparison tests without and with CSO were conducted with a developed BFA to evaluate the effects of CSO on the VRF performance. The following conclusions can be drawn from the tests and their results:

  • 1)

    The optimised control sequence obtained from CSO improves the BFA accuracy and uniformity. The average MRE and CV using the developed BFA with CSO were 3.75% and 3.54%, respectively, increasing the fertilisation accuracy by 2.77% and the uniformity by 1.51% compared with no optimisation.

  • 2)

    The rate

CRediT authorship contribution statement

Jiqin Zhang: Data curation, Validation, Visualization, Writing – original draft. Gang Liu: Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Research and Development Program of China (grant number 2017YFD0700503) and Ningxia Natural Science Funds (grant number 2021AAC03119).

References (30)

  • R.G.V. Bramley et al.

    Regional scale application of the precision agriculture thought process to promote improved fertilizer management in the Australian sugar industry

    Precis. Agric.

    (2019)
  • M. Chen et al.

    Design and experiment of variable rate fertilizer applicator based on crop canopy spectral reflectance

    Trans. Chinese Soc. Agric. Mach.

    (2015)
  • M. Chen et al.

    Response characteristics and efficiency of variable rate fertilization based on spectral reflectance

    Int. J. Agric. Biol. Eng.

    (2018)
  • J.P. Fulton et al.

    Rate Response Assessment From Various Granular Vrt Applicators

    Trans. ASAE

    (2005)
  • J.P. Fulton et al.

    Performance Assessment and Model Development of a Variable Rate, Spinner Disc Fertilizer Applicator

    Trans. ASAE

    (2001)
  • Cited by (6)

    View full text